mirrors/gstreamer
mirrors/gstreamer
1
1
Fork 0
mirror of https://gitlab.freedesktop.org/gstreamer/gstreamer.git synced 2024-11-30 13:41:48 +00:00
Code Issues 1 Projects Releases Wiki Activity
gstreamer/ext/webrtcdsp/gstwebrtcdsp.cpp

1138 lines
38 KiB
C++
Raw Normal View History

webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
/*
* WebRTC Audio Processing Elements
*
* Copyright 2016 Collabora Ltd
* @author: Nicolas Dufresne <nicolas.dufresne@collabora.com>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*
*/
/**
* SECTION:element-webrtcdsp
* @short_description: Audio Filter using WebRTC Audio Processing library
*
* A voice enhancement filter based on WebRTC Audio Processing library. This
* library provides a whide variety of enhancement algorithms. This element
* tries to enable as much as possible. The currently enabled enhancements are
* High Pass Filter, Echo Canceller, Noise Suppression, Automatic Gain Control,
* and some extended filters.
*
* While webrtcdsp element can be used alone, there is an exception for the
* echo canceller. The audio canceller need to be aware of the far end streams
* that are played to loud speakers. For this, you must place a webrtcechoprobe
* element at that far end. Note that the sample rate must match between
* webrtcdsp and the webrtechoprobe. Though, the number of channels can differ.
* The probe is found by the DSP element using it's object name. By default,
* webrtcdsp looks for webrtcechoprobe0, which means it just work if you have
* a single probe and DSP.
*
* The probe can only be used within the same top level GstPipeline.
documentation: fixed a heap o' typos
2019-09-02 19:08:44 +00:00
* Additionally, to simplify the code, the probe element must be created
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
* before the DSP sink pad is activated. It does not need to be in any
* particular state and does not even need to be added to the pipeline yet.
*
* # Example launch line
*
documentation: fixed a heap o' typos
2019-09-02 19:08:44 +00:00
* As a convenience, the echo canceller can be tested using an echo loop. In
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
* this configuration, one would expect a single echo to be heard.
*
* |[
* gst-launch-1.0 pulsesrc ! webrtcdsp ! webrtcechoprobe ! pulsesink
* ]|
*
* In real environment, you'll place the probe before the playback, but only
* process the far end streams. The DSP should be placed as close as possible
* to the audio capture. The following pipeline is astracted and does not
* represent a real pipeline.
*
* |[
* gst-launch-1.0 far-end-src ! audio/x-raw,rate=48000 ! webrtcechoprobe ! pulsesink \
* pulsesrc ! audio/x-raw,rate=48000 ! webrtcdsp ! far-end-sink
* ]|
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "gstwebrtcdsp.h"
#include "gstwebrtcechoprobe.h"
#include <webrtc/modules/audio_processing/include/audio_processing.h>
#include <webrtc/modules/interface/module_common_types.h>
#include <webrtc/system_wrappers/include/trace.h>
GST_DEBUG_CATEGORY (webrtc_dsp_debug);
#define GST_CAT_DEFAULT (webrtc_dsp_debug)
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
#define DEFAULT_TARGET_LEVEL_DBFS 3
#define DEFAULT_COMPRESSION_GAIN_DB 9
#define DEFAULT_STARTUP_MIN_VOLUME 12
#define DEFAULT_LIMITER TRUE
#define DEFAULT_GAIN_CONTROL_MODE webrtc::GainControl::kAdaptiveDigital
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
#define DEFAULT_VOICE_DETECTION FALSE
#define DEFAULT_VOICE_DETECTION_FRAME_SIZE_MS 10
#define DEFAULT_VOICE_DETECTION_LIKELIHOOD webrtc::VoiceDetection::kLowLikelihood
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
static GstStaticPadTemplate gst_webrtc_dsp_sink_template =
GST_STATIC_PAD_TEMPLATE ("sink",
GST_PAD_SINK,
GST_PAD_ALWAYS,
GST_STATIC_CAPS ("audio/x-raw, "
"format = (string) " GST_AUDIO_NE (S16) ", "
"layout = (string) interleaved, "
"rate = (int) { 48000, 32000, 16000, 8000 }, "
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
"channels = (int) [1, MAX];"
"audio/x-raw, "
"format = (string) " GST_AUDIO_NE (F32) ", "
"layout = (string) non-interleaved, "
"rate = (int) { 48000, 32000, 16000, 8000 }, "
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
"channels = (int) [1, MAX]")
);
static GstStaticPadTemplate gst_webrtc_dsp_src_template =
GST_STATIC_PAD_TEMPLATE ("src",
GST_PAD_SRC,
GST_PAD_ALWAYS,
GST_STATIC_CAPS ("audio/x-raw, "
"format = (string) " GST_AUDIO_NE (S16) ", "
"layout = (string) interleaved, "
"rate = (int) { 48000, 32000, 16000, 8000 }, "
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
"channels = (int) [1, MAX];"
"audio/x-raw, "
"format = (string) " GST_AUDIO_NE (F32) ", "
"layout = (string) non-interleaved, "
"rate = (int) { 48000, 32000, 16000, 8000 }, "
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
"channels = (int) [1, MAX]")
);
typedef webrtc::EchoCancellation::SuppressionLevel GstWebrtcEchoSuppressionLevel;
#define GST_TYPE_WEBRTC_ECHO_SUPPRESSION_LEVEL \
(gst_webrtc_echo_suppression_level_get_type ())
static GType
gst_webrtc_echo_suppression_level_get_type (void)
{
static GType suppression_level_type = 0;
static const GEnumValue level_types[] = {
{webrtc::EchoCancellation::kLowSuppression, "Low Suppression", "low"},
{webrtc::EchoCancellation::kModerateSuppression,
"Moderate Suppression", "moderate"},
{webrtc::EchoCancellation::kHighSuppression, "high Suppression", "high"},
{0, NULL, NULL}
};
if (!suppression_level_type) {
suppression_level_type =
g_enum_register_static ("GstWebrtcEchoSuppressionLevel", level_types);
}
return suppression_level_type;
}
typedef webrtc::NoiseSuppression::Level GstWebrtcNoiseSuppressionLevel;
#define GST_TYPE_WEBRTC_NOISE_SUPPRESSION_LEVEL \
(gst_webrtc_noise_suppression_level_get_type ())
static GType
gst_webrtc_noise_suppression_level_get_type (void)
{
static GType suppression_level_type = 0;
static const GEnumValue level_types[] = {
{webrtc::NoiseSuppression::kLow, "Low Suppression", "low"},
