hunting/utils/MediaBase/src/FfmpegEncoderV2.cpp

/*
 * Copyright (c) 2023 Fancy Code.
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
#include "FfmpegEncoderV2.h"
#include "ILog.h"
#ifdef __cplusplus
extern "C" {
#endif
#include <libavcodec/avcodec.h>
#include <libavcodec/codec.h>
#include <libavcodec/codec_id.h>
#include <libavcodec/packet.h>
#include <libavformat/avformat.h>
// #include <libavutil/avassert.h>
#include <libavutil/avutil.h>
#include <libavutil/channel_layout.h>
#include <libavutil/dict.h>
#include <libavutil/error.h>
#include <libavutil/frame.h>
#include <libavutil/mathematics.h>
#include <libavutil/opt.h>
#include <libavutil/pixfmt.h>
#include <libavutil/samplefmt.h>
#include <libswresample/swresample.h>
#ifdef __cplusplus
}
#endif
#include <cstdint>
#include <errno.h>
#include <functional>
constexpr long SOURCE_AUDIO_SAMPEL_RATE = 8000;
#define STREAM_DURATION 10.0
#define STREAM_FRAME_RATE 25              /* 25 images/s */
#define STREAM_PIX_FMT AV_PIX_FMT_YUV420P /* default pix_fmt */
FfmpegEncoderV2::FfmpegEncoderV2(const enum AVCodecID &codecId, const AVPixelFormat &encodePixelFormat,
                                 const int &width, const int &height)
    : mCodecId(codecId), mCodecCtx(nullptr), mCodec(nullptr), mFrame(nullptr), mTmpFrame(nullptr), mTmpPkt(nullptr),
      mSamplesCount(0), mSwrCtx(nullptr), next_pts(0), mVideoWidth(width), mVideoHeight(height),
      mEncodePixelFormat(encodePixelFormat)
{
}
bool FfmpegEncoderV2::Init(const int &outputFlags)
{
    mTmpPkt = av_packet_alloc();
    if (!mTmpPkt) {
        LogError("Could not allocate AVPacket\n");
        return false;
    }
    LogInfo("find encoder : %s\n", avcodec_get_name(mCodecId));
    int i = 0;
    /* find the encoder */
    mCodec = (AVCodec *)avcodec_find_encoder(mCodecId);
    if (!mCodec) {
        LogError("Could not find encoder for '%s'\n", avcodec_get_name(mCodecId));
        return false;
    }
    mCodecCtx = avcodec_alloc_context3(mCodec);
    if (!mCodecCtx) {
        LogError("Could not alloc an encoding context\n");
        return false;
    }
    const AVChannelLayout src = (AVChannelLayout)AV_CHANNEL_LAYOUT_MONO;
    switch (mCodec->type) {
    case AVMEDIA_TYPE_AUDIO:
        mCodecCtx->sample_fmt = mCodec->sample_fmts ? mCodec->sample_fmts[0] : AV_SAMPLE_FMT_FLTP;
        mCodecCtx->bit_rate = 64000;
        // mCodecCtx->bit_rate = 24000;
        mCodecCtx->sample_rate = 44100;
        if (mCodec->supported_samplerates) {
            mCodecCtx->sample_rate = mCodec->supported_samplerates[0];
            for (i = 0; mCodec->supported_samplerates[i]; i++) {
                if (mCodec->supported_samplerates[i] == 44100)
                    mCodecCtx->sample_rate = 44100;
            }
        }
        mCodecCtx->sample_rate = 16000;
        // mCodecCtx->time_base = (AVRational){1, mCodecCtx->sample_rate};
        // mCodecCtx->ch_layout.nb_channels = 1;
        // av_channel_layout_default(&mCodecCtx->ch_layout, 1);
        av_channel_layout_copy(&mCodecCtx->ch_layout, &src);
        break;

    case AVMEDIA_TYPE_VIDEO:
        mCodecCtx->codec_id = mCodecId;

        mCodecCtx->bit_rate = 300000;
        /* Resolution must be a multiple of two. */
        mCodecCtx->width = mVideoWidth;
        mCodecCtx->height = mVideoHeight;
        /* timebase: This is the fundamental unit of time (in seconds) in terms
         * of which frame timestamps are represented. For fixed-fps content,
         * timebase should be 1/framerate and timestamp increments should be
         * identical to 1. */
        mCodecCtx->time_base = (AVRational){1, STREAM_FRAME_RATE};

