gstreamer/ext/opus/gstopusdec.c
Vincent Penquerc'h f807905786 opus: properly create channel mapping tables
There are two of them, unintuitively enough; the one passed
to the encoder should not be the one that gets written to the
file. The former maps the input to an ordering which puts
paired channels first, while the latter moves the channels
to Vorbis order. So add code to calculate both, and we now
have properly paired channels where appropriate.

https://bugzilla.gnome.org/show_bug.cgi?id=665078
2011-12-09 15:04:20 +00:00

649 lines
19 KiB
C

/* GStreamer
* Copyright (C) 2004 Wim Taymans <wim@fluendo.com>
* Copyright (C) 2006 Tim-Philipp Müller <tim centricular net>
* Copyright (C) 2008 Sebastian Dröge <sebastian.droege@collabora.co.uk>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 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
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public
* License along with this library; if not, write to the
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
* Boston, MA 02111-1307, USA.
*/
/*
* Based on the speexdec element.
*/
/**
* SECTION:element-opusdec
* @see_also: opusenc, oggdemux
*
* This element decodes a OPUS stream to raw integer audio.
*
* <refsect2>
* <title>Example pipelines</title>
* |[
* gst-launch -v filesrc location=opus.ogg ! oggdemux ! opusdec ! audioconvert ! audioresample ! alsasink
* ]| Decode an Ogg/Opus file. To create an Ogg/Opus file refer to the documentation of opusenc.
* </refsect2>
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <math.h>
#include <string.h>
#include "gstopusheader.h"
#include "gstopuscommon.h"
#include "gstopusdec.h"
GST_DEBUG_CATEGORY_STATIC (opusdec_debug);
#define GST_CAT_DEFAULT opusdec_debug
static GstStaticPadTemplate opus_dec_src_factory =
GST_STATIC_PAD_TEMPLATE ("src",
GST_PAD_SRC,
GST_PAD_ALWAYS,
GST_STATIC_CAPS ("audio/x-raw-int, "
"rate = (int) { 48000, 24000, 16000, 12000, 8000 }, "
"channels = (int) [ 1, 8 ], "
"endianness = (int) BYTE_ORDER, "
"signed = (boolean) true, " "width = (int) 16, " "depth = (int) 16")
);
static GstStaticPadTemplate opus_dec_sink_factory =
GST_STATIC_PAD_TEMPLATE ("sink",
GST_PAD_SINK,
GST_PAD_ALWAYS,
GST_STATIC_CAPS ("audio/x-opus")
);
#define DB_TO_LINEAR(x) pow (10., (x) / 20.)
#define DEFAULT_USE_INBAND_FEC FALSE
#define DEFAULT_APPLY_GAIN TRUE
enum
{
PROP_0,
PROP_USE_INBAND_FEC,
PROP_APPLY_GAIN
};
GST_BOILERPLATE (GstOpusDec, gst_opus_dec, GstAudioDecoder,
GST_TYPE_AUDIO_DECODER);
static GstFlowReturn gst_opus_dec_parse_header (GstOpusDec * dec,
GstBuffer * buf);
static gboolean gst_opus_dec_start (GstAudioDecoder * dec);
static gboolean gst_opus_dec_stop (GstAudioDecoder * dec);
static GstFlowReturn gst_opus_dec_handle_frame (GstAudioDecoder * dec,
GstBuffer * buffer);
static gboolean gst_opus_dec_set_format (GstAudioDecoder * bdec,
GstCaps * caps);
static void gst_opus_dec_get_property (GObject * object, guint prop_id,
GValue * value, GParamSpec * pspec);