{webrtc::NoiseSuppression::kModerate, "Moderate Suppression", "moderate"},
{webrtc::NoiseSuppression::kHigh, "High Suppression", "high"},
{webrtc::NoiseSuppression::kVeryHigh, "Very High Suppression",
"very-high"},
{0, NULL, NULL}
};
if (!suppression_level_type) {
suppression_level_type =
g_enum_register_static ("GstWebrtcNoiseSuppressionLevel", level_types);
}
return suppression_level_type;
}
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
typedef webrtc::GainControl::Mode GstWebrtcGainControlMode;
#define GST_TYPE_WEBRTC_GAIN_CONTROL_MODE \
(gst_webrtc_gain_control_mode_get_type ())
static GType
gst_webrtc_gain_control_mode_get_type (void)
{
static GType gain_control_mode_type = 0;
static const GEnumValue mode_types[] = {
{webrtc::GainControl::kAdaptiveDigital, "Adaptive Digital", "adaptive-digital"},
{webrtc::GainControl::kFixedDigital, "Fixed Digital", "fixed-digital"},
{0, NULL, NULL}
};
if (!gain_control_mode_type) {
gain_control_mode_type =
g_enum_register_static ("GstWebrtcGainControlMode", mode_types);
}
return gain_control_mode_type;
}
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
typedef webrtc::VoiceDetection::Likelihood GstWebrtcVoiceDetectionLikelihood;
#define GST_TYPE_WEBRTC_VOICE_DETECTION_LIKELIHOOD \
(gst_webrtc_voice_detection_likelihood_get_type ())
static GType
gst_webrtc_voice_detection_likelihood_get_type (void)
{
static GType likelihood_type = 0;
static const GEnumValue likelihood_types[] = {
{webrtc::VoiceDetection::kVeryLowLikelihood, "Very Low Likelihood", "very-low"},
{webrtc::VoiceDetection::kLowLikelihood, "Low Likelihood", "low"},
{webrtc::VoiceDetection::kModerateLikelihood, "Moderate Likelihood", "moderate"},
{webrtc::VoiceDetection::kHighLikelihood, "High Likelihood", "high"},
{0, NULL, NULL}
};
if (!likelihood_type) {
likelihood_type =
g_enum_register_static ("GstWebrtcVoiceDetectionLikelihood", likelihood_types);
}
return likelihood_type;
}
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
enum
{
PROP_0,
PROP_PROBE,
PROP_HIGH_PASS_FILTER,
PROP_ECHO_CANCEL,
PROP_ECHO_SUPPRESSION_LEVEL,
PROP_NOISE_SUPPRESSION,
PROP_NOISE_SUPPRESSION_LEVEL,
PROP_GAIN_CONTROL,
PROP_EXPERIMENTAL_AGC,
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
PROP_EXTENDED_FILTER,
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
PROP_DELAY_AGNOSTIC,
PROP_TARGET_LEVEL_DBFS,
PROP_COMPRESSION_GAIN_DB,
PROP_STARTUP_MIN_VOLUME,
PROP_LIMITER,
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
PROP_GAIN_CONTROL_MODE,
PROP_VOICE_DETECTION,
PROP_VOICE_DETECTION_FRAME_SIZE_MS,
PROP_VOICE_DETECTION_LIKELIHOOD,
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
};
/**
* GstWebrtcDSP:
*
* The adder object structure.
*/
struct _GstWebrtcDsp
{
GstAudioFilter element;
/* Protected by the object lock */
GstAudioInfo info;
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
gboolean interleaved;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
guint period_size;
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
guint period_samples;
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
gboolean stream_has_voice;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
/* Protected by the stream lock */
GstAdapter *adapter;
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
GstPlanarAudioAdapter *padapter;
webrtcdsp: Style fix
2016-06-23 01:44:35 +00:00
webrtc::AudioProcessing * apm;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
/* Protected by the object lock */
gchar *probe_name;
GstWebrtcEchoProbe *probe;
/* Properties */
gboolean high_pass_filter;
gboolean echo_cancel;
webrtc::EchoCancellation::SuppressionLevel echo_suppression_level;
gboolean noise_suppression;
webrtc::NoiseSuppression::Level noise_suppression_level;
gboolean gain_control;
gboolean experimental_agc;
gboolean extended_filter;
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
gboolean delay_agnostic;
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
gint target_level_dbfs;
gint compression_gain_db;
gint startup_min_volume;
gboolean limiter;
webrtc::GainControl::Mode gain_control_mode;
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
gboolean voice_detection;
gint voice_detection_frame_size_ms;
webrtc::VoiceDetection::Likelihood voice_detection_likelihood;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
};
webrtcdsp: allow per feature registration Split plugin into features including dynamic types which can be indiviually registered during a static build. More details here: https://gitlab.freedesktop.org/gstreamer/gst-build/-/merge_requests/199 https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/661 Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2038>
2021-02-25 07:18:54 +00:00
G_DEFINE_TYPE_WITH_CODE (GstWebrtcDsp, gst_webrtc_dsp, GST_TYPE_AUDIO_FILTER,
GST_DEBUG_CATEGORY_INIT (webrtc_dsp_debug, "webrtcdsp", 0,
"libwebrtcdsp wrapping elements"););
GST_ELEMENT_REGISTER_DEFINE (webrtcdsp, "webrtcdsp", GST_RANK_NONE,
GST_TYPE_WEBRTC_DSP);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
static const gchar *
webrtc_error_to_string (gint err)
{
documentation: fixed a heap o' typos
2019-09-02 19:08:44 +00:00
const gchar *str = "unknown error";
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
switch (err) {
case webrtc::AudioProcessing::kNoError:
str = "success";
break;
case webrtc::AudioProcessing::kUnspecifiedError:
str = "unspecified error";
break;
case webrtc::AudioProcessing::kCreationFailedError:
str = "creating failed";
break;
case webrtc::AudioProcessing::kUnsupportedComponentError:
str = "unsupported component";
break;
case webrtc::AudioProcessing::kUnsupportedFunctionError:
str = "unsupported function";
break;
case webrtc::AudioProcessing::kNullPointerError:
str = "null pointer";
break;
case webrtc::AudioProcessing::kBadParameterError:
str = "bad parameter";
break;
case webrtc::AudioProcessing::kBadSampleRateError:
str = "bad sample rate";
break;
case webrtc::AudioProcessing::kBadDataLengthError:
str = "bad data length";
break;
