path: root/src/internal/audio/kitaudio.c

#include <assert.h>
#define __STDC_FORMAT_MACROS
#include <inttypes.h>

#include <libavformat/avformat.h>
#include <libavutil/samplefmt.h>
#include <libswresample/swresample.h>
#include <SDL.h>

#include "kitchensink/kiterror.h"
#include "kitchensink/internal/kitlibstate.h"
#include "kitchensink/internal/utils/kithelpers.h"
#include "kitchensink/internal/audio/kitaudio.h"
#include "kitchensink/internal/utils/kitringbuffer.h"

#if LIBAVUTIL_VERSION_MAJOR < 58
#define OLD_CHANNEL_LAYOUT
#endif

#define KIT_AUDIO_SYNC_THRESHOLD 0.05

typedef struct Kit_AudioDecoder {
    SwrContext *swr;
    AVFrame *scratch_frame;
} Kit_AudioDecoder;

typedef struct Kit_AudioPacket {
    double pts;
    size_t original_size;
    Kit_RingBuffer *rb;
} Kit_AudioPacket;


static Kit_AudioPacket* _CreateAudioPacket(const char* data, size_t len, double pts) {
    Kit_AudioPacket *p = calloc(1, sizeof(Kit_AudioPacket));
    p->rb = Kit_CreateRingBuffer(len);
    Kit_WriteRingBuffer(p->rb, data, len);
    p->pts = pts;
    return p;
}

static enum AVSampleFormat _FindAVSampleFormat(int format) {
    switch(format) {
        case AUDIO_U8: return AV_SAMPLE_FMT_U8;
        case AUDIO_S16SYS: return AV_SAMPLE_FMT_S16;
        case AUDIO_S32SYS: return AV_SAMPLE_FMT_S32;
        default: return AV_SAMPLE_FMT_NONE;
    }
}

#ifdef OLD_CHANNEL_LAYOUT
static int64_t _FindAVChannelLayout(int channels) {
    switch(channels) {
        case 1: return AV_CH_LAYOUT_MONO;
        case 2: return AV_CH_LAYOUT_STEREO;
        default: return AV_CH_LAYOUT_STEREO_DOWNMIX;
    }
}

static int _FindChannelLayout(uint64_t channel_layout) {
    switch(channel_layout) {
        case AV_CH_LAYOUT_MONO: return 1;
        case AV_CH_LAYOUT_STEREO: return 2;
        default: return 2;
    }
}
#else
static void _FindAVChannelLayout(AVChannelLayout *layout, int channels) {
    switch(channels) {
        case 1: *layout = (AVChannelLayout)AV_CHANNEL_LAYOUT_MONO;
        case 2: *layout = (AVChannelLayout)AV_CHANNEL_LAYOUT_STEREO;
        default: *layout = (AVChannelLayout)AV_CHANNEL_LAYOUT_STEREO;
    }
}

static int _FindChannelLayout(const AVChannelLayout *channel_layout) {
    if (channel_layout->nb_channels > 2)
        return 2;
    return channel_layout->nb_channels;
}
#endif

static int _FindBytes(enum AVSampleFormat fmt) {
    switch(fmt) {
        case AV_SAMPLE_FMT_U8P:
        case AV_SAMPLE_FMT_U8:
            return 1;
        case AV_SAMPLE_FMT_S32P:
        case AV_SAMPLE_FMT_S32:
            return 4;
        default:
            return 2;
    }
}

static int _FindSDLSampleFormat(enum AVSampleFormat fmt) {
    switch(fmt) {
        case AV_SAMPLE_FMT_U8P:
        case AV_SAMPLE_FMT_U8:
            return AUDIO_U8;
        case AV_SAMPLE_FMT_S32P:
        case AV_SAMPLE_FMT_S32:
            return AUDIO_S32SYS;
        default:
            return AUDIO_S16SYS;
    }
}

static int _FindSignedness(enum AVSampleFormat fmt) {
    switch(fmt) {
        case AV_SAMPLE_FMT_U8P:
        case AV_SAMPLE_FMT_U8:
            return 0;
        default:
            return 1;
    }
}

static void free_out_audio_packet_cb(void *packet) {
    Kit_AudioPacket *p = packet;
    Kit_DestroyRingBuffer(p->rb);
    free(p);
}

static void dec_read_audio(const Kit_Decoder *dec) {
    const Kit_AudioDecoder *audio_dec = dec->userdata;
    int len;
    int dst_linesize;
    int dst_nb_samples;
    int dst_bufsize;
    double pts;
    unsigned char **dst_data;
    Kit_AudioPacket *out_packet = NULL;
    int ret = 0;

