Skip to content
This repository has been archived by the owner on Oct 4, 2024. It is now read-only.

Commit

Permalink
audio/stream: support more files by transcoding input audio
Browse files Browse the repository at this point in the history
  • Loading branch information
@rr-
rr- committed Apr 30, 2024
1 parent 79f8a46 commit 388534f
Show file tree
Hide file tree
Showing 3 changed files with 124 additions and 52 deletions.
15 changes: 13 additions & 2 deletions src/engine/audio.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -91,17 +91,28 @@ bool Audio_Shutdown(void)
return true;
}

int32_t Audio_GetAVChannelLayout(const int32_t channels)
{
switch (channels) {
// clang-format off
case 1: return AV_CH_LAYOUT_MONO;
case 2: return AV_CH_LAYOUT_STEREO;
default: return AV_CH_LAYOUT_MONO;
// clang-format on
}
}

int32_t Audio_GetAVAudioFormat(const int32_t sample_fmt)
{
// clang-format off
switch (sample_fmt) {
// clang-format off
case AUDIO_U8: return AV_SAMPLE_FMT_U8;
case AUDIO_S16: return AV_SAMPLE_FMT_S16;
case AUDIO_S32: return AV_SAMPLE_FMT_S32;
case AUDIO_F32: return AV_SAMPLE_FMT_FLT;
default: return -1;
// clang-format on
}
// clang-format on
}

int32_t Audio_GetSDLAudioFormat(const enum AVSampleFormat sample_fmt)
Expand Down
5 changes: 3 additions & 2 deletions src/engine/audio.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -12,8 +12,9 @@

extern SDL_AudioDeviceID g_AudioDeviceID;

int32_t Audio_GetAVAudioFormat(const int32_t sample_fmt);
int32_t Audio_GetSDLAudioFormat(const enum AVSampleFormat sample_fmt);
int32_t Audio_GetAVChannelLayout(int32_t sample_fmt);
int32_t Audio_GetAVAudioFormat(int32_t sample_fmt);
int32_t Audio_GetSDLAudioFormat(enum AVSampleFormat sample_fmt);

void Audio_Sample_Init(void);
void Audio_Sample_Shutdown(void);
Expand Down
156 changes: 108 additions & 48 deletions src/engine/audio_stream.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@
#include <libavcodec/avcodec.h>
#include <libavcodec/codec.h>
#include <libavcodec/packet.h>
#include <libswresample/swresample.h>
#include <libavformat/avformat.h>
#include <libavformat/avio.h>
#include <libavutil/avutil.h>
Expand Down Expand Up @@ -51,6 +52,16 @@ typedef struct AUDIO_STREAM_SOUND {
AVFrame *frame;
} av;

struct {
int32_t src_format;
int32_t src_channels;
int32_t src_sample_rate;
int32_t dst_format;
int32_t dst_channels;
int32_t dst_sample_rate;
SwrContext *ctx;
} swr;

struct {
SDL_AudioStream *stream;
} sdl;
Expand All @@ -59,7 +70,10 @@ typedef struct AUDIO_STREAM_SOUND {
extern SDL_AudioDeviceID g_AudioDeviceID;

static AUDIO_STREAM_SOUND m_Streams[AUDIO_MAX_ACTIVE_STREAMS] = { 0 };
static float m_DecodeBuffer[AUDIO_SAMPLES * AUDIO_WORKING_CHANNELS] = { 0 };
static float m_MixBuffer[AUDIO_SAMPLES * AUDIO_WORKING_CHANNELS] = { 0 };

static size_t m_DecodeBufferCapacity = 0;
static float *m_DecodeBuffer = NULL;

static void Audio_Stream_SeekToStart(AUDIO_STREAM_SOUND *stream);
static bool Audio_Stream_DecodeFrame(AUDIO_STREAM_SOUND *stream);
Expand Down Expand Up @@ -132,44 +146,81 @@ static bool Audio_Stream_EnqueueFrame(AUDIO_STREAM_SOUND *stream)
{
assert(stream != NULL);

while (1) {
int32_t error_code =
avcodec_receive_frame(stream->av.codec_ctx, stream->av.frame);
if (error_code == AVERROR(EAGAIN)) {
av_frame_unref(stream->av.frame);
break;
}
int32_t error_code;