case webrtc::AudioProcessing::kBadNumberChannelsError:
str = "bad number of channels";
break;
case webrtc::AudioProcessing::kFileError:
str = "file IO error";
break;
case webrtc::AudioProcessing::kStreamParameterNotSetError:
str = "stream parameter not set";
break;
case webrtc::AudioProcessing::kNotEnabledError:
str = "not enabled";
break;
default:
break;
}
return str;
}
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
static GstBuffer *
gst_webrtc_dsp_take_buffer (GstWebrtcDsp * self)
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
{
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
GstBuffer *buffer;
GstClockTime timestamp;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
guint64 distance;
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
gboolean at_discont;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (self->interleaved) {
timestamp = gst_adapter_prev_pts (self->adapter, &distance);
distance /= self->info.bpf;
} else {
timestamp = gst_planar_audio_adapter_prev_pts (self->padapter, &distance);
}
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
timestamp += gst_util_uint64_scale_int (distance, GST_SECOND, self->info.rate);
if (self->interleaved) {
buffer = gst_adapter_take_buffer (self->adapter, self->period_size);
at_discont = (gst_adapter_pts_at_discont (self->adapter) == timestamp);
} else {
buffer = gst_planar_audio_adapter_take_buffer (self->padapter,
self->period_samples, GST_MAP_READWRITE);
at_discont =
(gst_planar_audio_adapter_pts_at_discont (self->padapter) == timestamp);
}
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
GST_BUFFER_PTS (buffer) = timestamp;
GST_BUFFER_DURATION (buffer) = 10 * GST_MSECOND;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (at_discont && distance == 0) {
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
GST_BUFFER_FLAG_SET (buffer, GST_BUFFER_FLAG_DISCONT);
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
} else {
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
GST_BUFFER_FLAG_UNSET (buffer, GST_BUFFER_FLAG_DISCONT);
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
}
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
return buffer;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
}
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
static GstFlowReturn
gst_webrtc_dsp_analyze_reverse_stream (GstWebrtcDsp * self,
GstClockTime rec_time)
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
{
GstWebrtcEchoProbe *probe = NULL;
webrtc::AudioProcessing * apm;
webrtc::AudioFrame frame;
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
GstBuffer *buf = NULL;
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
GstFlowReturn ret = GST_FLOW_OK;
gint err, delay;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
GST_OBJECT_LOCK (self);
webrtcdsp: We now fail if there is no echo probe When echo cancel is enabled, we now fail the pipeline if there is not echo probe. For this reason there is no need to check if probe pointer is set anymore.
2016-06-29 19:55:19 +00:00
if (self->echo_cancel)
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
probe = GST_WEBRTC_ECHO_PROBE (g_object_ref (self->probe));
GST_OBJECT_UNLOCK (self);
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
/* If echo cancellation is disabled */
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
if (!probe)
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
return GST_FLOW_OK;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
apm = self->apm;
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
if (self->delay_agnostic)
rec_time = GST_CLOCK_TIME_NONE;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
again:
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
delay = gst_webrtc_echo_probe_read (probe, rec_time, (gpointer) &frame, &buf);
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
apm->set_stream_delay_ms (delay);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
if (delay < 0)
goto done;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
if (frame.sample_rate_hz_ != self->info.rate) {
GST_ELEMENT_ERROR (self, STREAM, FORMAT,
("Echo Probe has rate %i , while the DSP is running at rate %i,"
" use a caps filter to ensure those are the same.",
frame.sample_rate_hz_, self->info.rate), (NULL));
ret = GST_FLOW_ERROR;
goto done;
}
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (buf) {
webrtc::StreamConfig config (frame.sample_rate_hz_, frame.num_channels_,
false);
GstAudioBuffer abuf;
float * const * data;
gst_audio_buffer_map (&abuf, &self->info, buf, GST_MAP_READWRITE);
data = (float * const *) abuf.planes;
if ((err = apm->ProcessReverseStream (data, config, config, data)) < 0)
GST_WARNING_OBJECT (self, "Reverse stream analyses failed: %s.",
webrtc_error_to_string (err));
gst_audio_buffer_unmap (&abuf);
gst_buffer_replace (&buf, NULL);
} else {
if ((err = apm->AnalyzeReverseStream (&frame)) < 0)
GST_WARNING_OBJECT (self, "Reverse stream analyses failed: %s.",
webrtc_error_to_string (err));
}
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
if (self->delay_agnostic)
goto again;
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
done:
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
gst_object_unref (probe);
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
gst_buffer_replace (&buf, NULL);
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
return ret;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
}
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
static void
webrtcdsp: Propagate VAD to audio level meta Whenever the voice activity changed on the stream, update or create an AudioLevelMeta and associate it to the corresponding buffer. Fixes #1073 Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2170>
2021-04-18 12:49:59 +00:00
gst_webrtc_vad_post_activity (GstWebrtcDsp *self, GstBuffer *buffer,
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
gboolean stream_has_voice)
{