    // Pull decoded frames out when ready and if we have room in decoder output buffer
    while(!ret && Kit_CanWriteDecoderOutput(dec)) {
        ret = avcodec_receive_frame(dec->codec_ctx, audio_dec->scratch_frame);
        if(!ret) {
            dst_nb_samples = av_rescale_rnd(
                audio_dec->scratch_frame->nb_samples,
                dec->output.samplerate,  // Target samplerate
                dec->codec_ctx->sample_rate,  // Source samplerate
                AV_ROUND_UP);

            av_samples_alloc_array_and_samples(
                &dst_data,
                &dst_linesize,
                dec->output.channels,
                dst_nb_samples,
                _FindAVSampleFormat(dec->output.format),
                0);

            len = swr_convert(
                audio_dec->swr,
                dst_data,
                dst_nb_samples,
                (const unsigned char **)audio_dec->scratch_frame->extended_data,
                audio_dec->scratch_frame->nb_samples);

            dst_bufsize = av_samples_get_buffer_size(
                &dst_linesize,
                dec->output.channels,
                len,
                _FindAVSampleFormat(dec->output.format), 1);

            // Get presentation timestamp
            pts = audio_dec->scratch_frame->best_effort_timestamp;
            pts *= av_q2d(dec->format_ctx->streams[dec->stream_index]->time_base);

            // Lock, write to audio buffer, unlock
            out_packet = _CreateAudioPacket(
                (char*)dst_data[0], (size_t)dst_bufsize, pts);
            Kit_WriteDecoderOutput(dec, out_packet);

            // Free temps
            av_freep(&dst_data[0]);
            av_freep(&dst_data);
        }
    }
}

static int dec_decode_audio_cb(Kit_Decoder *dec, AVPacket *in_packet) {
    assert(dec != NULL);
    assert(in_packet != NULL);

    // Try to clear the buffer first. We might have too much content in the ffmpeg buffer,
    // so we want to clear it of outgoing data if we can.
    dec_read_audio(dec);

    // Write packet to the decoder for handling.
    if(avcodec_send_packet(dec->codec_ctx, in_packet) < 0) {
        return 1;
    }

    // Some input data was put in successfully, so try again to get frames.
    dec_read_audio(dec);
    return 0;
}

static void dec_close_audio_cb(Kit_Decoder *dec) {
    if(dec == NULL) return;

    Kit_AudioDecoder *audio_dec = dec->userdata;
    if(audio_dec->scratch_frame != NULL) {
        av_frame_free(&audio_dec->scratch_frame);
    }
    if(audio_dec->swr != NULL) {
        swr_free(&audio_dec->swr);
    }
    free(audio_dec);
}

Kit_Decoder* Kit_CreateAudioDecoder(const Kit_Source *src, int stream_index) {
    assert(src != NULL);
    if(stream_index < 0) {
        return NULL;
    }

    const Kit_LibraryState *state = Kit_GetLibraryState();

    // First the generic decoder component ...
    Kit_Decoder *dec = Kit_CreateDecoder(
        src,
        stream_index,
        state->audio_buf_frames,
        free_out_audio_packet_cb,
        state->thread_count);
    if(dec == NULL) {
        goto EXIT_0;
    }

    // ... then allocate the audio decoder
    Kit_AudioDecoder *audio_dec = calloc(1, sizeof(Kit_AudioDecoder));
    if(audio_dec == NULL) {
        goto EXIT_1;
    }

    // Create temporary audio frame
    audio_dec->scratch_frame = av_frame_alloc();
    if(audio_dec->scratch_frame == NULL) {
        Kit_SetError("Unable to initialize temporary audio frame");
        goto EXIT_2;
    }

    // Set format configs
    Kit_OutputFormat output;
    memset(&output, 0, sizeof(Kit_OutputFormat));
    output.samplerate = dec->codec_ctx->sample_rate;
#ifdef OLD_CHANNEL_LAYOUT
    output.channels = _FindChannelLayout(dec->codec_ctx->channel_layout);
#else
    output.channels = _FindChannelLayout(&dec->codec_ctx->ch_layout);
#endif
    output.bytes = _FindBytes(dec->codec_ctx->sample_fmt);
    output.is_signed = _FindSignedness(dec->codec_ctx->sample_fmt);
    output.format = _FindSDLSampleFormat(dec->codec_ctx->sample_fmt);