        mCodecCtx->gop_size = 12; /* emit one intra frame every twelve frames at most */
        mCodecCtx->pix_fmt = mEncodePixelFormat;
        if (mCodecCtx->codec_id == AV_CODEC_ID_MPEG2VIDEO) {
            /* just for testing, we also add B-frames */
            mCodecCtx->max_b_frames = 2;
        }
        if (mCodecCtx->codec_id == AV_CODEC_ID_MPEG1VIDEO) {
            /* Needed to avoid using macroblocks in which some coeffs overflow.
             * This does not happen with normal video, it just happens here as
             * the motion of the chroma plane does not match the luma plane. */
            mCodecCtx->mb_decision = 2;
        }
        break;

    default:
        break;
    }
    /* Some formats want stream headers to be separate. */
    if (outputFlags & AVFMT_GLOBALHEADER) {
        mCodecCtx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
    }
    return true;
}
void FfmpegEncoderV2::UnInit(void)
{
    if (mFrame) {
        av_frame_free(&mFrame);
        mFrame = nullptr;
    }
    if (mTmpFrame) {
        av_frame_free(&mTmpFrame);
        mTmpFrame = nullptr;
    }
    if (mCodecCtx) {
        avcodec_free_context(&mCodecCtx);
        mCodecCtx = nullptr;
    }
    av_packet_free(&mTmpPkt);
    swr_free(&mSwrCtx);
}
AVRational FfmpegEncoderV2::GetTimeBase(void)
{
    switch (mCodec->type) {
    case AVMEDIA_TYPE_AUDIO:
        return (AVRational){1, mCodecCtx->sample_rate};

    case AVMEDIA_TYPE_VIDEO:
        return mCodecCtx->time_base;

    default:
        LogError("Unsupported media type.\n");
        return (AVRational){0, -1};
    }
}
bool FfmpegEncoderV2::OpenEncoder(AVDictionary *optArg, AVStream *stream)
{
    switch (mCodec->type) {
    case AVMEDIA_TYPE_AUDIO:
        return OpenAudio(optArg, stream);

    case AVMEDIA_TYPE_VIDEO:
        return OpenVideo(optArg, stream);

    default:
        LogError("Unsupported media type.\n");
        return false;
    }
}
// static void save_code_stream_file(const void *data, const size_t &size)
// {
//     char OutPath[16];
//     const void *pData = data;
//     FILE *file = NULL;
//     LogInfo("save_code_stream_file: %d\n", size);
//     sprintf(OutPath, "./test.jpg");
//     file = fopen(OutPath, "a+");

//     if (file) { // TODO: Don't open very time.
//         fwrite(pData, 1, size, file);
//         fflush(file);
//     }

//     if (file)
//         fclose(file);
// }
int FfmpegEncoderV2::EncodeData(AVFrame *frame, AVStream *stream, std::function<void(AVPacket *pkt)> callback)
{
    int ret = 0;
    AVFrame *tmpFrame = frame;
    if (AVMEDIA_TYPE_AUDIO == mCodec->type) {
        tmpFrame = ConvertAudioFrame(frame, mSwrCtx);
    }
    if (!tmpFrame) {
        LogError("Could not convert audio frame.\n");
        return AVERROR_EXIT;
    }
    // send the frame to the encoder
    ret = avcodec_send_frame(mCodecCtx, tmpFrame);
    if (ret < 0) {
        char error_str[AV_ERROR_MAX_STRING_SIZE] = {0};
        LogInfo("Error sending a frame to the encoder: %s\n",
                av_make_error_string(error_str, AV_ERROR_MAX_STRING_SIZE, ret));
        return AVERROR_EXIT;
    }

    while (ret >= 0) {
        ret = avcodec_receive_packet(mCodecCtx, mTmpPkt);
        if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF) {
            break;
        }
        if (ret < 0) {
            char error_str[AV_ERROR_MAX_STRING_SIZE] = {0};
            LogInfo("Error encoding a frame: %s\n", av_make_error_string(error_str, AV_ERROR_MAX_STRING_SIZE, ret));
            return AVERROR_EXIT;
        }

