Adding Opus unit test
This CL adds a unit test for Opus, as well as new APIs for true stereo decoding (skipping master/slave approach). BUG= Review URL: https://webrtc-codereview.appspot.com/1222006 git-svn-id: http://webrtc.googlecode.com/svn/trunk@3860 4adac7df-926f-26a2-2b94-8c16560cd09d
This commit is contained in:
parent
4392d5f9f8
commit
db11fab49e
@ -80,6 +80,7 @@ int WebRtcOpus_DecoderChannels(OpusDecInst* inst);
|
|||||||
* Return value : 0 - Success
|
* Return value : 0 - Success
|
||||||
* -1 - Error
|
* -1 - Error
|
||||||
*/
|
*/
|
||||||
|
int16_t WebRtcOpus_DecoderInitNew(OpusDecInst* inst);
|
||||||
int16_t WebRtcOpus_DecoderInit(OpusDecInst* inst);
|
int16_t WebRtcOpus_DecoderInit(OpusDecInst* inst);
|
||||||
int16_t WebRtcOpus_DecoderInitSlave(OpusDecInst* inst);
|
int16_t WebRtcOpus_DecoderInitSlave(OpusDecInst* inst);
|
||||||
|
|
||||||
@ -103,10 +104,13 @@ int16_t WebRtcOpus_DecoderInitSlave(OpusDecInst* inst);
|
|||||||
* Return value : >0 - Samples in decoded vector
|
* Return value : >0 - Samples in decoded vector
|
||||||
* -1 - Error
|
* -1 - Error
|
||||||
*/
|
*/
|
||||||
int16_t WebRtcOpus_Decode(OpusDecInst* inst, int16_t* encoded,
|
int16_t WebRtcOpus_DecodeNew(OpusDecInst* inst, const uint8_t* encoded,
|
||||||
int16_t encoded_bytes, int16_t* decoded,
|
int16_t encoded_bytes, int16_t* decoded,
|
||||||
int16_t* audio_type);
|
int16_t* audio_type);
|
||||||
int16_t WebRtcOpus_DecodeSlave(OpusDecInst* inst, int16_t* encoded,
|
int16_t WebRtcOpus_Decode(OpusDecInst* inst, const int16_t* encoded,
|
||||||
|
int16_t encoded_bytes, int16_t* decoded,
|
||||||
|
int16_t* audio_type);
|
||||||
|
int16_t WebRtcOpus_DecodeSlave(OpusDecInst* inst, const int16_t* encoded,
|
||||||
int16_t encoded_bytes, int16_t* decoded,
|
int16_t encoded_bytes, int16_t* decoded,
|
||||||
int16_t* audio_type);
|
int16_t* audio_type);
|
||||||
/****************************************************************************
|
/****************************************************************************
|
||||||
|
@ -8,15 +8,15 @@
|
|||||||
* be found in the AUTHORS file in the root of the source tree.
|
* be found in the AUTHORS file in the root of the source tree.
|
||||||
*/
|
*/
|
||||||
|
|
||||||
#include "modules/audio_coding/codecs/opus/interface/opus_interface.h"
|
#include "webrtc/modules/audio_coding/codecs/opus/interface/opus_interface.h"
|
||||||
|
|
||||||
#include <stdlib.h>
|
#include <stdlib.h>
|
||||||
#include <string.h>
|
#include <string.h>
|
||||||
|
|
||||||
#include "opus.h"
|
#include "opus.h"
|
||||||
|
|
||||||
#include "common_audio/signal_processing/resample_by_2_internal.h"
|
#include "webrtc/common_audio/signal_processing/resample_by_2_internal.h"
|
||||||
#include "common_audio/signal_processing/include/signal_processing_library.h"
|
#include "webrtc/common_audio/signal_processing/include/signal_processing_library.h"
|
||||||
|
|
||||||
enum {
|
enum {
|
||||||
/* Maximum supported frame size in WebRTC is 60 ms. */
|
/* Maximum supported frame size in WebRTC is 60 ms. */
|
||||||
@ -31,6 +31,9 @@ enum {
|
|||||||
|
|
||||||
/* Sample count is 48 kHz * samples per frame * stereo. */
|
/* Sample count is 48 kHz * samples per frame * stereo. */
|
||||||
kWebRtcOpusMaxFrameSize = 48 * kWebRtcOpusMaxDecodeFrameSizeMs * 2,
|
kWebRtcOpusMaxFrameSize = 48 * kWebRtcOpusMaxDecodeFrameSizeMs * 2,
|
||||||
|
|
||||||
|
/* Number of samples in resampler state. */
|
||||||
|
kWebRtcOpusStateSize = 7,
|
||||||
};
|
};
|
||||||
|
|
||||||
struct WebRtcOpusEncInst {
|
struct WebRtcOpusEncInst {
|
||||||
@ -39,27 +42,34 @@ struct WebRtcOpusEncInst {
|
|||||||
|
|
||||||
int16_t WebRtcOpus_EncoderCreate(OpusEncInst** inst, int32_t channels) {
|
int16_t WebRtcOpus_EncoderCreate(OpusEncInst** inst, int32_t channels) {
|
||||||
OpusEncInst* state;
|
OpusEncInst* state;
|
||||||
|
if (inst != NULL) {
|
||||||
state = (OpusEncInst*) calloc(1, sizeof(OpusEncInst));
|
state = (OpusEncInst*) calloc(1, sizeof(OpusEncInst));
|
||||||
if (state) {
|
if (state) {
|
||||||
int error;
|
int error;
|
||||||
// Default to VoIP application for mono, and AUDIO for stereo.
