ffmpeg - 使用 libav 将 RGBA 帧编码为 MP4 但输出一团糟
问题描述
我正在尝试将视频解码为 RGB 帧,然后对帧进行后处理,最后将帧编码为视频。但是输出的视频完全是一团糟:
我写了一个最小的例子来说明我的想法。首先,我从一些源视频中读取了一些信息:
AVFormatContext* inputFormatCtx = nullptr;
int ret = avformat_open_input(&inputFormatCtx, inputParamsVideo, nullptr, nullptr);
assert(ret >= 0);
ret = avformat_find_stream_info(inputFormatCtx, NULL);
av_dump_format(inputFormatCtx, 0, inputParamsVideo, 0);
assert(ret >= 0);
AVStream* inputVideoStream = nullptr;
for (int i = 0; i < inputFormatCtx->nb_streams; i++)
{
const auto inputStream = inputFormatCtx->streams[i];
if (inputStream->codec->codec_type == AVMEDIA_TYPE_VIDEO)
{
inputVideoStream = inputStream;
break;
}
}
assert(inputVideoStream != nullptr);
AVCodecParameters* inputParams = inputVideoStream->codecpar;
AVRational framerate = inputVideoStream->codec->framerate;
auto gop_size = inputVideoStream->codec->gop_size;
auto maxBFrames = inputVideoStream->codec->max_b_frames;
然后我将信息分配给输出流:
AVFormatContext *outputAVFormat = nullptr;
avformat_alloc_output_context2(&outputAVFormat, nullptr, nullptr, kOutputPath);
assert(outputAVFormat);
AVCodec* codec = avcodec_find_encoder(outputAVFormat->oformat->video_codec);
assert(codec);
AVCodecContext* encodingCtx = avcodec_alloc_context3(codec);
avcodec_parameters_to_context(encodingCtx, inputParams);
encodingCtx->time_base = av_inv_q(framerate);
encodingCtx->max_b_frames = maxBFrames;
encodingCtx->gop_size = gop_size;
if (outputAVFormat->oformat->flags & AVFMT_GLOBALHEADER)
encodingCtx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
AVStream* outStream = avformat_new_stream(outputAVFormat, nullptr);
assert(outStream != nullptr);
ret = avcodec_parameters_from_context(outStream->codecpar, encodingCtx);
assert(ret >= 0);
outStream->time_base = encodingCtx->time_base;
然后我将 RGBA 帧(从文件中读取)转换为 YUV420Psws_scale
和编码:
ret = avcodec_open2(encodingCtx, codec, nullptr);
assert(ret >= 0);
av_dump_format(outputAVFormat, 0, kOutputPath, 1);
ret = avio_open(&outputAVFormat->pb, kOutputPath, AVIO_FLAG_WRITE);
assert(ret >= 0);
ret = avformat_write_header(outputAVFormat, nullptr);
assert(ret >= 0);
AVFrame* frame = av_frame_alloc();
frame->width = inputParams->width;
frame->height = inputParams->height;
frame->format = inputParams->format;
frame->pts = 0;
assert(ret >= 0);
ret = av_frame_get_buffer(frame, 32);
int frameCount = 0;
assert(ret >= 0);
ret = av_frame_make_writable(frame);
assert(ret >= 0);
SwsContext* swsContext = sws_getContext(inputParams->width, inputParams->height,
AV_PIX_FMT_RGBA, frame->width,
frame->height, static_cast<AVPixelFormat>(inputParams->format),
SWS_BILINEAR, NULL, NULL, NULL);
for (auto inputPicPath : std::filesystem::directory_iterator(kInputDir))
{
int width, height, comp;
unsigned char* data = stbi_load(inputPicPath.path().string().c_str(), &width, &height, &comp, 4);
int srcStrides[1] = { 4 * width };
int ret = sws_scale(swsContext, &data, srcStrides, 0, height, frame->data,