static void gst_opus_dec_set_property (GObject * object, guint prop_id,
const GValue * value, GParamSpec * pspec);
static void
gst_opus_dec_base_init (gpointer g_class)
{
GstElementClass *element_class = GST_ELEMENT_CLASS (g_class);
gst_element_class_add_static_pad_template (element_class,
&opus_dec_src_factory);
gst_element_class_add_static_pad_template (element_class,
&opus_dec_sink_factory);
gst_element_class_set_details_simple (element_class, "Opus audio decoder",
"Codec/Decoder/Audio",
"decode opus streams to audio",
"Sebastian Dröge <sebastian.droege@collabora.co.uk>");
}
static void
gst_opus_dec_class_init (GstOpusDecClass * klass)
{
GObjectClass *gobject_class;
GstAudioDecoderClass *adclass;
gobject_class = (GObjectClass *) klass;
adclass = (GstAudioDecoderClass *) klass;
gobject_class->set_property = gst_opus_dec_set_property;
gobject_class->get_property = gst_opus_dec_get_property;
adclass->start = GST_DEBUG_FUNCPTR (gst_opus_dec_start);
adclass->stop = GST_DEBUG_FUNCPTR (gst_opus_dec_stop);
adclass->handle_frame = GST_DEBUG_FUNCPTR (gst_opus_dec_handle_frame);
adclass->set_format = GST_DEBUG_FUNCPTR (gst_opus_dec_set_format);
g_object_class_install_property (gobject_class, PROP_USE_INBAND_FEC,
g_param_spec_boolean ("use-inband-fec", "Use in-band FEC",
"Use forward error correction if available", DEFAULT_USE_INBAND_FEC,
G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
g_object_class_install_property (gobject_class, PROP_APPLY_GAIN,
g_param_spec_boolean ("apply-gain", "Apply gain",
"Apply gain if any is specified in the header", DEFAULT_APPLY_GAIN,
G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
GST_DEBUG_CATEGORY_INIT (opusdec_debug, "opusdec", 0,
"opus decoding element");
}
static void
gst_opus_dec_reset (GstOpusDec * dec)
{
dec->packetno = 0;
if (dec->state) {
opus_multistream_decoder_destroy (dec->state);
dec->state = NULL;
}
gst_buffer_replace (&dec->streamheader, NULL);
gst_buffer_replace (&dec->vorbiscomment, NULL);
gst_buffer_replace (&dec->last_buffer, NULL);
dec->primed = FALSE;
dec->pre_skip = 0;
dec->r128_gain = 0;
}
static void
gst_opus_dec_init (GstOpusDec * dec, GstOpusDecClass * g_class)
{
dec->sample_rate = 0;
dec->n_channels = 0;
dec->use_inband_fec = FALSE;
dec->apply_gain = DEFAULT_APPLY_GAIN;
gst_opus_dec_reset (dec);
}
static gboolean
gst_opus_dec_start (GstAudioDecoder * dec)
{
GstOpusDec *odec = GST_OPUS_DEC (dec);
gst_opus_dec_reset (odec);
/* we know about concealment */
gst_audio_decoder_set_plc_aware (dec, TRUE);
if (odec->use_inband_fec) {
gst_audio_decoder_set_latency (dec, 2 * GST_MSECOND + GST_MSECOND / 2,
120 * GST_MSECOND);
}
return TRUE;
}
static gboolean
gst_opus_dec_stop (GstAudioDecoder * dec)
{
GstOpusDec *odec = GST_OPUS_DEC (dec);
gst_opus_dec_reset (odec);
return TRUE;
}
static double
gst_opus_dec_get_r128_gain (gint16 r128_gain)
{
return r128_gain / (double) (1 << 8);
}
static double
gst_opus_dec_get_r128_volume (gint16 r128_gain)
{