webrtcdsp: Propagate VAD to audio level meta Whenever the voice activity changed on the stream, update or create an AudioLevelMeta and associate it to the corresponding buffer. Fixes #1073 Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2170>
2021-04-18 12:49:59 +00:00
GstClockTime timestamp = GST_BUFFER_PTS (buffer);
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
GstBaseTransform *trans = GST_BASE_TRANSFORM_CAST (self);
GstStructure *s;
GstClockTime stream_time;
webrtcdsp: Propagate VAD to audio level meta Whenever the voice activity changed on the stream, update or create an AudioLevelMeta and associate it to the corresponding buffer. Fixes #1073 Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2170>
2021-04-18 12:49:59 +00:00
GstAudioLevelMeta *meta;
guint8 level;
level = self->apm->level_estimator ()->RMS ();
meta = gst_buffer_get_audio_level_meta (buffer);
if (meta) {
meta->voice_activity = stream_has_voice;
meta->level = level;
} else {
gst_buffer_add_audio_level_meta (buffer, level, stream_has_voice);
}
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
stream_time = gst_segment_to_stream_time (&trans->segment, GST_FORMAT_TIME,
timestamp);
s = gst_structure_new ("voice-activity",
"stream-time", G_TYPE_UINT64, stream_time,
"stream-has-voice", G_TYPE_BOOLEAN, stream_has_voice, NULL);
GST_LOG_OBJECT (self, "Posting voice activity message, stream %s voice",
stream_has_voice ? "now has" : "no longer has");
gst_element_post_message (GST_ELEMENT (self),
gst_message_new_element (GST_OBJECT (self), s));
}
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
static GstFlowReturn
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
gst_webrtc_dsp_process_stream (GstWebrtcDsp * self,
GstBuffer * buffer)
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
{
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
GstAudioBuffer abuf;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtc::AudioProcessing * apm = self->apm;
gint err;
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (!gst_audio_buffer_map (&abuf, &self->info, buffer,
(GstMapFlags) GST_MAP_READWRITE)) {
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
gst_buffer_unref (buffer);
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
return GST_FLOW_ERROR;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
}
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (self->interleaved) {
webrtc::AudioFrame frame;
frame.num_channels_ = self->info.channels;
frame.sample_rate_hz_ = self->info.rate;
frame.samples_per_channel_ = self->period_samples;
memcpy (frame.data_, abuf.planes[0], self->period_size);
err = apm->ProcessStream (&frame);
if (err >= 0)
memcpy (abuf.planes[0], frame.data_, self->period_size);
} else {
float * const * data = (float * const *) abuf.planes;
webrtc::StreamConfig config (self->info.rate, self->info.channels, false);
err = apm->ProcessStream (data, config, config, data);
}
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (err < 0) {
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
GST_WARNING_OBJECT (self, "Failed to filter the audio: %s.",
webrtc_error_to_string (err));
} else {
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
if (self->voice_detection) {
gboolean stream_has_voice = apm->voice_detection ()->stream_has_voice ();
if (stream_has_voice != self->stream_has_voice)
webrtcdsp: Propagate VAD to audio level meta Whenever the voice activity changed on the stream, update or create an AudioLevelMeta and associate it to the corresponding buffer. Fixes #1073 Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2170>
2021-04-18 12:49:59 +00:00
gst_webrtc_vad_post_activity (self, buffer, stream_has_voice);
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
self->stream_has_voice = stream_has_voice;
}
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
}
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
gst_audio_buffer_unmap (&abuf);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
return GST_FLOW_OK;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
}
static GstFlowReturn
gst_webrtc_dsp_submit_input_buffer (GstBaseTransform * btrans,
gboolean is_discont, GstBuffer * buffer)
{
GstWebrtcDsp *self = GST_WEBRTC_DSP (btrans);
buffer = gst_buffer_make_writable (buffer);
GST_BUFFER_PTS (buffer) = gst_segment_to_running_time (&btrans->segment,
GST_FORMAT_TIME, GST_BUFFER_PTS (buffer));
if (is_discont) {
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
GST_DEBUG_OBJECT (self,
"Received discont, clearing adapter.");
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (self->interleaved)
gst_adapter_clear (self->adapter);
else
gst_planar_audio_adapter_clear (self->padapter);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
}
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (self->interleaved)
gst_adapter_push (self->adapter, buffer);
else
gst_planar_audio_adapter_push (self->padapter, buffer);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
return GST_FLOW_OK;
}
static GstFlowReturn
gst_webrtc_dsp_generate_output (GstBaseTransform * btrans, GstBuffer ** outbuf)
{
GstWebrtcDsp *self = GST_WEBRTC_DSP (btrans);
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
GstFlowReturn ret;
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
gboolean not_enough;
if (self->interleaved)
not_enough = gst_adapter_available (self->adapter) < self->period_size;
else
not_enough = gst_planar_audio_adapter_available (self->padapter) <
self->period_samples;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (not_enough) {
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
*outbuf = NULL;
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
return GST_FLOW_OK;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
}
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
*outbuf = gst_webrtc_dsp_take_buffer (self);
ret = gst_webrtc_dsp_analyze_reverse_stream (self, GST_BUFFER_PTS (*outbuf));
if (ret == GST_FLOW_OK)
ret = gst_webrtc_dsp_process_stream (self, *outbuf);