|
/* Default to VoIP application for mono, and AUDIO for stereo. */
|
||||||
int application = (channels == 1) ?
|
int application =
|
||||||
OPUS_APPLICATION_VOIP : OPUS_APPLICATION_AUDIO;
|
(channels == 1) ? OPUS_APPLICATION_VOIP : OPUS_APPLICATION_AUDIO;
|
||||||
|
|
||||||
state->encoder = opus_encoder_create(48000, channels, application, &error);
|
state->encoder = opus_encoder_create(48000, channels, application,
|
||||||
if (error == OPUS_OK || state->encoder != NULL ) {
|
&error);
|
||||||
|
if (error == OPUS_OK && state->encoder != NULL) {
|
||||||
*inst = state;
|
*inst = state;
|
||||||
return 0;
|
return 0;
|
||||||
}
|
}
|
||||||
free(state);
|
free(state);
|
||||||
}
|
}
|
||||||
|
}
|
||||||
return -1;
|
return -1;
|
||||||
}
|
}
|
||||||
|
|
||||||
int16_t WebRtcOpus_EncoderFree(OpusEncInst* inst) {
|
int16_t WebRtcOpus_EncoderFree(OpusEncInst* inst) {
|
||||||
|
if (inst) {
|
||||||
opus_encoder_destroy(inst->encoder);
|
opus_encoder_destroy(inst->encoder);
|
||||||
free(inst);
|
free(inst);
|
||||||
return 0;
|
return 0;
|
||||||
|
} else {
|
||||||
|
return -1;
|
||||||
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
int16_t WebRtcOpus_Encode(OpusEncInst* inst, int16_t* audio_in, int16_t samples,
|
int16_t WebRtcOpus_Encode(OpusEncInst* inst, int16_t* audio_in, int16_t samples,
|
||||||
@ -82,7 +92,11 @@ int16_t WebRtcOpus_Encode(OpusEncInst* inst, int16_t* audio_in, int16_t samples,
|
|||||||
}
|
}
|
||||||
|
|
||||||
int16_t WebRtcOpus_SetBitRate(OpusEncInst* inst, int32_t rate) {
|
int16_t WebRtcOpus_SetBitRate(OpusEncInst* inst, int32_t rate) {
|
||||||
|
if (inst) {
|
||||||
return opus_encoder_ctl(inst->encoder, OPUS_SET_BITRATE(rate));
|
return opus_encoder_ctl(inst->encoder, OPUS_SET_BITRATE(rate));
|
||||||
|
} else {
|
||||||
|
return -1;
|
||||||
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
struct WebRtcOpusDecInst {
|
struct WebRtcOpusDecInst {
|
||||||
@ -98,24 +112,25 @@ int16_t WebRtcOpus_DecoderCreate(OpusDecInst** inst, int channels) {
|
|||||||
int error_r;
|
int error_r;
|
||||||
OpusDecInst* state;
|
OpusDecInst* state;
|
||||||
|
|
||||||
// Create Opus decoder memory.
|
if (inst != NULL) {
|
||||||
|
/* Create Opus decoder memory. */
|
||||||
state = (OpusDecInst*) calloc(1, sizeof(OpusDecInst));
|
state = (OpusDecInst*) calloc(1, sizeof(OpusDecInst));
|
||||||
if (state == NULL) {
|
if (state == NULL) {
|
||||||
return -1;
|
return -1;
|
||||||
}
|
}
|
||||||
|
|
||||||
// Create new memory for left and right channel, always at 48000 Hz.
|
/* Create new memory for left and right channel, always at 48000 Hz. */
|
||||||
state->decoder_left = opus_decoder_create(48000, channels, &error_l);
|
state->decoder_left = opus_decoder_create(48000, channels, &error_l);
|
||||||
state->decoder_right = opus_decoder_create(48000, channels, &error_r);
|
state->decoder_right = opus_decoder_create(48000, channels, &error_r);
|
||||||
if (error_l == OPUS_OK && error_r == OPUS_OK && state->decoder_left != NULL
|
if (error_l == OPUS_OK && error_r == OPUS_OK && state->decoder_left != NULL
|
||||||
&& state->decoder_right != NULL) {
|
&& state->decoder_right != NULL) {
|
||||||
// Creation of memory all ok.
|
/* Creation of memory all ok. */
|
||||||
state->channels = channels;
|
state->channels = channels;
|
||||||
*inst = state;
|
*inst = state;
|
||||||
return 0;
|
return 0;
|
||||||
}
|
}
|
||||||
|
|
||||||
// If memory allocation was unsuccessful, free the entire state.