        /* rescale output packet timestamp values from codec to stream timebase */
        av_packet_rescale_ts(mTmpPkt, mCodecCtx->time_base, stream->time_base);
        // LogInfo("Write mCodecCtx->time_base.num: %d\n", mCodecCtx->time_base.num);
        // LogInfo("Write mCodecCtx->time_base.den: %d\n", mCodecCtx->time_base.den);
        // LogInfo("Write stream->time_base.num: %d\n", stream->time_base.num);
        // LogInfo("Write stream->time_base.den: %d\n", stream->time_base.den);
        mTmpPkt->stream_index = stream->index;
        // LogInfo(" Write frame mTmpPkt->pts: %llu\n", mTmpPkt->pts);

        if (callback) {
            // if (mCodecId == AV_CODEC_ID_MJPEG) {
            //     save_code_stream_file(mTmpPkt->data, mTmpPkt->size);
            // }
            callback(mTmpPkt);
        }
    }

    return ret == AVERROR_EOF ? 1 : 0;
}
bool FfmpegEncoderV2::OpenVideo(AVDictionary *optArg, AVStream *stream)
{
    int ret = 0;
    AVDictionary *opt = nullptr;
    av_dict_copy(&opt, optArg, 0);
    // av_dict_set(&opt, "strict_std_compliance", "experimental", 0);
    av_opt_set(mCodecCtx, "strict", "unofficial", 0); // Add for jpeg
    /* open the codec */
    ret = avcodec_open2(mCodecCtx, mCodec, &opt);
    av_dict_free(&opt);
    if (ret < 0) {
        char error_str[AV_ERROR_MAX_STRING_SIZE] = {0};
        LogError("Could not open video codec: %s\n", av_make_error_string(error_str, AV_ERROR_MAX_STRING_SIZE, ret));
        return false;
    }
    /* allocate and init a re-usable frame */
    mFrame = alloc_frame(mCodecCtx->pix_fmt, mCodecCtx->width, mCodecCtx->height);
    if (!mFrame) {
        LogError("Could not allocate video frame\n");
        return false;
    }
    if (mCodecCtx->pix_fmt != AV_PIX_FMT_YUV420P) {
        mTmpFrame = alloc_frame(AV_PIX_FMT_YUV420P, mCodecCtx->width, mCodecCtx->height);
        if (!mTmpFrame) {
            LogError("Could not allocate temporary video frame\n");
            return false;
        }
    }
    /* copy the stream parameters to the muxer */
    ret = avcodec_parameters_from_context(stream->codecpar, mCodecCtx);
    if (ret < 0) {
        LogError("Could not copy the stream parameters\n");
        return false;
    }
    LogInfo(" Open video success, mCodecCtx->pix_fmt = %d\n", mCodecCtx->pix_fmt);
    return true;
}
bool FfmpegEncoderV2::OpenAudio(AVDictionary *optArg, AVStream *stream)
{
    int nb_samples = 0;
    int ret = 0;
    AVDictionary *opt = nullptr;
    av_dict_copy(&opt, optArg, 0);
    /* open it */
    ret = avcodec_open2(mCodecCtx, mCodec, &opt);
    av_dict_free(&opt);
    if (ret < 0) {
        char error_str[AV_ERROR_MAX_STRING_SIZE] = {0};
        LogError("Could not open audio codec: %s\n", av_make_error_string(error_str, AV_ERROR_MAX_STRING_SIZE, ret));
        return false;
    }
    if (mCodecCtx->codec->capabilities & AV_CODEC_CAP_VARIABLE_FRAME_SIZE)
        nb_samples = 10000;
    else
        nb_samples = mCodecCtx->frame_size;
    mFrame = alloc_audio_frame(mCodecCtx->sample_fmt, &mCodecCtx->ch_layout, mCodecCtx->sample_rate, nb_samples);
    // mTmpFrame = alloc_audio_frame(AV_SAMPLE_FMT_S16, &mCodecCtx->ch_layout, mCodecCtx->sample_rate, nb_samples);
    /* copy the stream parameters to the muxer */
    ret = avcodec_parameters_from_context(stream->codecpar, mCodecCtx);
    if (ret < 0) {
        LogError("Could not copy the stream parameters\n");
        return false;
    }
    /* create resampler context */
    mSwrCtx = swr_alloc();
    if (!mSwrCtx) {
        LogError("Could not allocate resampler context\n");
        return false;
    }
    const AVChannelLayout src = (AVChannelLayout)AV_CHANNEL_LAYOUT_MONO;
    AVChannelLayout ch_layout;
    av_channel_layout_copy(&ch_layout, &src);
    /* set options */
    av_opt_set_chlayout(mSwrCtx, "in_chlayout", &ch_layout, 0);
    // av_opt_set_chlayout(mSwrCtx, "in_chlayout", &mCodecCtx->ch_layout, 0);
    av_opt_set_int(mSwrCtx, "in_sample_rate", SOURCE_AUDIO_SAMPEL_RATE, 0);
    av_opt_set_sample_fmt(mSwrCtx, "in_sample_fmt", AV_SAMPLE_FMT_S16, 0);
    av_opt_set_chlayout(mSwrCtx, "out_chlayout", &mCodecCtx->ch_layout, 0);
    av_opt_set_int(mSwrCtx, "out_sample_rate", mCodecCtx->sample_rate, 0);
    av_opt_set_sample_fmt(mSwrCtx, "out_sample_fmt", mCodecCtx->sample_fmt, 0);
    /* initialize the resampling context */
    if ((ret = swr_init(mSwrCtx)) < 0) {
        LogError("Failed to initialize the resampling context\n");
        return false;
    }
    return true;
}
AVFrame *FfmpegEncoderV2::ConvertAudioFrame(AVFrame *decodeFrame, struct SwrContext *swr_ctx)
{
    if (nullptr == decodeFrame) {
        LogError("decodeFrame is null\n");
        return nullptr;
    }
    // LogInfo("decodeFrame->pts = %d\n", decodeFrame->pts);
    // decodeFrame->pts = next_pts;
    // next_pts += decodeFrame->nb_samples;
    int ret = 0;
    int dst_nb_samples = 0;
    /* convert samples from native format to destination codec format, using the resampler */
    /* compute destination number of samples */
    dst_nb_samples = av_rescale_rnd(swr_get_delay(swr_ctx, mCodecCtx->sample_rate) + decodeFrame->nb_samples,
                                    mCodecCtx->sample_rate,
                                    SOURCE_AUDIO_SAMPEL_RATE,
                                    AV_ROUND_UP);
    // av_assert0(dst_nb_samples == decodeFrame->nb_samples);

