While trying to read rtsp stream I get some problems, with code and documentation alike. Short description: whatever I do, avcodec_open2 either fails (saying "codec type or id mismatches") or width and height of codec context after the call are 0 (thus making further code useless). Stream itself can be opened normally by VLC player and av_dump_format() displays correct info. My code is based on technique answer to this question.
Long description: my code is in C#, but here is C++-equivalent of FFMpeg calls (I actually reduced my code to this minimum and problem persists):
av_register_all();
avformat_network_init(); //return code ignored
AVFormatContext* formatContext = avformat_alloc_context();
if (avformat_open_input(&formatContext, stream_path, null, null) != 0) {
return;
}
if (avformat_find_stream_info(formatContext, null) < 0) {
return;
}
int videoStreamIndex = 0;
for (int i = 0; i < formatContext->nb_streams; ++i) {
AVStream* s = formatContext->streams[i];
if (s->codec == null) continue;
AVCodecContext c = *(s->codec);
if (c.codec_type == AVMEDIA_TYPE_VIDEO) videoStreamIndex = i;
}
//start reading packets from stream and write them to file
//av_read_play(formatContext); //return code ignored
//this call would print "method PLAY failed: 455 Method Not Valid in This State"
//seems to be the case that for rtsp stream it isn't needed
AVCodec* codec = null;
codec = avcodec_find_decoder(AV_CODEC_ID_H264);
if (codec == null) {
return;
}
AVCodecContext* codecContext = avcodec_alloc_context3(null);
avcodec_get_context_defaults3(codecContext, codec);//return code ignored
avcodec_copy_context(codecContext, formatContext->streams[videoStreamIndex]->codec); //return code ignored
av_dump_format(formatContext, videoStreamIndex, stream_path, 0);
if (avcodec_open2(codecContext, codec, null) < 0) {
return;
}
The code actually uses DLL version of FFMpeg library; avcodec-55.dll and avformat-55.dll are used.
Documentation says something weird about which calls can be made in which succession (that copy_context should be called before get_context_defaults), current code is left close as possible to technique version. As written, it results in non-zero return from avcodec_open2 with "codec type or id mismatches" message. Changing the order does little good: now avcodec_open2 executes successfully, but both codecContext->width and codecContext->height are 0 afterwards.
Also documentation doesn't mention which is default value for the third argument of avcodec_open2 should be, but source code seems to taking into account that options can be NULL.
Output of av_dump_format is as follows:
Input #0, rtsp, from 'rtsp://xx.xx.xx.xx:xx/video.pro1':
Metadata:
title : QStream
comment : QStreaming Media
Duration: N/A, start: 0.000000, bitrate: 64 kb/s
Stream #0:0: Video: h264 (Baseline), yuvj420p(pc), 1920x1080, 30 fps, 25 tbr, 90k tbn, 60 tbc
Stream #0:1: Audio: pcm_mulaw, 8000 Hz, 1 channels, s16, 64 kb/s
First, what does the av_dump_format shows? Are you sure your video stream codec is h264, because you try to open the codec as if it were H264.
In order to open any codec, change your avcodec_find_decoder to pass it the source codec id:
codec = avcodec_find_decoder(formatContext->streams[videoStreamIndex]->codec->codec_id);
By the way, (forget this one if you do not use the c++ code but stick with c#): you do not need to make a copy of the initial AVCodecContext when you are looking for the video stream. You can do: (note that you may want to keep a pointer to the inital codec context, see below).
AVCodecContext* c = s->codec;
if (c->codec_type == AVMEDIA_TYPE_VIDEO) {
videoStreamIndex = i;
initialVideoCodecCtx = c;
}
Next point, not really relevant in this case: instead of looping through all the steams, FFmpeg has a helper function for it:
int videoStreamIndex = av_find_best_stream(formatContext, AVMEDIA_TYPE_VIDEO, -1, -1, NULL, 0);
Last point: I think only the first point should do the trick to make avcodec_open2 work, but you might not be able to decode your stream. You opened the codec for the new codec context, but no codec is opened for the inital context. Why did you make a copy of the initial codec context? It is usefull if you want to record your stream in another file (i.e. transcode), but if you only want to decode your stream, it is much easier to use the initial context, and use it instead of the new one as a parameter for avcodec_decode_video2.