|
/* If memory allocation was unsuccessful, free the entire state. */
|
||||||
if (state->decoder_left) {
|
if (state->decoder_left) {
|
||||||
opus_decoder_destroy(state->decoder_left);
|
opus_decoder_destroy(state->decoder_left);
|
||||||
}
|
}
|
||||||
@ -123,21 +138,35 @@ int16_t WebRtcOpus_DecoderCreate(OpusDecInst** inst, int channels) {
|
|||||||
opus_decoder_destroy(state->decoder_right);
|
opus_decoder_destroy(state->decoder_right);
|
||||||
}
|
}
|
||||||
free(state);
|
free(state);
|
||||||
state = NULL;
|
}
|
||||||
return -1;
|
return -1;
|
||||||
}
|
}
|
||||||
|
|
||||||
int16_t WebRtcOpus_DecoderFree(OpusDecInst* inst) {
|
int16_t WebRtcOpus_DecoderFree(OpusDecInst* inst) {
|
||||||
|
if (inst) {
|
||||||
opus_decoder_destroy(inst->decoder_left);
|
opus_decoder_destroy(inst->decoder_left);
|
||||||
opus_decoder_destroy(inst->decoder_right);
|
opus_decoder_destroy(inst->decoder_right);
|
||||||
free(inst);
|
free(inst);
|
||||||
return 0;
|
return 0;
|
||||||
|
} else {
|
||||||
|
return -1;
|
||||||
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
int WebRtcOpus_DecoderChannels(OpusDecInst* inst) {
|
int WebRtcOpus_DecoderChannels(OpusDecInst* inst) {
|
||||||
return inst->channels;
|
return inst->channels;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
int16_t WebRtcOpus_DecoderInitNew(OpusDecInst* inst) {
|
||||||
|
int error = opus_decoder_ctl(inst->decoder_left, OPUS_RESET_STATE);
|
||||||
|
if (error == OPUS_OK) {
|
||||||
|
memset(inst->state_48_32_left, 0, sizeof(inst->state_48_32_left));
|
||||||
|
memset(inst->state_48_32_right, 0, sizeof(inst->state_48_32_right));
|
||||||
|
return 0;
|
||||||
|
}
|
||||||
|
return -1;
|
||||||
|
}
|
||||||
|
|
||||||
int16_t WebRtcOpus_DecoderInit(OpusDecInst* inst) {
|
int16_t WebRtcOpus_DecoderInit(OpusDecInst* inst) {
|
||||||
int error = opus_decoder_ctl(inst->decoder_left, OPUS_RESET_STATE);
|
int error = opus_decoder_ctl(inst->decoder_left, OPUS_RESET_STATE);
|
||||||
if (error == OPUS_OK) {
|
if (error == OPUS_OK) {
|
||||||
@ -156,7 +185,7 @@ int16_t WebRtcOpus_DecoderInitSlave(OpusDecInst* inst) {
|
|||||||
return -1;
|
return -1;
|
||||||
}
|
}
|
||||||
|
|
||||||
static int DecodeNative(OpusDecoder* inst, int16_t* encoded,
|
static int DecodeNative(OpusDecoder* inst, const int16_t* encoded,
|
||||||
int16_t encoded_bytes, int16_t* decoded,
|
int16_t encoded_bytes, int16_t* decoded,
|
||||||
int16_t* audio_type) {
|
int16_t* audio_type) {
|
||||||
unsigned char* coded = (unsigned char*) encoded;
|
unsigned char* coded = (unsigned char*) encoded;
|
||||||
@ -173,16 +202,113 @@ static int DecodeNative(OpusDecoder* inst, int16_t* encoded,
|
|||||||
return -1;
|
return -1;
|
||||||
}
|
}
|
||||||
|
|
||||||
int16_t WebRtcOpus_Decode(OpusDecInst* inst, int16_t* encoded,
|
/* Resample from 48 to 32 kHz. Length of state is assumed to be
|
||||||
|
* kWebRtcOpusStateSize (7).
|
||||||
|
*/
|
||||||
|
static int WebRtcOpus_Resample48to32(const int16_t* samples_in, int length,
|
||||||
|
int16_t* state, int16_t* samples_out) {
|
||||||
|
int i;
|
||||||
|
int blocks;
|
||||||
|
int16_t output_samples;
|
||||||
|
int32_t buffer32[kWebRtcOpusMaxFrameSize + kWebRtcOpusStateSize];
|
||||||
|
|
||||||
|
/* Resample from 48 kHz to 32 kHz. */
|
||||||
|
for (i = 0; i < kWebRtcOpusStateSize; i++) {
|
||||||
|
buffer32[i] = state[i];
|
||||||
|
state[i] = samples_in[length - kWebRtcOpusStateSize + i];
|
||||||
|
}
|
||||||
|
for (i = 0; i < length; i++) {
|
||||||
|
buffer32[kWebRtcOpusStateSize + i] = samples_in[i];
|
||||||
|
}
|
||||||
|
/* Resampling 3 samples to 2. Function divides the input in |blocks| number
|
||||||
|
* of 3-sample groups, and output is |blocks| number of 2-sample groups.
|
||||||
|
* When this is removed, the compensation in WebRtcOpus_DurationEst should be
|
||||||
|
* removed too. */
|
||||||
|
blocks = length / 3;
|
||||||
|
WebRtcSpl_Resample48khzTo32khz(buffer32, buffer32, blocks);
|
||||||
|
output_samples = (int16_t) (blocks * 2);
|
||||||
|
WebRtcSpl_VectorBitShiftW32ToW16(samples_out, output_samples, buffer32, 15);
|
||||||
|
|
||||||
|
return output_samples;
|
||||||
|
}
|
||||||
|
|
||||||
|
int16_t WebRtcOpus_DecodeNew(OpusDecInst* inst, const uint8_t* encoded,
|
||||||
|
int16_t encoded_bytes, int16_t* decoded,
|
||||||
|
int16_t* audio_type) {
|
||||||
|
/* Enough for 120 ms (the largest Opus packet size) of mono audio at 48 kHz
|
||||||
|
* and resampler overlap. This will need to be enlarged for stereo decoding.
|
||||||
|
*/
|
||||||
|
int16_t buffer16_left[kWebRtcOpusMaxFrameSize];
|
||||||
|
int16_t buffer16_right[kWebRtcOpusMaxFrameSize];
|
||||||
|
int16_t buffer_out[kWebRtcOpusMaxFrameSize];
|
||||||
|
int16_t* coded = (int16_t*) encoded;
|
||||||
|
int decoded_samples;
|
||||||
|
int resampled_samples;
|
||||||
|
int i;
|
||||||
|
|
||||||
|
/* If mono case, just do a regular call to the decoder.