return DB_TO_LINEAR (gst_opus_dec_get_r128_gain (r128_gain));
}
static GstCaps *
gst_opus_dec_negotiate (GstOpusDec * dec)
{
GstCaps *caps = gst_pad_get_allowed_caps (GST_AUDIO_DECODER_SRC_PAD (dec));
GstStructure *s;
caps = gst_caps_make_writable (caps);
gst_caps_truncate (caps);
s = gst_caps_get_structure (caps, 0);
gst_structure_fixate_field_nearest_int (s, "rate", 48000);
gst_structure_get_int (s, "rate", &dec->sample_rate);
gst_structure_fixate_field_nearest_int (s, "channels", dec->n_channels);
gst_structure_get_int (s, "channels", &dec->n_channels);
GST_INFO_OBJECT (dec, "Negotiated %d channels, %d Hz", dec->n_channels,
dec->sample_rate);
return caps;
}
static GstFlowReturn
gst_opus_dec_parse_header (GstOpusDec * dec, GstBuffer * buf)
{
const guint8 *data = GST_BUFFER_DATA (buf);
GstCaps *caps;
const GstAudioChannelPosition *pos = NULL;
g_return_val_if_fail (gst_opus_header_is_id_header (buf), GST_FLOW_ERROR);
g_return_val_if_fail (dec->n_channels == 0
|| dec->n_channels == data[9], GST_FLOW_ERROR);
dec->n_channels = data[9];
dec->pre_skip = GST_READ_UINT16_LE (data + 10);
dec->r128_gain = GST_READ_UINT16_LE (data + 14);
dec->r128_gain_volume = gst_opus_dec_get_r128_volume (dec->r128_gain);
GST_INFO_OBJECT (dec,
"Found pre-skip of %u samples, R128 gain %d (volume %f)",
dec->pre_skip, dec->r128_gain, dec->r128_gain_volume);
dec->channel_mapping_family = data[18];
if (dec->channel_mapping_family == 0) {
/* implicit mapping */
GST_INFO_OBJECT (dec, "Channel mapping family 0, implicit mapping");
dec->n_streams = dec->n_stereo_streams = 1;
dec->channel_mapping[0] = 0;
dec->channel_mapping[1] = 1;
} else {
dec->n_streams = data[19];
dec->n_stereo_streams = data[20];
memcpy (dec->channel_mapping, data + 21, dec->n_channels);
if (dec->channel_mapping_family == 1) {
GST_INFO_OBJECT (dec, "Channel mapping family 1, Vorbis mapping");
switch (dec->n_channels) {
case 1:
case 2:
/* nothing */
break;
case 3:
case 4:
case 5:
case 6:
case 7:
case 8:
pos = gst_opus_channel_positions[dec->n_channels - 1];
break;
default:{
gint i;
GstAudioChannelPosition *posn =
g_new (GstAudioChannelPosition, dec->n_channels);
GST_ELEMENT_WARNING (GST_ELEMENT (dec), STREAM, DECODE,
(NULL), ("Using NONE channel layout for more than 8 channels"));
for (i = 0; i < dec->n_channels; i++)
posn[i] = GST_AUDIO_CHANNEL_POSITION_NONE;
pos = posn;
}
}
} else {
GST_INFO_OBJECT (dec, "Channel mapping family %d",
dec->channel_mapping_family);
}
}
caps = gst_opus_dec_negotiate (dec);
if (pos) {
GST_DEBUG_OBJECT (dec, "Setting channel positions on caps");
gst_audio_set_channel_positions (gst_caps_get_structure (caps, 0), pos);
}
if (dec->n_channels > 8) {
g_free ((GstAudioChannelPosition *) pos);
}
GST_INFO_OBJECT (dec, "Setting src caps to %" GST_PTR_FORMAT, caps);
gst_pad_set_caps (GST_AUDIO_DECODER_SRC_PAD (dec), caps);
gst_caps_unref (caps);
return GST_FLOW_OK;
}
static GstFlowReturn
gst_opus_dec_parse_comments (GstOpusDec * dec, GstBuffer * buf)
{
return GST_FLOW_OK;
}
static GstFlowReturn