if (error_code < 0) {
if (!stream->swr.ctx) {
stream->swr.src_sample_rate = stream->av.codec_ctx->sample_rate;
stream->swr.src_channels = stream->av.codec_ctx->channels;
stream->swr.src_format = stream->av.codec_ctx->sample_fmt;
stream->swr.dst_sample_rate = AUDIO_WORKING_RATE;
stream->swr.dst_channels = AUDIO_WORKING_CHANNELS;
stream->swr.dst_format = Audio_GetAVAudioFormat(AUDIO_WORKING_FORMAT);
stream->swr.ctx = swr_alloc_set_opts(
stream->swr.ctx, Audio_GetAVChannelLayout(stream->swr.dst_channels),
stream->swr.dst_format, stream->swr.dst_sample_rate,
stream->swr.src_channels, stream->swr.src_format,
stream->swr.src_sample_rate, 0, 0);
if (!stream->swr.ctx) {
av_packet_unref(stream->av.packet);
av_frame_unref(stream->av.frame);
LOG_ERROR(
"Got an error when decoding frame: %d, %s", error_code,
av_err2str(error_code));
break;
error_code = AVERROR(ENOMEM);
goto cleanup;
}

error_code = av_samples_get_buffer_size(
NULL, stream->av.codec_ctx->channels, stream->av.frame->nb_samples,
stream->av.codec_ctx->sample_fmt, 1);
error_code = swr_init(stream->swr.ctx);
if (error_code != 0) {
av_packet_unref(stream->av.packet);
goto cleanup;
}
}

while (1) {
error_code =
avcodec_receive_frame(stream->av.codec_ctx, stream->av.frame);
if (error_code == AVERROR(EAGAIN)) {
av_frame_unref(stream->av.frame);
error_code = 0;
break;
}

if (error_code < 0) {
LOG_ERROR(
"Got an error when decoding frame: %d, %s", error_code,
av_err2str(error_code));
av_frame_unref(stream->av.frame);
break;
}

int32_t data_size = error_code;
uint8_t *out_buffer = NULL;
const int32_t out_samples =
swr_get_out_samples(stream->swr.ctx, stream->av.frame->nb_samples);
av_samples_alloc(
&out_buffer, NULL, stream->swr.dst_channels, out_samples,
stream->swr.dst_format, 1);
int32_t resampled_size = swr_convert(
stream->swr.ctx, &out_buffer, out_samples,
(const uint8_t **)stream->av.frame->data,
stream->av.frame->nb_samples);

size_t out_pos = 0;
while (resampled_size > 0) {
const size_t out_buffer_size = av_samples_get_buffer_size(
NULL, stream->swr.dst_channels, resampled_size,
stream->swr.dst_format, 1);

if (out_buffer_size > m_DecodeBufferCapacity) {
m_DecodeBufferCapacity += out_buffer_size;
m_DecodeBuffer =
Memory_Realloc(m_DecodeBuffer, m_DecodeBufferCapacity);
}
if (out_buffer) {
memcpy(
(uint8_t *)m_DecodeBuffer + out_pos, out_buffer,
out_buffer_size);
}
out_pos += out_buffer_size;

resampled_size =
swr_convert(stream->swr.ctx, &out_buffer, out_samples, NULL, 0);
}

if (SDL_AudioStreamPut(
stream->sdl.stream, stream->av.frame->data[0], data_size)) {
if (SDL_AudioStreamPut(stream->sdl.stream, m_DecodeBuffer, out_pos)) {
LOG_ERROR("Got an error when decoding frame: %s", SDL_GetError());
av_frame_unref(stream->av.frame);
break;
Expand All @@ -178,10 +229,19 @@ static bool Audio_Stream_EnqueueFrame(AUDIO_STREAM_SOUND *stream)
double time_base_sec = av_q2d(stream->av.stream->time_base);
stream->timestamp =
stream->av.frame->best_effort_timestamp * time_base_sec;
av_freep(&out_buffer);
av_frame_unref(stream->av.frame);
}

av_packet_unref(stream->av.packet);

cleanup:
if (error_code > 0) {
LOG_ERROR(
"Got an error when decoding frame: %d, %s", error_code,
av_err2str(error_code));
}

return true;
}

Expand Down Expand Up @@ -265,14 +325,6 @@ static bool Audio_Stream_InitialiseFromPath(

Audio_Stream_DecodeFrame(stream);

int32_t sdl_format =
Audio_GetSDLAudioFormat(stream->av.codec_ctx->sample_fmt);
if (sdl_format < 0) {
LOG_ERROR(
"Unknown sample format: %d", stream->av.codec_ctx->sample_fmt);
goto cleanup;
}

int32_t sdl_sample_rate = stream->av.codec_ctx->sample_rate;
int32_t sdl_channels = stream->av.codec_ctx->channels;