|
||||||
|
* If stereo, we need to de-interleave the stereo output in to blocks with
|
||||||
|
* left and right channel. Each block is resampled to 32 kHz, and then
|
||||||
|
* interleaved again. */
|
||||||
|
|
||||||
|
/* Decode to a temporary buffer. */
|
||||||
|
decoded_samples = DecodeNative(inst->decoder_left, coded, encoded_bytes,
|
||||||
|
buffer16_left, audio_type);
|
||||||
|
if (decoded_samples < 0) {
|
||||||
|
return -1;
|
||||||
|
}
|
||||||
|
|
||||||
|
/* De-interleave if stereo. */
|
||||||
|
if (inst->channels == 2) {
|
||||||
|
/* The parameter |decoded_samples| holds the number of samples pairs, in
|
||||||
|
* case of stereo. Number of samples in |buffer16| equals |decoded_samples|
|
||||||
|
* times 2. */
|
||||||
|
for (i = 0; i < decoded_samples; i++) {
|
||||||
|
/* Take every second sample, starting at the first sample. */
|
||||||
|
buffer16_left[i] = buffer16_left[i * 2];
|
||||||
|
buffer16_right[i] = buffer16_left[i * 2 + 1];
|
||||||
|
}
|
||||||
|
|
||||||
|
/* Resample from 48 kHz to 32 kHz for left channel. */
|
||||||
|
resampled_samples = WebRtcOpus_Resample48to32(buffer16_left,
|
||||||
|
decoded_samples,
|
||||||
|
inst->state_48_32_left,
|
||||||
|
buffer_out);
|
||||||
|
|
||||||
|
/* Add samples interleaved to output vector. */
|
||||||
|
for (i = 0; i < resampled_samples; i++) {
|
||||||
|
decoded[i * 2] = buffer_out[i];
|
||||||
|
}
|
||||||
|
|
||||||
|
/* Resample from 48 kHz to 32 kHz for right channel. */
|
||||||
|
resampled_samples = WebRtcOpus_Resample48to32(buffer16_right,
|
||||||
|
decoded_samples,
|
||||||
|
inst->state_48_32_right,
|
||||||
|
buffer_out);
|
||||||
|
|
||||||
|
/* Add samples interleaved to output vector. */
|
||||||
|
for (i = 0; i < decoded_samples; i++) {
|
||||||
|
decoded[i * 2 + 1] = buffer_out[i];
|
||||||
|
}
|
||||||
|
} else {
|
||||||
|
/* Resample from 48 kHz to 32 kHz for left channel. */
|
||||||
|
resampled_samples = WebRtcOpus_Resample48to32(buffer16_left,
|
||||||
|
decoded_samples,
|
||||||
|
inst->state_48_32_left,
|
||||||
|
decoded);
|
||||||
|
}
|
||||||
|
return resampled_samples;
|
||||||
|
}
|
||||||
|
|
||||||
|
|
||||||
|
int16_t WebRtcOpus_Decode(OpusDecInst* inst, const int16_t* encoded,
|
||||||
int16_t encoded_bytes, int16_t* decoded,
|
int16_t encoded_bytes, int16_t* decoded,
|
||||||
int16_t* audio_type) {
|
int16_t* audio_type) {
|
||||||
/* Enough for 120 ms (the largest Opus packet size) of mono audio at 48 kHz
|
/* Enough for 120 ms (the largest Opus packet size) of mono audio at 48 kHz
|
||||||
* and resampler overlap. This will need to be enlarged for stereo decoding.
|
* and resampler overlap. This will need to be enlarged for stereo decoding.
|
||||||
*/
|
*/
|
||||||
int16_t buffer16[kWebRtcOpusMaxFrameSize];
|
int16_t buffer16[kWebRtcOpusMaxFrameSize];
|
||||||
int32_t buffer32[kWebRtcOpusMaxFrameSize + 7];
|
|
||||||
int decoded_samples;
|
int decoded_samples;
|
||||||
int blocks;
|
|
||||||
int16_t output_samples;
|
int16_t output_samples;
|
||||||
int i;
|
int i;
|
||||||
|
|
||||||
@ -208,36 +334,22 @@ int16_t WebRtcOpus_Decode(OpusDecInst* inst, int16_t* encoded,
|
|||||||
buffer16[i] = buffer16[i * 2];
|
buffer16[i] = buffer16[i * 2];
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
/* Resample from 48 kHz to 32 kHz. */
|
/* Resample from 48 kHz to 32 kHz. */
|
||||||
for (i = 0; i < 7; i++) {
|
output_samples = WebRtcOpus_Resample48to32(buffer16, decoded_samples,
|
||||||
buffer32[i] = inst->state_48_32_left[i];
|
inst->state_48_32_left, decoded);
|
||||||
inst->state_48_32_left[i] = buffer16[decoded_samples - 7 + i];
|
|
||||||
}
|
|
||||||
for (i = 0; i < decoded_samples; i++) {
|
|
||||||
buffer32[7 + i] = buffer16[i];
|
|
||||||
}
|
|
||||||
/* Resampling 3 samples to 2. Function divides the input in |blocks| number
|
|
||||||
* of 3-sample groups, and output is |blocks| number of 2-sample groups.
