mirror of
https://gitlab.freedesktop.org/gstreamer/gstreamer.git
synced 2024-11-14 13:21:28 +00:00
7ce3fccf25
* Use GST_PARAM_DOC_SHOW_DEFAULT flags for GPU ID related properties * Add doc caps * Add since markers Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/3250>
346 lines
9.6 KiB
C++
346 lines
9.6 KiB
C++
/* GStreamer
|
|
* Copyright (C) 2020 Seungha Yang <seungha@centricular.com>
|
|
*
|
|
* 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., 51 Franklin St, Fifth Floor,
|
|
* Boston, MA 02110-1301, USA.
|
|
*/
|
|
|
|
#ifdef HAVE_CONFIG_H
|
|
#include "config.h"
|
|
#endif
|
|
|
|
#include <gst/gst.h>
|
|
#include "gstmfaudioencoder.h"
|
|
#include <wrl.h>
|
|
#include <string.h>
|
|
|
|
/* *INDENT-OFF* */
|
|
using namespace Microsoft::WRL;
|
|
/* *INDENT-ON* */
|
|
|
|
GST_DEBUG_CATEGORY (gst_mf_audio_encoder_debug);
|
|
#define GST_CAT_DEFAULT gst_mf_audio_encoder_debug
|
|
|
|
/**
|
|
* GstMFAudioEncoder:
|
|
*
|
|
* Base class for MediaFoundation audio encoders
|
|
*
|
|
* Since: 1.22
|
|
*/
|
|
#define gst_mf_audio_encoder_parent_class parent_class
|
|
G_DEFINE_ABSTRACT_TYPE_WITH_CODE (GstMFAudioEncoder, gst_mf_audio_encoder,
|
|
GST_TYPE_AUDIO_ENCODER,
|
|
GST_DEBUG_CATEGORY_INIT (gst_mf_audio_encoder_debug, "mfaudioencoder", 0,
|
|
"mfaudioencoder"));
|
|
|
|
static gboolean gst_mf_audio_encoder_open (GstAudioEncoder * enc);
|
|
static gboolean gst_mf_audio_encoder_close (GstAudioEncoder * enc);
|
|
static gboolean gst_mf_audio_encoder_set_format (GstAudioEncoder * enc,
|
|
GstAudioInfo * info);
|
|
static GstFlowReturn gst_mf_audio_encoder_handle_frame (GstAudioEncoder * enc,
|
|
GstBuffer * buffer);
|
|
static GstFlowReturn gst_mf_audio_encoder_drain (GstAudioEncoder * enc);
|
|
static void gst_mf_audio_encoder_flush (GstAudioEncoder * enc);
|
|
|
|
static void
|
|
gst_mf_audio_encoder_class_init (GstMFAudioEncoderClass * klass)
|
|
{
|
|
GstAudioEncoderClass *audioenc_class = GST_AUDIO_ENCODER_CLASS (klass);
|
|
|
|
audioenc_class->open = GST_DEBUG_FUNCPTR (gst_mf_audio_encoder_open);
|
|
audioenc_class->close = GST_DEBUG_FUNCPTR (gst_mf_audio_encoder_close);
|
|
audioenc_class->set_format =
|
|
GST_DEBUG_FUNCPTR (gst_mf_audio_encoder_set_format);
|
|
audioenc_class->handle_frame =
|
|
GST_DEBUG_FUNCPTR (gst_mf_audio_encoder_handle_frame);
|
|
audioenc_class->flush = GST_DEBUG_FUNCPTR (gst_mf_audio_encoder_flush);
|
|
|
|
gst_type_mark_as_plugin_api (GST_TYPE_MF_AUDIO_ENCODER,
|
|
(GstPluginAPIFlags) 0);
|
|
}
|
|
|
|
static void
|
|
gst_mf_audio_encoder_init (GstMFAudioEncoder * self)