    /* when we pass a frame to the encoder, it may keep a reference to it
     * internally;
     * make sure we do not overwrite it here
     */
    ret = av_frame_make_writable(mFrame);
    if (ret < 0) {
        LogError("av_frame_make_writable failed\n");
        return nullptr;
    }

    /* convert to destination format */
    ret = swr_convert(
        swr_ctx, mFrame->data, dst_nb_samples, (const uint8_t **)decodeFrame->data, decodeFrame->nb_samples);
    if (ret < 0) {
        LogError("Error while converting\n");
        return nullptr;
    }
    // LogInfo("mCodecCtx->time_base.num = %d, mCodecCtx->time_base.den=%d\n",
    //         mCodecCtx->time_base.num,
    //         mCodecCtx->time_base.den);
    mFrame->pts = av_rescale_q(decodeFrame->pts, (AVRational){1, 1000000}, mCodecCtx->time_base);
    // LogInfo("decodeFrame->pts = %d\n", decodeFrame->pts);
    // LogInfo("mFrame->pts = %d\n", mFrame->pts);
    mSamplesCount += dst_nb_samples;
    return mFrame;
}
AVFrame *FfmpegEncoderV2::alloc_frame(enum AVPixelFormat pix_fmt, int width, int height)
{
    AVFrame *frame;
    int ret;

    frame = av_frame_alloc();
    if (!frame)
        return nullptr;

    frame->format = pix_fmt;
    frame->width = width;
    frame->height = height;

    /* allocate the buffers for the frame data */
    ret = av_frame_get_buffer(frame, 0);
    if (ret < 0) {
        LogInfo("Could not allocate frame data.\n");
        return nullptr;
    }

    return frame;
}
AVFrame *FfmpegEncoderV2::alloc_audio_frame(enum AVSampleFormat sample_fmt, const AVChannelLayout *channel_layout,
                                            int sample_rate, int nb_samples)
{
    AVFrame *frame = av_frame_alloc();
    if (!frame) {
        LogError("Error allocating an audio frame\n");
        return nullptr;
    }

    frame->format = sample_fmt;
    av_channel_layout_copy(&frame->ch_layout, channel_layout);
    frame->sample_rate = sample_rate;
    frame->nb_samples = nb_samples;

    if (nb_samples) {
        if (av_frame_get_buffer(frame, 0) < 0) {
            LogError("Error allocating an audio buffer\n");
            return nullptr;
        }
    }

    return frame;
}