|
|
||||||
* When this is removed, the compensation in WebRtcOpus_DurationEst should be
|
|
||||||
* removed too. */
|
|
||||||
blocks = decoded_samples / 3;
|
|
||||||
WebRtcSpl_Resample48khzTo32khz(buffer32, buffer32, blocks);
|
|
||||||
output_samples = (int16_t) (blocks * 2);
|
|
||||||
WebRtcSpl_VectorBitShiftW32ToW16(decoded, output_samples, buffer32, 15);
|
|
||||||
|
|
||||||
return output_samples;
|
return output_samples;
|
||||||
}
|
}
|
||||||
|
|
||||||
int16_t WebRtcOpus_DecodeSlave(OpusDecInst* inst, int16_t* encoded,
|
int16_t WebRtcOpus_DecodeSlave(OpusDecInst* inst, const int16_t* encoded,
|
||||||
int16_t encoded_bytes, int16_t* decoded,
|
int16_t encoded_bytes, int16_t* decoded,
|
||||||
int16_t* audio_type) {
|
int16_t* audio_type) {
|
||||||
/* Enough for 120 ms (the largest Opus packet size) of mono audio at 48 kHz
|
/* Enough for 120 ms (the largest Opus packet size) of mono audio at 48 kHz
|
||||||
* and resampler overlap. This will need to be enlarged for stereo decoding.
|
* and resampler overlap. This will need to be enlarged for stereo decoding.
|
||||||
*/
|
*/
|
||||||
int16_t buffer16[kWebRtcOpusMaxFrameSize];
|
int16_t buffer16[kWebRtcOpusMaxFrameSize];
|
||||||
int32_t buffer32[kWebRtcOpusMaxFrameSize + 7];
|
|
||||||
int decoded_samples;
|
int decoded_samples;
|
||||||
int blocks;
|
|
||||||
int16_t output_samples;
|
int16_t output_samples;
|
||||||
int i;
|
int i;
|
||||||
|
|
||||||
@ -261,19 +373,8 @@ int16_t WebRtcOpus_DecodeSlave(OpusDecInst* inst, int16_t* encoded,
|
|||||||
return -1;
|
return -1;
|
||||||
}
|
}
|
||||||
/* Resample from 48 kHz to 32 kHz. */
|
/* Resample from 48 kHz to 32 kHz. */
|
||||||
for (i = 0; i < 7; i++) {
|
output_samples = WebRtcOpus_Resample48to32(buffer16, decoded_samples,
|
||||||
buffer32[i] = inst->state_48_32_right[i];
|
inst->state_48_32_right, decoded);
|
||||||
inst->state_48_32_right[i] = buffer16[decoded_samples - 7 + i];
|
|
||||||
}
|
|
||||||
for (i = 0; i < decoded_samples; i++) {
|
|
||||||
buffer32[7 + i] = buffer16[i];
|
|
||||||
}
|
|
||||||
/* Resampling 3 samples to 2. Function divides the input in |blocks| number
|
|
||||||
* of 3-sample groups, and output is |blocks| number of 2-sample groups. */
|
|
||||||
blocks = decoded_samples / 3;
|
|
||||||
WebRtcSpl_Resample48khzTo32khz(buffer32, buffer32, blocks);
|
|
||||||
output_samples = (int16_t) (blocks * 2);
|
|
||||||
WebRtcSpl_VectorBitShiftW32ToW16(decoded, output_samples, buffer32, 15);
|
|
||||||
|
|
||||||
return output_samples;
|
return output_samples;
|
||||||
}
|
}
|
||||||
|
301
webrtc/modules/audio_coding/codecs/opus/opus_unittest.cc
Normal file
301
webrtc/modules/audio_coding/codecs/opus/opus_unittest.cc
Normal file
@ -0,0 +1,301 @@
|
|||||||
|
/*
|
||||||
|
* Copyright (c) 2013 The WebRTC project authors. All Rights Reserved.
|
||||||
|
*
|
||||||
|
* Use of this source code is governed by a BSD-style license
|
||||||
|
* that can be found in the LICENSE file in the root of the source
|
||||||
|
* tree. An additional intellectual property rights grant can be found
|
||||||
|
* in the file PATENTS. All contributing project authors may
|
||||||
|
* be found in the AUTHORS file in the root of the source tree.
|
||||||
|
*/
|
||||||
|
#include <string>
|
||||||
|
|
||||||
|
#include "testing/gtest/include/gtest/gtest.h"
|
||||||
|
#include "webrtc/modules/audio_coding/codecs/opus/interface/opus_interface.h"
|
||||||
|
#include "webrtc/test/testsupport/fileutils.h"
|
||||||
|
|
||||||
|
struct WebRtcOpusEncInst;
|
||||||
|
struct WebRtcOpusDecInst;
|
||||||
|
|
||||||
|
namespace webrtc {
|
||||||
|
|
||||||
|
// Number of samples in a 60 ms stereo frame, sampled at 48 kHz.
|
||||||
|
enum { kOpusNumberOfSamples = 480 * 6 * 2 };
|
||||||
|
// Maximum number of bytes in output bitstream.