|
|
{
|
|
gst_audio_encoder_set_drainable (GST_AUDIO_ENCODER (self), TRUE);
|
|
}
|
|
|
|
static gboolean
|
|
gst_mf_audio_encoder_open (GstAudioEncoder * enc)
|
|
{
|
|
GstMFAudioEncoder *self = GST_MF_AUDIO_ENCODER (enc);
|
|
GstMFAudioEncoderClass *klass = GST_MF_AUDIO_ENCODER_GET_CLASS (enc);
|
|
GstMFTransformEnumParams enum_params = { 0, };
|
|
MFT_REGISTER_TYPE_INFO output_type;
|
|
|
|
output_type.guidMajorType = MFMediaType_Audio;
|
|
output_type.guidSubtype = klass->codec_id;
|
|
|
|
enum_params.category = MFT_CATEGORY_AUDIO_ENCODER;
|
|
enum_params.enum_flags = klass->enum_flags;
|
|
enum_params.output_typeinfo = &output_type;
|
|
enum_params.device_index = klass->device_index;
|
|
|
|
GST_DEBUG_OBJECT (self, "Create MFT with enum flags 0x%x, device index %d",
|
|
klass->enum_flags, klass->device_index);
|
|
|
|
self->transform = gst_mf_transform_new (&enum_params);
|
|
if (!self->transform) {
|
|
GST_ERROR_OBJECT (self, "Cannot create MFT object");
|
|
return FALSE;
|
|
}
|
|
|
|
return TRUE;
|
|
}
|
|
|
|
static gboolean
|
|
gst_mf_audio_encoder_close (GstAudioEncoder * enc)
|
|
{
|
|
GstMFAudioEncoder *self = GST_MF_AUDIO_ENCODER (enc);
|
|
|
|
gst_clear_object (&self->transform);
|
|
|
|
return TRUE;
|
|
}
|
|
|
|
static gboolean
|
|
gst_mf_audio_encoder_set_format (GstAudioEncoder * enc, GstAudioInfo * info)
|
|
{
|
|
GstMFAudioEncoder *self = GST_MF_AUDIO_ENCODER (enc);
|
|
GstMFAudioEncoderClass *klass = GST_MF_AUDIO_ENCODER_GET_CLASS (enc);
|
|
ComPtr < IMFMediaType > in_type;
|
|
ComPtr < IMFMediaType > out_type;
|
|
|
|
GST_DEBUG_OBJECT (self, "Set format");
|
|
|
|
gst_mf_audio_encoder_drain (enc);
|
|
|
|
if (!gst_mf_transform_open (self->transform)) {
|
|
GST_ERROR_OBJECT (self, "Failed to open MFT");
|
|
return FALSE;
|
|
}
|
|
|
|
g_assert (klass->get_output_type != nullptr);
|
|
if (!klass->get_output_type (self, info, &out_type)) {
|
|
GST_ERROR_OBJECT (self, "subclass failed to set output type");
|
|
return FALSE;
|
|
}
|
|
|
|
gst_mf_dump_attributes (out_type.Get (), "Set output type", GST_LEVEL_DEBUG);
|
|
|
|
if (!gst_mf_transform_set_output_type (self->transform, out_type.Get ())) {
|
|
GST_ERROR_OBJECT (self, "Couldn't set output type");
|
|
return FALSE;
|
|
}
|
|
|
|
g_assert (klass->get_input_type != nullptr);
|
|
if (!klass->get_input_type (self, info, &in_type)) {
|
|
GST_ERROR_OBJECT (self, "subclass didn't provide input type");
|
|
return FALSE;
|
|
}
|
|
|
|
gst_mf_dump_attributes (in_type.Get (), "Set input type", GST_LEVEL_DEBUG);
|
|
|
|