opus_dec_chain_parse_data (GstOpusDec * dec, GstBuffer * buffer)
{
GstFlowReturn res = GST_FLOW_OK;
gint size;
guint8 *data;
GstBuffer *outbuf;
gint16 *out_data;
int n, err;
int samples;
unsigned int packet_size;
GstBuffer *buf;
if (dec->state == NULL) {
/* If we did not get any headers, default to 2 channels */
if (dec->n_channels == 0) {
GstCaps *caps;
GST_INFO_OBJECT (dec, "No header, assuming single stream");
dec->n_channels = 2;
dec->sample_rate = 48000;
caps = gst_opus_dec_negotiate (dec);
GST_INFO_OBJECT (dec, "Setting src caps to %" GST_PTR_FORMAT, caps);
gst_pad_set_caps (GST_AUDIO_DECODER_SRC_PAD (dec), caps);
gst_caps_unref (caps);
/* default stereo mapping */
dec->channel_mapping_family = 0;
dec->channel_mapping[0] = 0;
dec->channel_mapping[1] = 1;
dec->n_streams = 1;
dec->n_stereo_streams = 1;
}
GST_DEBUG_OBJECT (dec, "Creating decoder with %d channels, %d Hz",
dec->n_channels, dec->sample_rate);
#ifndef GST_DISABLE_DEBUG
gst_opus_common_log_channel_mapping_table (GST_ELEMENT (dec), opusdec_debug,
"Mapping table", dec->n_channels, dec->channel_mapping);
#endif
GST_DEBUG_OBJECT (dec, "%d streams, %d stereo", dec->n_streams,
dec->n_stereo_streams);
dec->state =
opus_multistream_decoder_create (dec->sample_rate, dec->n_channels,
dec->n_streams, dec->n_stereo_streams, dec->channel_mapping, &err);
if (!dec->state || err != OPUS_OK)
goto creation_failed;
}
if (buffer) {
GST_DEBUG_OBJECT (dec, "Received buffer of size %u",
GST_BUFFER_SIZE (buffer));
} else {
GST_DEBUG_OBJECT (dec, "Received missing buffer");
}
/* if using in-band FEC, we introdude one extra frame's delay as we need
to potentially wait for next buffer to decode a missing buffer */
if (dec->use_inband_fec && !dec->primed) {
GST_DEBUG_OBJECT (dec, "First buffer received in FEC mode, early out");
goto done;
}
/* That's the buffer we'll be sending to the opus decoder. */
buf = dec->use_inband_fec && dec->last_buffer ? dec->last_buffer : buffer;
if (buf) {
data = GST_BUFFER_DATA (buf);
size = GST_BUFFER_SIZE (buf);
GST_DEBUG_OBJECT (dec, "Using buffer of size %u", size);
} else {
/* concealment data, pass NULL as the bits parameters */
GST_DEBUG_OBJECT (dec, "Using NULL buffer");
data = NULL;
size = 0;
}
/* use maximum size (120 ms) as the number of returned samples is
not constant over the stream. */
samples = 120 * dec->sample_rate / 1000;
packet_size = samples * dec->n_channels * 2;
res = gst_pad_alloc_buffer_and_set_caps (GST_AUDIO_DECODER_SRC_PAD (dec),
GST_BUFFER_OFFSET_NONE, packet_size,
GST_PAD_CAPS (GST_AUDIO_DECODER_SRC_PAD (dec)), &outbuf);
if (res != GST_FLOW_OK) {
GST_DEBUG_OBJECT (dec, "buf alloc flow: %s", gst_flow_get_name (res));
return res;
}
out_data = (gint16 *) GST_BUFFER_DATA (outbuf);
if (dec->use_inband_fec) {
if (dec->last_buffer) {
/* normal delayed decode */
n = opus_multistream_decode (dec->state, data, size, out_data, samples,
0);
} else {
/* FEC reconstruction decode */
n = opus_multistream_decode (dec->state, data, size, out_data, samples,