|
||||||
|
enum { kMaxBytes = 1000 };
|
||||||
|
|
||||||
|
class OpusTest : public ::testing::Test {
|
||||||
|
protected:
|
||||||
|
OpusTest();
|
||||||
|
virtual void SetUp();
|
||||||
|
|
||||||
|
WebRtcOpusEncInst* opus_mono_encoder_;
|
||||||
|
WebRtcOpusEncInst* opus_stereo_encoder_;
|
||||||
|
WebRtcOpusDecInst* opus_mono_decoder_;
|
||||||
|
WebRtcOpusDecInst* opus_mono_decoder_new_;
|
||||||
|
WebRtcOpusDecInst* opus_stereo_decoder_;
|
||||||
|
WebRtcOpusDecInst* opus_stereo_decoder_new_;
|
||||||
|
|
||||||
|
int16_t speech_data_[kOpusNumberOfSamples];
|
||||||
|
int16_t output_data_[kOpusNumberOfSamples];
|
||||||
|
uint8_t bitstream_[kMaxBytes];
|
||||||
|
};
|
||||||
|
|
||||||
|
OpusTest::OpusTest()
|
||||||
|
: opus_mono_encoder_(NULL),
|
||||||
|
opus_stereo_encoder_(NULL),
|
||||||
|
opus_mono_decoder_(NULL),
|
||||||
|
opus_mono_decoder_new_(NULL),
|
||||||
|
opus_stereo_decoder_(NULL),
|
||||||
|
opus_stereo_decoder_new_(NULL) {
|
||||||
|
}
|
||||||
|
|
||||||
|
void OpusTest::SetUp() {
|
||||||
|
// Read some samples from a speech file, to be used in the encode test.
|
||||||
|
// In this test we do not care that the sampling frequency of the file is
|
||||||
|
// really 32000 Hz. We pretend that it is 48000 Hz.
|
||||||
|
FILE* input_file;
|
||||||
|
const std::string file_name =
|
||||||
|
webrtc::test::ResourcePath("audio_coding/testfile32kHz", "pcm");
|
||||||
|
input_file = fopen(file_name.c_str(), "rb");
|
||||||
|
ASSERT_TRUE(input_file != NULL);
|
||||||
|
ASSERT_EQ(kOpusNumberOfSamples,
|
||||||
|
static_cast<int32_t>(fread(speech_data_, sizeof(int16_t),
|
||||||
|
kOpusNumberOfSamples, input_file)));
|
||||||
|
fclose(input_file);
|
||||||
|
input_file = NULL;
|
||||||
|
}
|
||||||
|
|
||||||
|
// Test failing Create.
|
||||||
|
TEST_F(OpusTest, OpusCreateFail) {
|
||||||
|
// Test to see that an invalid pointer is caught.
|
||||||
|
EXPECT_EQ(-1, WebRtcOpus_EncoderCreate(NULL, 1));
|
||||||
|
EXPECT_EQ(-1, WebRtcOpus_EncoderCreate(&opus_mono_encoder_, 3));
|
||||||
|
EXPECT_EQ(-1, WebRtcOpus_DecoderCreate(NULL, 1));
|
||||||
|
EXPECT_EQ(-1, WebRtcOpus_DecoderCreate(&opus_mono_decoder_, 3));
|
||||||
|
}
|
||||||
|
|
||||||
|
// Test failing Free.
|
||||||
|
TEST_F(OpusTest, OpusFreeFail) {
|
||||||
|
// Test to see that an invalid pointer is caught.
|
||||||
|
EXPECT_EQ(-1, WebRtcOpus_EncoderFree(NULL));
|
||||||
|
EXPECT_EQ(-1, WebRtcOpus_DecoderFree(NULL));
|
||||||
|
}
|
||||||
|
|
||||||
|
// Test normal Create and Free.
|
||||||
|
TEST_F(OpusTest, OpusCreateFree) {
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderCreate(&opus_mono_encoder_, 1));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderCreate(&opus_mono_decoder_, 1));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderCreate(&opus_stereo_encoder_, 2));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderCreate(&opus_stereo_decoder_, 2));
|
||||||
|
EXPECT_TRUE(opus_mono_encoder_ != NULL);
|
||||||
|
EXPECT_TRUE(opus_mono_decoder_ != NULL);
|
||||||
|
EXPECT_TRUE(opus_stereo_encoder_ != NULL);
|
||||||
|
EXPECT_TRUE(opus_stereo_decoder_ != NULL);
|
||||||
|
// Free encoder and decoder memory.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderFree(opus_mono_encoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderFree(opus_mono_decoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderFree(opus_stereo_encoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderFree(opus_stereo_decoder_));
|
||||||
|
}
|
||||||
|
|
||||||
|
TEST_F(OpusTest, OpusEncodeDecodeMono) {
|
||||||
|
// Create encoder memory.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderCreate(&opus_mono_encoder_, 1));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderCreate(&opus_mono_decoder_, 1));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderCreate(&opus_mono_decoder_new_, 1));
|
||||||
|
|
||||||
|
// Set bitrate.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_SetBitRate(opus_mono_encoder_, 32000));
|
||||||
|
|
||||||
|
// Check number of channels for decoder.
|
||||||
|
EXPECT_EQ(1, WebRtcOpus_DecoderChannels(opus_mono_decoder_));
|
||||||
|
EXPECT_EQ(1, WebRtcOpus_DecoderChannels(opus_mono_decoder_new_));
|
||||||
|
|
||||||
|
// Encode & decode.
|
||||||
|
int16_t encoded_bytes;
|
||||||
|
int16_t audio_type;
|
||||||
|
int16_t output_data_decode_new[kOpusNumberOfSamples];
|
||||||
|
int16_t output_data_decode[kOpusNumberOfSamples];
|
||||||
|
int16_t* coded = reinterpret_cast<int16_t*>(bitstream_);
|
||||||
|
encoded_bytes = WebRtcOpus_Encode(opus_mono_encoder_, speech_data_, 960,
|
||||||
|
kMaxBytes, bitstream_);
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_DecodeNew(opus_mono_decoder_new_, bitstream_,
|
||||||
|
encoded_bytes, output_data_decode_new,
|
||||||
|
&audio_type));
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_Decode(opus_mono_decoder_, coded,
|
||||||
|
encoded_bytes, output_data_decode,
|
||||||
|
&audio_type));
|
||||||
|
|
||||||
|
// Data in |output_data_decode_new| should be the same as in
|
||||||
|
// |output_data_decode|.