    // Create resampler
#ifdef OLD_CHANNEL_LAYOUT
    audio_dec->swr = swr_alloc_set_opts(
        NULL,
        _FindAVChannelLayout(output.channels), // Target channel layout
        _FindAVSampleFormat(output.format), // Target fmt
        output.samplerate, // Target samplerate
        dec->codec_ctx->channel_layout, // Source channel layout
        dec->codec_ctx->sample_fmt, // Source fmt
        dec->codec_ctx->sample_rate, // Source samplerate
        0, NULL);
#else
    AVChannelLayout layout;
    _FindAVChannelLayout(&layout, output.channels);
     int swr_ok = swr_alloc_set_opts2(
            &audio_dec->swr,
            &layout, // Target channel layout
            _FindAVSampleFormat(output.format), // Target fmt
            output.samplerate, // Target sample rate
            &dec->codec_ctx->ch_layout, // Source channel layout
            dec->codec_ctx->sample_fmt, // Source fmt
            dec->codec_ctx->sample_rate, // Source sample rate
            0, NULL);
     if (swr_ok != 0) {
         Kit_SetError("Unable to initialize audio resampler context");
         goto EXIT_3;
     }
#endif

    if(swr_init(audio_dec->swr) != 0) {
        Kit_SetError("Unable to initialize audio resampler context");
        goto EXIT_3;
    }

    // Set callbacks and userdata, and we're go
    dec->dec_decode = dec_decode_audio_cb;
    dec->dec_close = dec_close_audio_cb;
    dec->userdata = audio_dec;
    dec->output = output;
    return dec;

EXIT_3:
    av_frame_free(&audio_dec->scratch_frame);
EXIT_2:
    free(audio_dec);
EXIT_1:
    Kit_CloseDecoder(dec);
EXIT_0:
    return NULL;
}

double Kit_GetAudioDecoderPTS(const Kit_Decoder *dec) {
    const Kit_AudioPacket *packet = Kit_PeekDecoderOutput(dec);
    if(packet == NULL) {
        return -1.0;
    }
    return packet->pts;
}

int Kit_GetAudioDecoderData(Kit_Decoder *dec, unsigned char *buf, int len) {
    assert(dec != NULL);

    Kit_AudioPacket *packet = NULL;
    int ret = 0;
    int bytes_per_sample = 0;
    double bytes_per_second = 0;
    double sync_ts = 0;

    // First, peek the next packet. Make sure we have something to read.
    packet = Kit_PeekDecoderOutput(dec);
    if(packet == NULL) {
        return 0;
    }

    // If packet should not yet be played, stop here and wait.
    // If packet should have already been played, skip it and try to find a better packet.
    // For audio, it is possible that we cannot find good packet. Then just don't read anything.
    sync_ts = _GetSystemTime() - dec->clock_sync;
    if(packet->pts > sync_ts + KIT_AUDIO_SYNC_THRESHOLD) {
        return 0;
    }
    while(packet != NULL && packet->pts < sync_ts - KIT_AUDIO_SYNC_THRESHOLD) {
        Kit_AdvanceDecoderOutput(dec);
        free_out_audio_packet_cb(packet);
        packet = Kit_PeekDecoderOutput(dec);
    }
    if(packet == NULL) {
        return 0;
    }

    // Read data from packet ringbuffer
    if(len > 0) {
        ret = Kit_ReadRingBuffer(packet->rb, (char*)buf, len);
        if(ret) {
            bytes_per_sample = dec->output.bytes * dec->output.channels;
            bytes_per_second = bytes_per_sample * dec->output.samplerate;
            packet->pts += ((double)ret) / bytes_per_second;
        }
    }
    dec->clock_pos = packet->pts;

    // If ringbuffer is cleared, kill packet and advance buffer.
    // Otherwise, forward the pts value for the current packet.
    if(Kit_GetRingBufferLength(packet->rb) == 0) {
        Kit_AdvanceDecoderOutput(dec);
        free_out_audio_packet_cb(packet);
    }
    return ret;
}