if (!gst_mf_transform_set_input_type (self->transform, in_type.Get ())) {
|
|
GST_ERROR_OBJECT (self, "Couldn't set input media type");
|
|
return FALSE;
|
|
}
|
|
|
|
g_assert (klass->set_src_caps != nullptr);
|
|
if (!klass->set_src_caps (self, info))
|
|
return FALSE;
|
|
|
|
g_assert (klass->frame_samples > 0);
|
|
gst_audio_encoder_set_frame_samples_min (enc, klass->frame_samples);
|
|
gst_audio_encoder_set_frame_samples_max (enc, klass->frame_samples);
|
|
gst_audio_encoder_set_frame_max (enc, 1);
|
|
|
|
/* mediafoundation encoder needs timestamp and duration */
|
|
self->sample_count = 0;
|
|
self->sample_duration_in_mf = gst_util_uint64_scale (klass->frame_samples,
|
|
10000000, GST_AUDIO_INFO_RATE (info));
|
|
|
|
GST_DEBUG_OBJECT (self,
|
|
"Calculated sample duration %" GST_TIME_FORMAT,
|
|
GST_TIME_ARGS (self->sample_duration_in_mf * 100));
|
|
|
|
return TRUE;
|
|
}
|
|
|
|
static gboolean
|
|
gst_mf_audio_encoder_process_input (GstMFAudioEncoder * self,
|
|
GstBuffer * buffer)
|
|
{
|
|
HRESULT hr;
|
|
ComPtr < IMFSample > sample;
|
|
ComPtr < IMFMediaBuffer > media_buffer;
|
|
BYTE *data;
|
|
gboolean res = FALSE;
|
|
GstMapInfo info;
|
|
guint64 timestamp;
|
|
|
|
if (!gst_buffer_map (buffer, &info, GST_MAP_READ)) {
|
|
GST_ELEMENT_ERROR (self,
|
|
RESOURCE, READ, ("Couldn't map input buffer"), (nullptr));
|
|
return FALSE;
|
|
}
|
|
|
|
GST_TRACE_OBJECT (self, "Process buffer %" GST_PTR_FORMAT, buffer);
|
|
|
|
timestamp = self->sample_count * self->sample_duration_in_mf;
|
|
|
|
hr = MFCreateSample (&sample);
|
|
if (!gst_mf_result (hr))
|
|
goto done;
|
|
|
|
hr = MFCreateMemoryBuffer (info.size, &media_buffer);
|
|
if (!gst_mf_result (hr))
|
|
goto done;
|
|
|
|
hr = media_buffer->Lock (&data, nullptr, nullptr);
|
|
if (!gst_mf_result (hr))
|
|
goto done;
|
|
|
|
memcpy (data, info.data, info.size);
|
|
media_buffer->Unlock ();
|
|
|
|
hr = media_buffer->SetCurrentLength (info.size);
|
|
if (!gst_mf_result (hr))
|
|
goto done;
|
|
|
|
hr = sample->AddBuffer (media_buffer.Get ());
|
|
if (!gst_mf_result (hr))
|
|
goto done;
|
|
|
|
hr = sample->SetSampleTime (timestamp);
|
|
if (!gst_mf_result (hr))
|
|
goto done;
|
|
|
|
hr = sample->SetSampleDuration (self->sample_duration_in_mf);
|
|
if (!gst_mf_result (hr))
|
|
goto done;
|
|
|
|
if (!gst_mf_transform_process_input (self->transform, sample.Get ())) {
|
|
GST_ERROR_OBJECT (self, "Failed to process input");
|
|
goto done;
|
|
}
|
|
|
|
self->sample_count++;
|
|
|
|
res = TRUE;
|
|
|
|
done:
|
|
gst_buffer_unmap (buffer, &info);
|
|
|
|
return res;
|
|
}
|
|
|
|