|
||||||
|
for (int i = 0; i < 640; i++) {
|
||||||
|
EXPECT_EQ(output_data_decode_new[i], output_data_decode[i]);
|
||||||
|
}
|
||||||
|
|
||||||
|
// Free memory.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderFree(opus_mono_encoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderFree(opus_mono_decoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderFree(opus_mono_decoder_new_));
|
||||||
|
}
|
||||||
|
|
||||||
|
TEST_F(OpusTest, OpusEncodeDecodeStereo) {
|
||||||
|
// Create encoder memory.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderCreate(&opus_stereo_encoder_, 2));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderCreate(&opus_stereo_decoder_, 2));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderCreate(&opus_stereo_decoder_new_, 2));
|
||||||
|
|
||||||
|
// Set bitrate.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_SetBitRate(opus_stereo_encoder_, 64000));
|
||||||
|
|
||||||
|
// Check number of channels for decoder.
|
||||||
|
EXPECT_EQ(2, WebRtcOpus_DecoderChannels(opus_stereo_decoder_));
|
||||||
|
EXPECT_EQ(2, WebRtcOpus_DecoderChannels(opus_stereo_decoder_new_));
|
||||||
|
|
||||||
|
// Encode & decode.
|
||||||
|
int16_t encoded_bytes;
|
||||||
|
int16_t audio_type;
|
||||||
|
int16_t output_data_decode_new[kOpusNumberOfSamples];
|
||||||
|
int16_t output_data_decode[kOpusNumberOfSamples];
|
||||||
|
int16_t output_data_decode_slave[kOpusNumberOfSamples];
|
||||||
|
int16_t* coded = reinterpret_cast<int16_t*>(bitstream_);
|
||||||
|
encoded_bytes = WebRtcOpus_Encode(opus_stereo_encoder_, speech_data_, 960,
|
||||||
|
kMaxBytes, bitstream_);
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_DecodeNew(opus_stereo_decoder_new_, bitstream_,
|
||||||
|
encoded_bytes, output_data_decode_new,
|
||||||
|
&audio_type));
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_Decode(opus_stereo_decoder_, coded,
|
||||||
|
encoded_bytes, output_data_decode,
|
||||||
|
&audio_type));
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_DecodeSlave(opus_stereo_decoder_, coded,
|
||||||
|
encoded_bytes, output_data_decode_slave,
|
||||||
|
&audio_type));
|
||||||
|
|
||||||
|
// Data in |output_data_decode_new| should be the same as in
|
||||||
|
// |output_data_decode| and |output_data_decode_slave| interleaved to a
|
||||||
|
// stereo signal.
|
||||||
|
for (int i = 0; i < 640; i++) {
|
||||||
|
EXPECT_EQ(output_data_decode_new[i * 2], output_data_decode[i]);
|
||||||
|
EXPECT_EQ(output_data_decode_new[i * 2 + 1], output_data_decode_slave[i]);
|
||||||
|
}
|
||||||
|
|
||||||
|
// Free memory.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderFree(opus_stereo_encoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderFree(opus_stereo_decoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderFree(opus_stereo_decoder_new_));
|
||||||
|
}
|
||||||
|
|
||||||
|
TEST_F(OpusTest, OpusSetBitRate) {
|
||||||
|
// Test without creating encoder memory.
|
||||||
|
EXPECT_EQ(-1, WebRtcOpus_SetBitRate(opus_mono_encoder_, 60000));
|
||||||
|
EXPECT_EQ(-1, WebRtcOpus_SetBitRate(opus_stereo_encoder_, 60000));
|
||||||
|
|
||||||
|
// Create encoder memory, try with different bitrates.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderCreate(&opus_mono_encoder_, 1));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderCreate(&opus_stereo_encoder_, 2));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_SetBitRate(opus_mono_encoder_, 30000));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_SetBitRate(opus_stereo_encoder_, 60000));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_SetBitRate(opus_mono_encoder_, 300000));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_SetBitRate(opus_stereo_encoder_, 600000));
|
||||||
|
|
||||||
|
// Free memory.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderFree(opus_mono_encoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderFree(opus_stereo_encoder_));
|
||||||
|
}
|
||||||
|
|
||||||
|
// Encode and decode one frame (stereo), initialize the decoder and
|
||||||
|
// decode once more.
|
||||||
|
TEST_F(OpusTest, OpusDecodeInit) {
|
||||||
|
// Create encoder memory.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderCreate(&opus_stereo_encoder_, 2));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderCreate(&opus_stereo_decoder_, 2));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderCreate(&opus_stereo_decoder_new_, 2));
|
||||||
|
|
||||||
|
// Encode & decode.