frame->linesize);
assert(ret >= 0);
frame->pts = frameCount;
//frame->pict_type = AV_PICTURE_TYPE_I;
frameCount += 1;
encode(encodingCtx, frame, 0, outputAVFormat);
stbi_image_free(data);
}
while (encode(encodingCtx, nullptr, 0, outputAVFormat))
{
;
}
static bool encode(AVCodecContext* enc_ctx, AVFrame* frame, std::uint32_t streamIndex,
AVFormatContext * formatCtx)
{
int ret;
int got_output = 0;
AVPacket packet = {};
av_init_packet(&packet);
ret = avcodec_encode_video2(enc_ctx, &packet, frame, &got_output);
assert(ret >= 0);
if (got_output) {
packet.stream_index = streamIndex;
av_packet_rescale_ts(&packet, enc_ctx->time_base, formatCtx->streams[streamIndex]->time_base);
ret = av_interleaved_write_frame(formatCtx, &packet);
assert(ret >= 0);
return true;
}
else {
return false;
}
}
最后我清理了一些东西:
av_write_trailer(outputAVFormat);
sws_freeContext(swsContext);
avcodec_free_context(&encodingCtx);
avio_closep(&outputAVFormat->pb);
avformat_free_context(outputAVFormat);
av_frame_free(&frame);
我转储了输入格式和输出格式:
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'H:\Me.MP4':
Metadata:
major_brand : mp42
minor_version : 1
compatible_brands: mp41mp42isom
creation_time : 2019-04-03T05:44:22.000000Z
Duration: 00:00:06.90, start: 0.000000, bitrate: 1268 kb/s
Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p(tv, bt709), 540x960, 1238 kb/s, 29.86 fps, 30 tbr, 600 tbn, 1200 tbc (default)
Metadata:
creation_time : 2019-04-03T05:44:22.000000Z
handler_name : Core Media Video
Stream #0:1(und): Audio: aac (LC) (mp4a / 0x6134706D), 8000 Hz, stereo, fltp, 24 kb/s (default)
Metadata:
creation_time : 2019-04-03T05:44:22.000000Z
handler_name : Core Media Audio
[libx264 @ 000002126F90C1C0] using cpu capabilities: MMX2 SSE2Fast SSSE3 SSE4.2 AVX FMA3 BMI2 AVX2
[libx264 @ 000002126F90C1C0] profile High, level 3.1, 4:2:0, 8-bit
[libx264 @ 000002126F90C1C0] 264 - core 157 - H.264/MPEG-4 AVC codec - Copyleft 2003-2018 - http://www.videolan.org/x264.html - options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=12 lookahead_threads=2 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=0 weightp=2 keyint=12 keyint_min=1 scenecut=40 intra_refresh=0 rc_lookahead=12 rc=abr mbtree=1 bitrate=1238 ratetol=1.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, mp4, to './output.mp4':
Stream #0:0: Video: h264 (High) (avc1 / 0x31637661), yuv420p(tv, bt709), 540x960, q=2-31, 1238 kb/s, 29.86 tbn
更新:
我删除后
encodingCtx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
输出视频是正确的。此外,输出 avi 也可以。
解决方案
推荐阅读
- python - 在python中请求URL并制作目录错误
- java - 初学者 Java 编码器试图弄清楚如果问题被回答否如何结束程序
- android - 如何解决android spring框架ResourceAccessException?
- c# - 如何将 sqlite 表数据翻译成印地语或马拉地语等区域语言?
- wcf - HTTP 请求未经授权...错误
- ios - 如何在 Flutter 获得 iOS 设备的权限?
- reactjs - 反应对象没有得到更新的 redux 存储值
- hsts - 如何在过时的 SSL 上访问受 HSTS 保护的私人网页?
- amazon-web-services - 为什么私有子网中的 EMR 需要完整的出站 Internet 访问权限
- java - 如何使用 Java 调用支持 RQL(资源查询语言)的 REST API