static GstFlowReturn
|
|
gst_mf_audio_encoder_process_output (GstMFAudioEncoder * self)
|
|
{
|
|
GstMFAudioEncoderClass *klass = GST_MF_AUDIO_ENCODER_GET_CLASS (self);
|
|
HRESULT hr;
|
|
BYTE *data = nullptr;
|
|
ComPtr < IMFMediaBuffer > media_buffer;
|
|
ComPtr < IMFSample > sample;
|
|
GstBuffer *buffer;
|
|
GstFlowReturn res = GST_FLOW_ERROR;
|
|
DWORD buffer_len = 0;
|
|
|
|
res = gst_mf_transform_get_output (self->transform, &sample);
|
|
|
|
if (res != GST_FLOW_OK)
|
|
return res;
|
|
|
|
hr = sample->GetBufferByIndex (0, &media_buffer);
|
|
if (!gst_mf_result (hr))
|
|
return GST_FLOW_ERROR;
|
|
|
|
hr = media_buffer->Lock (&data, nullptr, &buffer_len);
|
|
if (!gst_mf_result (hr))
|
|
return GST_FLOW_ERROR;
|
|
|
|
/* Can happen while draining */
|
|
if (buffer_len == 0 || !data) {
|
|
GST_DEBUG_OBJECT (self, "Empty media buffer");
|
|
media_buffer->Unlock ();
|
|
return GST_FLOW_OK;
|
|
}
|
|
|
|
buffer = gst_audio_encoder_allocate_output_buffer (GST_AUDIO_ENCODER (self),
|
|
buffer_len);
|
|
gst_buffer_fill (buffer, 0, data, buffer_len);
|
|
media_buffer->Unlock ();
|
|
|
|
return gst_audio_encoder_finish_frame (GST_AUDIO_ENCODER (self), buffer,
|
|
klass->frame_samples);
|
|
}
|
|
|
|
static GstFlowReturn
|
|
gst_mf_audio_encoder_handle_frame (GstAudioEncoder * enc, GstBuffer * buffer)
|
|
{
|
|
GstMFAudioEncoder *self = GST_MF_AUDIO_ENCODER (enc);
|
|
GstFlowReturn ret;
|
|
|
|
if (!buffer)
|
|
return gst_mf_audio_encoder_drain (enc);
|
|
|
|
if (!gst_mf_audio_encoder_process_input (self, buffer)) {
|
|
GST_ERROR_OBJECT (self, "Failed to process input");
|
|
return GST_FLOW_ERROR;
|
|
}
|
|
|
|
do {
|
|
ret = gst_mf_audio_encoder_process_output (self);
|
|
} while (ret == GST_FLOW_OK);
|
|
|
|
if (ret == GST_MF_TRANSFORM_FLOW_NEED_DATA)
|
|
ret = GST_FLOW_OK;
|
|
|
|
return ret;
|
|
}
|
|
|
|
static GstFlowReturn
|
|
gst_mf_audio_encoder_drain (GstAudioEncoder * enc)
|
|
{
|
|
GstMFAudioEncoder *self = GST_MF_AUDIO_ENCODER (enc);
|
|
GstFlowReturn ret = GST_FLOW_OK;
|
|
|
|
if (!self->transform)
|
|
return GST_FLOW_OK;
|
|
|
|
gst_mf_transform_drain (self->transform);
|
|
|
|
do {
|
|
ret = gst_mf_audio_encoder_process_output (self);
|
|
} while (ret == GST_FLOW_OK);
|
|
|
|
if (ret == GST_MF_TRANSFORM_FLOW_NEED_DATA)
|
|
ret = GST_FLOW_OK;
|
|
|
|
return ret;
|
|
}
|
|
|
|
static void
|
|
gst_mf_audio_encoder_flush (GstAudioEncoder * enc)
|
|
{
|
|
GstMFAudioEncoder *self = GST_MF_AUDIO_ENCODER (enc);
|
|
|
|
if (!self->transform)
|
|
return;
|
|
|
|
gst_mf_transform_flush (self->transform);
|
|
}
|