|
||||||
|
int16_t encoded_bytes;
|
||||||
|
int16_t audio_type;
|
||||||
|
int16_t output_data_decode_new[kOpusNumberOfSamples];
|
||||||
|
int16_t output_data_decode[kOpusNumberOfSamples];
|
||||||
|
int16_t output_data_decode_slave[kOpusNumberOfSamples];
|
||||||
|
int16_t* coded = reinterpret_cast<int16_t*>(bitstream_);
|
||||||
|
encoded_bytes = WebRtcOpus_Encode(opus_stereo_encoder_, speech_data_, 960,
|
||||||
|
kMaxBytes, bitstream_);
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_DecodeNew(opus_stereo_decoder_new_, bitstream_,
|
||||||
|
encoded_bytes, output_data_decode_new,
|
||||||
|
&audio_type));
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_Decode(opus_stereo_decoder_, coded,
|
||||||
|
encoded_bytes, output_data_decode,
|
||||||
|
&audio_type));
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_DecodeSlave(opus_stereo_decoder_, coded,
|
||||||
|
encoded_bytes, output_data_decode_slave,
|
||||||
|
&audio_type));
|
||||||
|
|
||||||
|
// Data in |output_data_decode_new| should be the same as in
|
||||||
|
// |output_data_decode| and |output_data_decode_slave| interleaved to a
|
||||||
|
// stereo signal.
|
||||||
|
for (int i = 0; i < 640; i++) {
|
||||||
|
EXPECT_EQ(output_data_decode_new[i * 2], output_data_decode[i]);
|
||||||
|
EXPECT_EQ(output_data_decode_new[i * 2 + 1], output_data_decode_slave[i]);
|
||||||
|
}
|
||||||
|
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderInitNew(opus_stereo_decoder_new_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderInit(opus_stereo_decoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderInitSlave(opus_stereo_decoder_));
|
||||||
|
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_DecodeNew(opus_stereo_decoder_new_, bitstream_,
|
||||||
|
encoded_bytes, output_data_decode_new,
|
||||||
|
&audio_type));
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_Decode(opus_stereo_decoder_, coded,
|
||||||
|
encoded_bytes, output_data_decode,
|
||||||
|
&audio_type));
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_DecodeSlave(opus_stereo_decoder_, coded,
|
||||||
|
encoded_bytes, output_data_decode_slave,
|
||||||
|
&audio_type));
|
||||||
|
|
||||||
|
// Data in |output_data_decode_new| should be the same as in
|
||||||
|
// |output_data_decode| and |output_data_decode_slave| interleaved to a
|
||||||
|
// stereo signal.
|
||||||
|
for (int i = 0; i < 640; i++) {
|
||||||
|
EXPECT_EQ(output_data_decode_new[i * 2], output_data_decode[i]);
|
||||||
|
EXPECT_EQ(output_data_decode_new[i * 2 + 1], output_data_decode_slave[i]);
|
||||||
|
}
|
||||||
|
|
||||||
|
// Free memory.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderFree(opus_stereo_encoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderFree(opus_stereo_decoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderFree(opus_stereo_decoder_new_));
|
||||||
|
}
|
||||||
|
|
||||||
|
// PLC not implemented.
|
||||||
|
TEST_F(OpusTest, OpusDecodePlc) {
|
||||||
|
int16_t plc_buffer[kOpusNumberOfSamples];
|
||||||
|
EXPECT_EQ(-1, WebRtcOpus_DecodePlc(opus_stereo_decoder_, plc_buffer, 1));
|
||||||
|
}
|
||||||
|
|
||||||
|
// Duration estimation.
|
||||||
|
TEST_F(OpusTest, OpusDurationEstimation) {
|
||||||
|
// Create.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderCreate(&opus_stereo_encoder_, 2));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderCreate(&opus_stereo_decoder_, 2));
|
||||||
|
|
||||||
|
// Encode with different packet sizes (input 48 kHz, output in 32 kHz).
|
||||||
|
int16_t encoded_bytes;
|
||||||
|
|
||||||
|
// 10 ms.
|
||||||
|
encoded_bytes = WebRtcOpus_Encode(opus_stereo_encoder_, speech_data_, 480,
|
||||||
|
kMaxBytes, bitstream_);
|
||||||
|
EXPECT_EQ(320, WebRtcOpus_DurationEst(opus_stereo_decoder_, bitstream_,
|
||||||
|
encoded_bytes));
|
||||||
|
|
||||||
|
// 20 ms
|
||||||
|
encoded_bytes = WebRtcOpus_Encode(opus_stereo_encoder_, speech_data_, 960,
|
||||||
|
kMaxBytes, bitstream_);
|
||||||
|
EXPECT_EQ(640, WebRtcOpus_DurationEst(opus_stereo_decoder_, bitstream_,
|
||||||
|
encoded_bytes));
|
||||||
|
|
||||||
|
|
||||||
|
// Free memory.
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_EncoderFree(opus_stereo_encoder_));
|
||||||
|
EXPECT_EQ(0, WebRtcOpus_DecoderFree(opus_stereo_decoder_));
|
||||||
|
}
|
||||||
|
|
||||||
|
} // namespace webrtc
|
@ -190,6 +190,7 @@
|
|||||||
'../../codecs/isac/fix/source/filterbanks_unittest.cc',
|
'../../codecs/isac/fix/source/filterbanks_unittest.cc',
|
||||||
'../../codecs/isac/fix/source/lpc_masking_model_unittest.cc',
|
'../../codecs/isac/fix/source/lpc_masking_model_unittest.cc',
|
||||||
'../../codecs/isac/fix/source/transform_unittest.cc',
|
'../../codecs/isac/fix/source/transform_unittest.cc',
|
||||||
|
'../../codecs/opus/opus_unittest.cc',
|
||||||
# Test for NetEq 4.
|
# Test for NetEq 4.
|
||||||
'../../neteq4/audio_multi_vector_unittest.cc',
|
'../../neteq4/audio_multi_vector_unittest.cc',
|
||||||
'../../neteq4/audio_vector_unittest.cc',
|
'../../neteq4/audio_vector_unittest.cc',
|
||||||
|
Loading…
x
Reference in New Issue
Block a user