Expand All @@ -290,8 +342,8 @@ static bool Audio_Stream_InitialiseFromPath(
stream->stop_at = -1.0; // negative value means unset

stream->sdl.stream = SDL_NewAudioStream(
sdl_format, sdl_channels, sdl_sample_rate, AUDIO_WORKING_FORMAT,
sdl_channels, AUDIO_WORKING_RATE);
AUDIO_WORKING_FORMAT, sdl_channels, sdl_sample_rate,
AUDIO_WORKING_FORMAT, sdl_channels, AUDIO_WORKING_RATE);
if (!stream->sdl.stream) {
LOG_ERROR("Failed to create SDL stream: %s", SDL_GetError());
goto cleanup;
Expand Down Expand Up @@ -342,6 +394,7 @@ void Audio_Stream_Init(void)

void Audio_Stream_Shutdown(void)
{
Memory_FreePointer(&m_DecodeBuffer);
if (!g_AudioDeviceID) {
return;
}
Expand Down Expand Up @@ -429,16 +482,20 @@ bool Audio_Stream_Close(int32_t sound_id)
stream->av.format_ctx = NULL;
}

if (stream->av.packet) {
av_packet_free(&stream->av.packet);
stream->av.packet = NULL;
if (stream->swr.ctx) {
swr_free(&stream->swr.ctx);
}

if (stream->av.frame) {
av_frame_free(&stream->av.frame);
stream->av.frame = NULL;
}

if (stream->av.packet) {
av_packet_free(&stream->av.packet);
stream->av.packet = NULL;
}

stream->av.stream = NULL;
stream->av.codec = NULL;

Expand Down Expand Up @@ -527,9 +584,9 @@ void Audio_Stream_Mix(float *dst_buffer, size_t len)
}
}

memset(m_DecodeBuffer, 0, READ_BUFFER_SIZE);
memset(m_MixBuffer, 0, READ_BUFFER_SIZE);
int32_t bytes_gotten = SDL_AudioStreamGet(
stream->sdl.stream, m_DecodeBuffer, READ_BUFFER_SIZE);
stream->sdl.stream, m_MixBuffer, READ_BUFFER_SIZE);
if (bytes_gotten < 0) {
LOG_ERROR("Error reading from sdl.stream: %s", SDL_GetError());
stream->is_playing = false;
Expand All @@ -543,21 +600,23 @@ void Audio_Stream_Mix(float *dst_buffer, size_t len)
stream->is_read_done = true;
} else {
int32_t samples_gotten = bytes_gotten
/ (stream->av.codec_ctx->channels
* sizeof(AUDIO_WORKING_FORMAT));
/ (AUDIO_WORKING_CHANNELS * sizeof(AUDIO_WORKING_FORMAT));

const float *src_ptr = &m_DecodeBuffer[0];
const float *src_ptr = &m_MixBuffer[0];
float *dst_ptr = dst_buffer;

if (stream->av.codec_ctx->channels == 2) {
if (stream->av.codec_ctx->channels == AUDIO_WORKING_CHANNELS) {
for (int32_t s = 0; s < samples_gotten; s++) {
*dst_ptr++ += *src_ptr++ * stream->volume;
*dst_ptr++ += *src_ptr++ * stream->volume;
for (int32_t c = 0; c < AUDIO_WORKING_CHANNELS; c++) {
*dst_ptr++ += *src_ptr++ * stream->volume;
}
}
} else if (stream->av.codec_ctx->channels == 1) {
for (int32_t s = 0; s < samples_gotten; s++) {
*dst_ptr++ += *src_ptr * stream->volume;
*dst_ptr++ += *src_ptr++ * stream->volume;
for (int32_t c = 0; c < AUDIO_WORKING_CHANNELS; c++) {
*dst_ptr++ += *src_ptr * stream->volume;
}
src_ptr++;
}
} else {
for (int32_t s = 0; s < samples_gotten; s++) {
Expand All @@ -568,8 +627,9 @@ void Audio_Stream_Mix(float *dst_buffer, size_t len)
src_sample += *src_ptr++;
}
src_sample /= (float)stream->av.codec_ctx->channels;
*dst_ptr++ += src_sample * stream->volume;
*dst_ptr++ += src_sample * stream->volume;
for (int32_t c = 0; c < AUDIO_WORKING_CHANNELS; c++) {
*dst_ptr++ += src_sample * stream->volume;
}
}
}
}
Expand Down

0 comments on commit 388534f

Please sign in to comment.