AnotherFoxGuy/webrtc-audio-processing
1
0
Fork 0
Code Pull Requests Activity

Update audio_processing module

Browse Source 
Corresponds to upstream commit 524e9b043e7e86fd72353b987c9d5f6a1ebf83e1

Update notes:

 * Pull in third party license file

 * Replace .gypi files with BUILD.gn to keep track of what changes
   upstream

 * Bunch of new filse pulled in as dependencies

 * Won't build yet due to changes needed on top of these
This commit is contained in:
Arun Raghavan
2015-10-13 17:25:22 +05:30
parent 5ae7a5d6cd
commit 753eada3aa
324 changed files with 52533 additions and 16117 deletions
Download Patch File Download Diff File Expand all files Collapse all files

162
webrtc/modules/audio_processing/audio_buffer.h
View File

@ -8,64 +8,156 @@
* be found in the AUTHORS file in the root of the source tree.
*/
#ifndef WEBRTC_MODULES_AUDIO_PROCESSING_MAIN_SOURCE_AUDIO_BUFFER_H_
#define WEBRTC_MODULES_AUDIO_PROCESSING_MAIN_SOURCE_AUDIO_BUFFER_H_
#ifndef WEBRTC_MODULES_AUDIO_PROCESSING_AUDIO_BUFFER_H_
#define WEBRTC_MODULES_AUDIO_PROCESSING_AUDIO_BUFFER_H_
#include "module_common_types.h"
#include "typedefs.h"
#include "webrtc/base/scoped_ptr.h"
#include "webrtc/common_audio/channel_buffer.h"
#include "webrtc/modules/audio_processing/include/audio_processing.h"
#include "webrtc/modules/audio_processing/splitting_filter.h"
#include "webrtc/modules/interface/module_common_types.h"
#include "webrtc/system_wrappers/interface/scoped_vector.h"
#include "webrtc/typedefs.h"
namespace webrtc {
struct AudioChannel;
struct SplitAudioChannel;
class PushSincResampler;
class IFChannelBuffer;
enum Band {
kBand0To8kHz = 0,
kBand8To16kHz = 1,
kBand16To24kHz = 2
};
class AudioBuffer {
public:
AudioBuffer(int max_num_channels, int samples_per_channel);
// TODO(ajm): Switch to take ChannelLayouts.
AudioBuffer(size_t input_num_frames,
int num_input_channels,
size_t process_num_frames,
int num_process_channels,
size_t output_num_frames);
virtual ~AudioBuffer();
int num_channels() const;
int samples_per_channel() const;
int samples_per_split_channel() const;
void set_num_channels(int num_channels);
size_t num_frames() const;
size_t num_frames_per_band() const;
size_t num_keyboard_frames() const;
size_t num_bands() const;
WebRtc_Word16* data(int channel) const;
WebRtc_Word16* low_pass_split_data(int channel) const;
WebRtc_Word16* high_pass_split_data(int channel) const;
WebRtc_Word16* mixed_low_pass_data(int channel) const;
WebRtc_Word16* low_pass_reference(int channel) const;
// Returns a pointer array to the full-band channels.
// Usage:
// channels()[channel][sample].
// Where:
// 0 <= channel < |num_proc_channels_|
// 0 <= sample < |proc_num_frames_|
int16_t* const* channels();
const int16_t* const* channels_const() const;
float* const* channels_f();
const float* const* channels_const_f() const;
WebRtc_Word32* analysis_filter_state1(int channel) const;
WebRtc_Word32* analysis_filter_state2(int channel) const;
WebRtc_Word32* synthesis_filter_state1(int channel) const;
WebRtc_Word32* synthesis_filter_state2(int channel) const;
// Returns a pointer array to the bands for a specific channel.
// Usage:
// split_bands(channel)[band][sample].
// Where:
// 0 <= channel < |num_proc_channels_|
// 0 <= band < |num_bands_|
// 0 <= sample < |num_split_frames_|
int16_t* const* split_bands(int channel);
const int16_t* const* split_bands_const(int channel) const;
float* const* split_bands_f(int channel);
const float* const* split_bands_const_f(int channel) const;
// Returns a pointer array to the channels for a specific band.
// Usage:
// split_channels(band)[channel][sample].
// Where:
// 0 <= band < |num_bands_|
// 0 <= channel < |num_proc_channels_|
// 0 <= sample < |num_split_frames_|