return ret;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
}
static gboolean
gst_webrtc_dsp_start (GstBaseTransform * btrans)
{
GstWebrtcDsp *self = GST_WEBRTC_DSP (btrans);
webrtc::Config config;
GST_OBJECT_LOCK (self);
config.Set < webrtc::ExtendedFilter >
(new webrtc::ExtendedFilter (self->extended_filter));
config.Set < webrtc::ExperimentalAgc >
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
(new webrtc::ExperimentalAgc (self->experimental_agc, self->startup_min_volume));
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
config.Set < webrtc::DelayAgnostic >
(new webrtc::DelayAgnostic (self->delay_agnostic));
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
/* TODO Intelligibility enhancer, Beamforming, etc. */
self->apm = webrtc::AudioProcessing::Create (config);
if (self->echo_cancel) {
self->probe = gst_webrtc_acquire_echo_probe (self->probe_name);
if (self->probe == NULL) {
GST_OBJECT_UNLOCK (self);
GST_ELEMENT_ERROR (self, RESOURCE, NOT_FOUND,
("No echo probe with name %s found.", self->probe_name), (NULL));
return FALSE;
}
}
GST_OBJECT_UNLOCK (self);
return TRUE;
}
static gboolean
gst_webrtc_dsp_setup (GstAudioFilter * filter, const GstAudioInfo * info)
{
GstWebrtcDsp *self = GST_WEBRTC_DSP (filter);
webrtc::AudioProcessing * apm;
webrtc::ProcessingConfig pconfig;
GstAudioInfo probe_info = *info;
gint err = 0;
GST_LOG_OBJECT (self, "setting format to %s with %i Hz and %i channels",
info->finfo->description, info->rate, info->channels);
GST_OBJECT_LOCK (self);
gst_adapter_clear (self->adapter);
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
gst_planar_audio_adapter_clear (self->padapter);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
self->info = *info;
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
self->interleaved = (info->layout == GST_AUDIO_LAYOUT_INTERLEAVED);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
apm = self->apm;
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (!self->interleaved)
gst_planar_audio_adapter_configure (self->padapter, info);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
/* WebRTC library works with 10ms buffers, compute once this size */
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
self->period_samples = info->rate / 100;
self->period_size = self->period_samples * info->bpf;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
if (self->interleaved &&
(webrtc::AudioFrame::kMaxDataSizeSamples * 2) < self->period_size)
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
goto period_too_big;
if (self->probe) {
GST_WEBRTC_ECHO_PROBE_LOCK (self->probe);
if (self->probe->info.rate != 0) {
if (self->probe->info.rate != info->rate)
goto probe_has_wrong_rate;
probe_info = self->probe->info;
}
GST_WEBRTC_ECHO_PROBE_UNLOCK (self->probe);
}
webrtcdsp: Workaround GCC/MingW 4.7 limitations It refuses to initialize a classes using brackets notation. This is to allow building using our mingw version.
2016-07-05 13:21:58 +00:00
/* input stream */
pconfig.streams[webrtc::ProcessingConfig::kInputStream] =
webrtc::StreamConfig (info->rate, info->channels, false);
/* output stream */
pconfig.streams[webrtc::ProcessingConfig::kOutputStream] =
webrtc::StreamConfig (info->rate, info->channels, false);
/* reverse input stream */
pconfig.streams[webrtc::ProcessingConfig::kReverseInputStream] =
webrtc::StreamConfig (probe_info.rate, probe_info.channels, false);
/* reverse output stream */
pconfig.streams[webrtc::ProcessingConfig::kReverseOutputStream] =
webrtc::StreamConfig (probe_info.rate, probe_info.channels, false);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
if ((err = apm->Initialize (pconfig)) < 0)
goto initialize_failed;
/* Setup Filters */
if (self->high_pass_filter) {
GST_DEBUG_OBJECT (self, "Enabling High Pass filter");
apm->high_pass_filter ()->Enable (true);
}
webrtcdsp: Rewrite echo data synchronization The previous code would run out of sync if there was packet lost or clock skews. When that happened, the echo cancellation feature would completely stop working. As this is crucial for audio calls, this patch re-implement synchronization completely. Instead of letting it drift until next discont, we now synchronize against the record data at every iteration. This way we simply never let the stream drift for longer then 10ms period. We also shorter the delay by using the latency up the probe (basically excluding the sink latency. This is a decent delay to avoid starving in the probe queue. https://bugzilla.gnome.org/show_bug.cgi?id=768009
2016-06-29 19:56:47 +00:00
if (self->echo_cancel) {
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
GST_DEBUG_OBJECT (self, "Enabling Echo Cancellation");
apm->echo_cancellation ()->enable_drift_compensation (false);
apm->echo_cancellation ()
->set_suppression_level (self->echo_suppression_level);
apm->echo_cancellation ()->Enable (true);
}
if (self->noise_suppression) {
GST_DEBUG_OBJECT (self, "Enabling Noise Suppression");
apm->noise_suppression ()->set_level (self->noise_suppression_level);
apm->noise_suppression ()->Enable (true);
}
if (self->gain_control) {
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
GEnumClass *mode_class = (GEnumClass *)
g_type_class_ref (GST_TYPE_WEBRTC_GAIN_CONTROL_MODE);
GST_DEBUG_OBJECT (self, "Enabling Digital Gain Control, target level "
"dBFS %d, compression gain dB %d, limiter %senabled, mode: %s",
self->target_level_dbfs, self->compression_gain_db,
self->limiter ? "" : "NOT ",
g_enum_get_value (mode_class, self->gain_control_mode)->value_name);
g_type_class_unref (mode_class);
apm->gain_control ()->set_mode (self->gain_control_mode);
apm->gain_control ()->set_target_level_dbfs (self->target_level_dbfs);
apm->gain_control ()->set_compression_gain_db (self->compression_gain_db);
apm->gain_control ()->enable_limiter (self->limiter);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
apm->gain_control ()->Enable (true);
}
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