1);
}
} else {
/* normal decode */
n = opus_multistream_decode (dec->state, data, size, out_data, samples, 0);
}
if (n < 0) {
GST_ELEMENT_ERROR (dec, STREAM, DECODE, ("Decoding error: %d", n), (NULL));
return GST_FLOW_ERROR;
}
GST_DEBUG_OBJECT (dec, "decoded %d samples", n);
GST_BUFFER_SIZE (outbuf) = n * 2 * dec->n_channels;
/* Skip any samples that need skipping */
if (dec->pre_skip > 0) {
guint scaled_pre_skip = dec->pre_skip * dec->sample_rate / 48000;
guint skip = scaled_pre_skip > n ? n : scaled_pre_skip;
guint scaled_skip = skip * 48000 / dec->sample_rate;
GST_BUFFER_SIZE (outbuf) -= skip * 2 * dec->n_channels;
GST_BUFFER_DATA (outbuf) += skip * 2 * dec->n_channels;
dec->pre_skip -= scaled_skip;
GST_INFO_OBJECT (dec,
"Skipping %u samples (%u at 48000 Hz, %u left to skip)", skip,
scaled_skip, dec->pre_skip);
}
if (GST_BUFFER_SIZE (outbuf) == 0) {
gst_buffer_unref (outbuf);
outbuf = NULL;
}
/* Apply gain */
/* Would be better off leaving this to a volume element, as this is
a naive conversion that does too many int/float conversions.
However, we don't have control over the pipeline...
So make it optional if the user program wants to use a volume,
but do it by default so the correct volume goes out by default */
if (dec->apply_gain && outbuf && dec->r128_gain) {
unsigned int i, nsamples = GST_BUFFER_SIZE (outbuf) / 2;
double volume = dec->r128_gain_volume;
gint16 *samples = (gint16 *) GST_BUFFER_DATA (outbuf);
GST_DEBUG_OBJECT (dec, "Applying gain: volume %f", volume);
for (i = 0; i < nsamples; ++i) {
int sample = (int) (samples[i] * volume + 0.5);
samples[i] = sample < -32768 ? -32768 : sample > 32767 ? 32767 : sample;
}
}
res = gst_audio_decoder_finish_frame (GST_AUDIO_DECODER (dec), outbuf, 1);
if (res != GST_FLOW_OK)
GST_DEBUG_OBJECT (dec, "flow: %s", gst_flow_get_name (res));
done:
if (dec->use_inband_fec) {
gst_buffer_replace (&dec->last_buffer, buffer);
dec->primed = TRUE;
}
return res;
creation_failed:
GST_ERROR_OBJECT (dec, "Failed to create Opus decoder: %d", err);
return GST_FLOW_ERROR;
}
static gboolean
gst_opus_dec_set_format (GstAudioDecoder * bdec, GstCaps * caps)
{
GstOpusDec *dec = GST_OPUS_DEC (bdec);
gboolean ret = TRUE;
GstStructure *s;
const GValue *streamheader;
GST_DEBUG_OBJECT (dec, "set_format: %" GST_PTR_FORMAT, caps);
s = gst_caps_get_structure (caps, 0);
if ((streamheader = gst_structure_get_value (s, "streamheader")) &&
G_VALUE_HOLDS (streamheader, GST_TYPE_ARRAY) &&
gst_value_array_get_size (streamheader) >= 2) {
const GValue *header, *vorbiscomment;
GstBuffer *buf;
GstFlowReturn res = GST_FLOW_OK;
header = gst_value_array_get_value (streamheader, 0);
if (header && G_VALUE_HOLDS (header, GST_TYPE_BUFFER)) {
buf = gst_value_get_buffer (header);
res = gst_opus_dec_parse_header (dec, buf);
if (res != GST_FLOW_OK)
goto done;
gst_buffer_replace (&dec->streamheader, buf);
}
vorbiscomment = gst_value_array_get_value (streamheader, 1);