int16_t* const* split_channels(Band band);
const int16_t* const* split_channels_const(Band band) const;
float* const* split_channels_f(Band band);
const float* const* split_channels_const_f(Band band) const;
// Returns a pointer to the ChannelBuffer that encapsulates the full-band
// data.
ChannelBuffer<int16_t>* data();
const ChannelBuffer<int16_t>* data() const;
ChannelBuffer<float>* data_f();
const ChannelBuffer<float>* data_f() const;
// Returns a pointer to the ChannelBuffer that encapsulates the split data.
ChannelBuffer<int16_t>* split_data();
const ChannelBuffer<int16_t>* split_data() const;
ChannelBuffer<float>* split_data_f();
const ChannelBuffer<float>* split_data_f() const;
// Returns a pointer to the low-pass data downmixed to mono. If this data
// isn't already available it re-calculates it.
const int16_t* mixed_low_pass_data();
const int16_t* low_pass_reference(int channel) const;
const float* keyboard_data() const;
void set_activity(AudioFrame::VADActivity activity);
AudioFrame::VADActivity activity();
AudioFrame::VADActivity activity() const;
// Use for int16 interleaved data.
void DeinterleaveFrom(AudioFrame* audioFrame);
void InterleaveTo(AudioFrame* audioFrame) const;
void Mix(int num_mixed_channels);
void CopyAndMixLowPass(int num_mixed_channels);
// If |data_changed| is false, only the non-audio data members will be copied
// to |frame|.
void InterleaveTo(AudioFrame* frame, bool data_changed);
// Use for float deinterleaved data.
void CopyFrom(const float* const* data, const StreamConfig& stream_config);
void CopyTo(const StreamConfig& stream_config, float* const* data);
void CopyLowPassToReference();
// Splits the signal into different bands.
void SplitIntoFrequencyBands();
// Recombine the different bands into one signal.
void MergeFrequencyBands();
private:
const int max_num_channels_;
// Called from DeinterleaveFrom() and CopyFrom().
void InitForNewData();
// The audio is passed into DeinterleaveFrom() or CopyFrom() with input
// format (samples per channel and number of channels).
const size_t input_num_frames_;
const int num_input_channels_;
// The audio is stored by DeinterleaveFrom() or CopyFrom() with processing
// format.
const size_t proc_num_frames_;
const int num_proc_channels_;
// The audio is returned by InterleaveTo() and CopyTo() with output samples
// per channels and the current number of channels. This last one can be
// changed at any time using set_num_channels().
const size_t output_num_frames_;
int num_channels_;
int num_mixed_channels_;
int num_mixed_low_pass_channels_;
const int samples_per_channel_;
int samples_per_split_channel_;
size_t num_bands_;
size_t num_split_frames_;
bool mixed_low_pass_valid_;
bool reference_copied_;
AudioFrame::VADActivity activity_;
WebRtc_Word16* data_;
// TODO(andrew): use vectors here.
AudioChannel* channels_;
SplitAudioChannel* split_channels_;
// TODO(andrew): improve this, we don't need the full 32 kHz space here.
AudioChannel* mixed_low_pass_channels_;
AudioChannel* low_pass_reference_channels_;
const float* keyboard_data_;
rtc::scoped_ptr<IFChannelBuffer> data_;
rtc::scoped_ptr<IFChannelBuffer> split_data_;
rtc::scoped_ptr<SplittingFilter> splitting_filter_;
rtc::scoped_ptr<ChannelBuffer<int16_t> > mixed_low_pass_channels_;
rtc::scoped_ptr<ChannelBuffer<int16_t> > low_pass_reference_channels_;
rtc::scoped_ptr<IFChannelBuffer> input_buffer_;
rtc::scoped_ptr<IFChannelBuffer> output_buffer_;
rtc::scoped_ptr<ChannelBuffer<float> > process_buffer_;
ScopedVector<PushSincResampler> input_resamplers_;
ScopedVector<PushSincResampler> output_resamplers_;
};
} // namespace webrtc
#endif // WEBRTC_MODULES_AUDIO_PROCESSING_MAIN_SOURCE_AUDIO_BUFFER_H_
#endif // WEBRTC_MODULES_AUDIO_PROCESSING_AUDIO_BUFFER_H_