if (self->voice_detection) {
GEnumClass *likelihood_class = (GEnumClass *)
g_type_class_ref (GST_TYPE_WEBRTC_VOICE_DETECTION_LIKELIHOOD);
GST_DEBUG_OBJECT (self, "Enabling Voice Activity Detection, frame size "
"%d milliseconds, likelihood: %s", self->voice_detection_frame_size_ms,
g_enum_get_value (likelihood_class,
self->voice_detection_likelihood)->value_name);
g_type_class_unref (likelihood_class);
self->stream_has_voice = FALSE;
apm->voice_detection ()->Enable (true);
apm->voice_detection ()->set_likelihood (self->voice_detection_likelihood);
apm->voice_detection ()->set_frame_size_ms (
self->voice_detection_frame_size_ms);
webrtcdsp: Propagate VAD to audio level meta Whenever the voice activity changed on the stream, update or create an AudioLevelMeta and associate it to the corresponding buffer. Fixes #1073 Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2170>
2021-04-18 12:49:59 +00:00
apm->level_estimator ()->Enable (true);
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
}
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
GST_OBJECT_UNLOCK (self);
return TRUE;
period_too_big:
GST_OBJECT_UNLOCK (self);
GST_WARNING_OBJECT (self, "webrtcdsp format produce too big period "
"(maximum is %" G_GSIZE_FORMAT " samples and we have %u samples), "
"reduce the number of channels or the rate.",
webrtc::AudioFrame::kMaxDataSizeSamples, self->period_size / 2);
return FALSE;
probe_has_wrong_rate:
GST_WEBRTC_ECHO_PROBE_UNLOCK (self->probe);
GST_OBJECT_UNLOCK (self);
GST_ELEMENT_ERROR (self, STREAM, FORMAT,
("Echo Probe has rate %i , while the DSP is running at rate %i,"
" use a caps filter to ensure those are the same.",
probe_info.rate, info->rate), (NULL));
return FALSE;
initialize_failed:
GST_OBJECT_UNLOCK (self);
GST_ELEMENT_ERROR (self, LIBRARY, INIT,
("Failed to initialize WebRTC Audio Processing library"),
("webrtc::AudioProcessing::Initialize() failed: %s",
webrtc_error_to_string (err)));
return FALSE;
}
static gboolean
gst_webrtc_dsp_stop (GstBaseTransform * btrans)
{
GstWebrtcDsp *self = GST_WEBRTC_DSP (btrans);
GST_OBJECT_LOCK (self);
gst_adapter_clear (self->adapter);
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
gst_planar_audio_adapter_clear (self->padapter);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
if (self->probe) {
gst_webrtc_release_echo_probe (self->probe);
self->probe = NULL;
}
delete self->apm;
self->apm = NULL;
GST_OBJECT_UNLOCK (self);
return TRUE;
}
static void
gst_webrtc_dsp_set_property (GObject * object,
guint prop_id, const GValue * value, GParamSpec * pspec)
{
GstWebrtcDsp *self = GST_WEBRTC_DSP (object);
GST_OBJECT_LOCK (self);
switch (prop_id) {
case PROP_PROBE:
g_free (self->probe_name);
self->probe_name = g_value_dup_string (value);
break;
case PROP_HIGH_PASS_FILTER:
self->high_pass_filter = g_value_get_boolean (value);
break;
case PROP_ECHO_CANCEL:
self->echo_cancel = g_value_get_boolean (value);
break;
case PROP_ECHO_SUPPRESSION_LEVEL:
self->echo_suppression_level =
(GstWebrtcEchoSuppressionLevel) g_value_get_enum (value);
break;
case PROP_NOISE_SUPPRESSION:
self->noise_suppression = g_value_get_boolean (value);
break;
case PROP_NOISE_SUPPRESSION_LEVEL:
self->noise_suppression_level =
(GstWebrtcNoiseSuppressionLevel) g_value_get_enum (value);
break;
case PROP_GAIN_CONTROL:
self->gain_control = g_value_get_boolean (value);
break;
case PROP_EXPERIMENTAL_AGC:
self->experimental_agc = g_value_get_boolean (value);
break;
case PROP_EXTENDED_FILTER:
self->extended_filter = g_value_get_boolean (value);
break;
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
case PROP_DELAY_AGNOSTIC:
self->delay_agnostic = g_value_get_boolean (value);
break;
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
case PROP_TARGET_LEVEL_DBFS:
self->target_level_dbfs = g_value_get_int (value);
break;
case PROP_COMPRESSION_GAIN_DB:
self->compression_gain_db = g_value_get_int (value);
break;
case PROP_STARTUP_MIN_VOLUME:
self->startup_min_volume = g_value_get_int (value);
break;
case PROP_LIMITER:
self->limiter = g_value_get_boolean (value);
break;
case PROP_GAIN_CONTROL_MODE:
self->gain_control_mode =
(GstWebrtcGainControlMode) g_value_get_enum (value);
break;
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
case PROP_VOICE_DETECTION:
self->voice_detection = g_value_get_boolean (value);
break;
case PROP_VOICE_DETECTION_FRAME_SIZE_MS:
self->voice_detection_frame_size_ms = g_value_get_int (value);
break;
case PROP_VOICE_DETECTION_LIKELIHOOD:
self->voice_detection_likelihood =
(GstWebrtcVoiceDetectionLikelihood) g_value_get_enum (value);
break;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
GST_OBJECT_UNLOCK (self);
}
static void
gst_webrtc_dsp_get_property (GObject * object,
guint prop_id, GValue * value, GParamSpec * pspec)
{
GstWebrtcDsp *self = GST_WEBRTC_DSP (object);
GST_OBJECT_LOCK (self);
switch (prop_id) {
case PROP_PROBE:
g_value_set_string (value, self->probe_name);
break;
case PROP_HIGH_PASS_FILTER:
g_value_set_boolean (value, self->high_pass_filter);
break;
case PROP_ECHO_CANCEL:
g_value_set_boolean (value, self->echo_cancel);
break;
case PROP_ECHO_SUPPRESSION_LEVEL:
g_value_set_enum (value, self->echo_suppression_level);
break;
case PROP_NOISE_SUPPRESSION:
g_value_set_boolean (value, self->noise_suppression);
break;
case PROP_NOISE_SUPPRESSION_LEVEL:
g_value_set_enum (value, self->noise_suppression_level);
break;
case PROP_GAIN_CONTROL:
g_value_set_boolean (value, self->gain_control);
break;
case PROP_EXPERIMENTAL_AGC:
g_value_set_boolean (value, self->experimental_agc);
break;
case PROP_EXTENDED_FILTER:
g_value_set_boolean (value, self->extended_filter);
break;
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
case PROP_DELAY_AGNOSTIC:
g_value_set_boolean (value, self->delay_agnostic);
break;
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
case PROP_TARGET_LEVEL_DBFS:
g_value_set_int (value, self->target_level_dbfs);
break;
case PROP_COMPRESSION_GAIN_DB:
g_value_set_int (value, self->compression_gain_db);
break;
case PROP_STARTUP_MIN_VOLUME:
g_value_set_int (value, self->startup_min_volume);
break;
case PROP_LIMITER:
g_value_set_boolean (value, self->limiter);
break;
case PROP_GAIN_CONTROL_MODE:
g_value_set_enum (value, self->gain_control_mode);
break;
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
case PROP_VOICE_DETECTION:
g_value_set_boolean (value, self->voice_detection);
break;
case PROP_VOICE_DETECTION_FRAME_SIZE_MS:
g_value_set_int (value, self->voice_detection_frame_size_ms);
break;
case PROP_VOICE_DETECTION_LIKELIHOOD:
g_value_set_enum (value, self->voice_detection_likelihood);
break;
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
GST_OBJECT_UNLOCK (self);
}
static void
gst_webrtc_dsp_finalize (GObject * object)
{
GstWebrtcDsp *self = GST_WEBRTC_DSP (object);
gst_object_unref (self->adapter);
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
gst_object_unref (self->padapter);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
g_free (self->probe_name);
G_OBJECT_CLASS (gst_webrtc_dsp_parent_class)->finalize (object);
}
static void
gst_webrtc_dsp_init (GstWebrtcDsp * self)
{
self->adapter = gst_adapter_new ();
webrtcdsp: add support for using F32/non-interleaved buffers This is the native format that is in use by the webrtc audio processing library internally, so this avoids internal {de,}interleaving and format conversion (S16->F32 and back) https://bugzilla.gnome.org/show_bug.cgi?id=793605
2018-02-19 16:30:13 +00:00
self->padapter = gst_planar_audio_adapter_new ();
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
gst_audio_info_init (&self->info);
}
static void
gst_webrtc_dsp_class_init (GstWebrtcDspClass * klass)
{
GObjectClass *gobject_class = G_OBJECT_CLASS (klass);
GstElementClass *element_class = GST_ELEMENT_CLASS (klass);
GstBaseTransformClass *btrans_class = GST_BASE_TRANSFORM_CLASS (klass);
GstAudioFilterClass *audiofilter_class = GST_AUDIO_FILTER_CLASS (klass);
gobject_class->finalize = GST_DEBUG_FUNCPTR (gst_webrtc_dsp_finalize);
gobject_class->set_property = GST_DEBUG_FUNCPTR (gst_webrtc_dsp_set_property);
gobject_class->get_property = GST_DEBUG_FUNCPTR (gst_webrtc_dsp_get_property);
btrans_class->passthrough_on_same_caps = FALSE;
btrans_class->start = GST_DEBUG_FUNCPTR (gst_webrtc_dsp_start);
btrans_class->stop = GST_DEBUG_FUNCPTR (gst_webrtc_dsp_stop);
btrans_class->submit_input_buffer =
GST_DEBUG_FUNCPTR (gst_webrtc_dsp_submit_input_buffer);
btrans_class->generate_output =
GST_DEBUG_FUNCPTR (gst_webrtc_dsp_generate_output);
audiofilter_class->setup = GST_DEBUG_FUNCPTR (gst_webrtc_dsp_setup);
gst_element_class_add_static_pad_template (element_class,
&gst_webrtc_dsp_src_template);
gst_element_class_add_static_pad_template (element_class,
&gst_webrtc_dsp_sink_template);
gst_element_class_set_static_metadata (element_class,
"Voice Processor (AGC, AEC, filters, etc.)",
"Generic/Audio",
"Pre-processes voice with WebRTC Audio Processing Library",
"Nicolas Dufresne <nicolas.dufresne@collabora.com>");
g_object_class_install_property (gobject_class,
PROP_PROBE,
g_param_spec_string ("probe", "Echo Probe",
"The name of the webrtcechoprobe element that record the audio being "
"played through loud speakers. Must be set before PAUSED state.",
"webrtcechoprobe0",
(GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_HIGH_PASS_FILTER,
g_param_spec_boolean ("high-pass-filter", "High Pass Filter",
"Enable or disable high pass filtering", TRUE,
(GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_ECHO_CANCEL,
g_param_spec_boolean ("echo-cancel", "Echo Cancel",
webrtcdsp: fix doc string of echo-cancel property If the echo probe element is not found, initialization actually fails instead of silently working with echo-cancel disabled. https://bugzilla.gnome.org/show_bug.cgi?id=780976
2017-04-07 14:13:52 +00:00
"Enable or disable echo canceller", TRUE,
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
(GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_ECHO_SUPPRESSION_LEVEL,
g_param_spec_enum ("echo-suppression-level", "Echo Suppression Level",
"Controls the aggressiveness of the suppressor. A higher level "
"trades off double-talk performance for increased echo suppression.",
GST_TYPE_WEBRTC_ECHO_SUPPRESSION_LEVEL,
webrtc::EchoCancellation::kModerateSuppression,
(GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_NOISE_SUPPRESSION,
g_param_spec_boolean ("noise-suppression", "Noise Suppression",
"Enable or disable noise suppression", TRUE,
(GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_NOISE_SUPPRESSION_LEVEL,
g_param_spec_enum ("noise-suppression-level", "Noise Suppression Level",
"Controls the aggressiveness of the suppression. Increasing the "
"level will reduce the noise level at the expense of a higher "
"speech distortion.", GST_TYPE_WEBRTC_NOISE_SUPPRESSION_LEVEL,
webrtc::EchoCancellation::kModerateSuppression,
(GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_GAIN_CONTROL,
g_param_spec_boolean ("gain-control", "Gain Control",
"Enable or disable automatic digital gain control",
TRUE, (GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_EXPERIMENTAL_AGC,
g_param_spec_boolean ("experimental-agc", "Experimental AGC",
"Enable or disable experimental automatic gain control.",
FALSE, (GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_EXTENDED_FILTER,
g_param_spec_boolean ("extended-filter", "Extended Filter",
"Enable or disable the extended filter.",
TRUE, (GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
webrtcdsp: Add delay-agnostic property In this mode, we let WebRTC Audio Processing figure-out the delay. This is useful when the latency reported by the stack cannot be trusted. Note that in this mode, the leaking of echo during packet lost is much worst. It is recommanded to use PLC (e.g. spanplc, or opus built-in plc). In this mode, we don't do any synchronization. Instead, we simply process all the available reverse stream data as it comes.