if (vorbiscomment && G_VALUE_HOLDS (vorbiscomment, GST_TYPE_BUFFER)) {
buf = gst_value_get_buffer (vorbiscomment);
res = gst_opus_dec_parse_comments (dec, buf);
if (res != GST_FLOW_OK)
goto done;
gst_buffer_replace (&dec->vorbiscomment, buf);
}
}
done:
return ret;
}
static gboolean
memcmp_buffers (GstBuffer * buf1, GstBuffer * buf2)
{
gsize size1, size2;
size1 = GST_BUFFER_SIZE (buf1);
size2 = GST_BUFFER_SIZE (buf2);
if (size1 != size2)
return FALSE;
return !memcmp (GST_BUFFER_DATA (buf1), GST_BUFFER_DATA (buf2), size1);
}
static GstFlowReturn
gst_opus_dec_handle_frame (GstAudioDecoder * adec, GstBuffer * buf)
{
GstFlowReturn res;
GstOpusDec *dec;
/* no fancy draining */
if (G_UNLIKELY (!buf))
return GST_FLOW_OK;
dec = GST_OPUS_DEC (adec);
GST_LOG_OBJECT (dec,
"Got buffer ts %" GST_TIME_FORMAT ", duration %" GST_TIME_FORMAT,
GST_TIME_ARGS (GST_BUFFER_TIMESTAMP (buf)),
GST_TIME_ARGS (GST_BUFFER_DURATION (buf)));
/* If we have the streamheader and vorbiscomment from the caps already
* ignore them here */
if (dec->streamheader && dec->vorbiscomment) {
if (memcmp_buffers (dec->streamheader, buf)) {
GST_DEBUG_OBJECT (dec, "found streamheader");
gst_audio_decoder_finish_frame (adec, NULL, 1);
res = GST_FLOW_OK;
} else if (memcmp_buffers (dec->vorbiscomment, buf)) {
GST_DEBUG_OBJECT (dec, "found vorbiscomments");
gst_audio_decoder_finish_frame (adec, NULL, 1);
res = GST_FLOW_OK;
} else {
res = opus_dec_chain_parse_data (dec, buf);
}
} else {
/* Otherwise fall back to packet counting and assume that the
* first two packets might be the headers, checking magic. */
switch (dec->packetno) {
case 0:
if (gst_opus_header_is_header (buf, "OpusHead", 8)) {
GST_DEBUG_OBJECT (dec, "found streamheader");
res = gst_opus_dec_parse_header (dec, buf);
gst_audio_decoder_finish_frame (adec, NULL, 1);
} else {
res = opus_dec_chain_parse_data (dec, buf);
}
break;
case 1:
if (gst_opus_header_is_header (buf, "OpusTags", 8)) {
GST_DEBUG_OBJECT (dec, "counted vorbiscomments");
res = gst_opus_dec_parse_comments (dec, buf);
gst_audio_decoder_finish_frame (adec, NULL, 1);
} else {
res = opus_dec_chain_parse_data (dec, buf);
}
break;
default:
{
res = opus_dec_chain_parse_data (dec, buf);
break;
}
}
}
dec->packetno++;
return res;
}
static void
gst_opus_dec_get_property (GObject * object, guint prop_id, GValue * value,
GParamSpec * pspec)
{
GstOpusDec *dec = GST_OPUS_DEC (object);
switch (prop_id) {
case PROP_USE_INBAND_FEC:
g_value_set_boolean (value, dec->use_inband_fec);
break;
case PROP_APPLY_GAIN:
g_value_set_boolean (value, dec->apply_gain);
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
static void
gst_opus_dec_set_property (GObject * object, guint prop_id,
const GValue * value, GParamSpec * pspec)
{
GstOpusDec *dec = GST_OPUS_DEC (object);
switch (prop_id) {
case PROP_USE_INBAND_FEC:
dec->use_inband_fec = g_value_get_boolean (value);
break;
case PROP_APPLY_GAIN:
dec->apply_gain = g_value_get_boolean (value);
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}