2016-07-14 03:00:51 +00:00
g_object_class_install_property (gobject_class,
PROP_DELAY_AGNOSTIC,
g_param_spec_boolean ("delay-agnostic", "Delay Agnostic",
"Enable or disable the delay agnostic mode.",
FALSE, (GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
webrtcdsp: expose new properties. This gives the user more control over the gain control and experimental gain control behaviour. https://bugzilla.gnome.org/show_bug.cgi?id=779103
2017-02-22 01:17:35 +00:00
g_object_class_install_property (gobject_class,
PROP_TARGET_LEVEL_DBFS,
g_param_spec_int ("target-level-dbfs", "Target Level dBFS",
"Sets the target peak |level| (or envelope) of the gain control in "
"dBFS (decibels from digital full-scale).",
0, 31, DEFAULT_TARGET_LEVEL_DBFS, (GParamFlags) (G_PARAM_READWRITE |
G_PARAM_STATIC_STRINGS | G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_COMPRESSION_GAIN_DB,
g_param_spec_int ("compression-gain-db", "Compression Gain dB",
"Sets the maximum |gain| the digital compression stage may apply, "
"in dB.",
0, 90, DEFAULT_COMPRESSION_GAIN_DB, (GParamFlags) (G_PARAM_READWRITE |
G_PARAM_STATIC_STRINGS | G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_STARTUP_MIN_VOLUME,
g_param_spec_int ("startup-min-volume", "Startup Minimum Volume",
"At startup the experimental AGC moves the microphone volume up to "
"|startup_min_volume| if the current microphone volume is set too "
"low. No effect if experimental-agc isn't enabled.",
12, 255, DEFAULT_STARTUP_MIN_VOLUME, (GParamFlags) (G_PARAM_READWRITE |
G_PARAM_STATIC_STRINGS | G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_LIMITER,
g_param_spec_boolean ("limiter", "Limiter",
"When enabled, the compression stage will hard limit the signal to "
"the target level. Otherwise, the signal will be compressed but not "
"limited above the target level.",
DEFAULT_LIMITER, (GParamFlags) (G_PARAM_READWRITE |
G_PARAM_STATIC_STRINGS | G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_GAIN_CONTROL_MODE,
g_param_spec_enum ("gain-control-mode", "Gain Control Mode",
"Controls the mode of the compression stage",
GST_TYPE_WEBRTC_GAIN_CONTROL_MODE,
DEFAULT_GAIN_CONTROL_MODE,
(GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
webrtcdsp: expose voice activity detection. The element now exposes properties to enable and configure voice activity detection, and posts "voice-activity" messages when the return value of stream_has_voice () changes. https://bugzilla.gnome.org/show_bug.cgi?id=779138
2017-02-23 15:12:38 +00:00
g_object_class_install_property (gobject_class,
PROP_VOICE_DETECTION,
g_param_spec_boolean ("voice-detection", "Voice Detection",
"Enable or disable the voice activity detector",
DEFAULT_VOICE_DETECTION, (GParamFlags) (G_PARAM_READWRITE |
G_PARAM_STATIC_STRINGS | G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_VOICE_DETECTION_FRAME_SIZE_MS,
g_param_spec_int ("voice-detection-frame-size-ms",
"Voice Detection Frame Size Milliseconds",
"Sets the |size| of the frames in ms on which the VAD will operate. "
"Larger frames will improve detection accuracy, but reduce the "
"frequency of updates",
10, 30, DEFAULT_VOICE_DETECTION_FRAME_SIZE_MS,
(GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
g_object_class_install_property (gobject_class,
PROP_VOICE_DETECTION_LIKELIHOOD,
g_param_spec_enum ("voice-detection-likelihood",
"Voice Detection Likelihood",
"Specifies the likelihood that a frame will be declared to contain "
"voice.",
GST_TYPE_WEBRTC_VOICE_DETECTION_LIKELIHOOD,
DEFAULT_VOICE_DETECTION_LIKELIHOOD,
(GParamFlags) (G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS |
G_PARAM_CONSTRUCT)));
plugins: uddate gst_type_mark_as_plugin_api() calls
2020-06-05 22:40:42 +00:00
gst_type_mark_as_plugin_api (GST_TYPE_WEBRTC_GAIN_CONTROL_MODE, (GstPluginAPIFlags) 0);
gst_type_mark_as_plugin_api (GST_TYPE_WEBRTC_NOISE_SUPPRESSION_LEVEL, (GstPluginAPIFlags) 0);
gst_type_mark_as_plugin_api (GST_TYPE_WEBRTC_ECHO_SUPPRESSION_LEVEL, (GstPluginAPIFlags) 0);
gst_type_mark_as_plugin_api (GST_TYPE_WEBRTC_VOICE_DETECTION_LIKELIHOOD, (GstPluginAPIFlags) 0);
webrtcdsp: Add WebRTC Audio Processing support This DSP library can be used to enhance voice signal for real time communication call. In implements multiple filters like noise reduction, high pass filter, echo cancellation, automatic gain control, etc. The webrtcdsp element can be used along, or with the help of the webrtcechoprobe if echo cancellation is enabled. The echo probe should be placed as close as possible to the audio sink, while the DSP is generally place close to the audio capture. For local testing, one can use an echo loop pipeline like the following: autoaudiosrc ! webrtcdsp ! webrtcechoprobe ! autoaudiosink This pipeline should produce a single echo rather then repeated echo. Those elements works if they are placed in the same top level pipeline. https://bugzilla.gnome.org/show_bug.cgi?id=767800
2016-06-06 19:10:05 +00:00
}
Reference in a new issue Copy permalink