To sum it up, replace your code after avformat_find_stream_info by (warning: no error check):
int videoStreamIndex = av_find_best_stream(formatContext, AVMEDIA_TYPE_VIDEO, -1, -1, NULL, 0);
AVCodecContext* codecCtx = formatContext->streams[videoStreamIndex]->codec;
AVCodec* codec = avcodec_find_decoder(codecCtx->codec_id);
// tune codecCtx if you want special decoding options. See FFmpeg docs for a list of members
if (avcodec_open2(codecCtx, codec, null) < 0) {
return;
}
// use av_read_frame(formatContext, ...) to read packets
// use avcodec_decode_video2(codecCtx, ...) to decode packets
If avcodec_open2 does not fail, and you still see width and height being 0 this might be expected. Notice that the stream (frame) dimensions are not always known until you actually start decoding.
You should use the AVFrame values in order to initialize your decoding buffers, after your first avcodec_decode_video2 decoding call.
Related
I am trying to create a live RTMP stream containing the animation generated with NVIDIA OptiX. The stream is to be received by nginx + rtmp module and broadcasted in MPEG-DASH format. Full chain up to dash.js player is working if the video is first saved to .flv file and then I send it with ffmpeg without any reformatting using command:
ffmpeg -re -i my_video.flv -c:v copy -f flv rtmp://x.x.x.x:1935/dash/test
But I want to stream directly from the code. And with this I am failng... Nginx logs an error "dash: invalid avcc received (2: No such file or directory)". Then it seems to receive the stream correctly (segments are rolling, dash manifest is there), however the stream is not possible to play in the browser.
I can see only one difference in the manifest between direct stream and stream from file. Codecs attribute of the representation in the direct stream is missed: codecs="avcc1.000000" instead of "avc1.640028" which I get when streaming from file.
My code opens the stream:
av_register_all();
AVOutputFormat* fmt = av_guess_format("flv",
file_name, nullptr);
fmt->video_codec = AV_CODEC_ID_H264;
AVFormatContext* _oc;
avformat_alloc_output_context2(&_oc, fmt, nullptr, "rtmp://x.x.x.x:1935/dash/test");
AVStream* _vs = avformat_new_stream(_oc, nullptr);
_vs->id = 0;
_vs->time_base = AVRational { 1, 25 };
_vs->avg_frame_rate = AVRational{ 25, 1 };
AVCodecParameters *vpar = _vs->codecpar;
vpar->codec_id = fmt->video_codec;
vpar->codec_type = AVMEDIA_TYPE_VIDEO;
vpar->format = AV_PIX_FMT_YUV420P;
vpar->profile = FF_PROFILE_H264_HIGH;
vpar->level = _level;
vpar->width = _width;
vpar->height = _height;
vpar->bit_rate = _avg_bitrate;
avio_open(&_oc->pb, _oc->filename, AVIO_FLAG_WRITE);
avformat_write_header(_oc, nullptr);
Width, height, bitrate, level and profile I get from NVENC encoder settings. I also do the error checking, ommited here. Then I have a loop writing each encoded packets, with IDR frames etc all prepared on the fly with NVENC. The loop body is:
auto & pkt_data = _packets[i];
AVPacket pkt = { 0 };
av_init_packet(&pkt);
pkt.pts = av_rescale_q(_n_frames++, AVRational{ 1, 25 }, _vs->time_base);
pkt.duration = av_rescale_q(1, AVRational{ 1, 25 }, _vs->time_base);
pkt.dts = pkt.pts;
pkt.stream_index = _vs->index;
pkt.data = pkt_data.data();
pkt.size = (int)pkt_data.size();
if (!memcmp(pkt_data.data(), "\x00\x00\x00\x01\x67", 5))
{
pkt.flags |= AV_PKT_FLAG_KEY;
}
av_write_frame(_oc, &pkt);
Obviously ffmpeg is writing avcc code somewhere... I have no clue where to add this code so the RTMP server can recognize it. Or I am missing something else?
Any hint greatly appreciated, folks!
Thanks to Gyan's comment I was able to solve the issue. Following the AV_CODEC_FLAG_GLOBAL_HEADER flag in the wrapper one can see how the global header is added, which was missing in my case. You can use directly the NVENC API function nvEncGetSequenceParams, but since I am anyway using SDK, it is a bit cleaner.
So I had to attach the header to AVCodecParameters::extradata:
std::vector<uint8_t> payload;
_encoder->GetSequenceParams(payload);
vpar->extradata_size = payload.size();
vpar->extradata = (uint8_t*)av_mallocz(payload.size() + AV_INPUT_BUFFER_PADDING_SIZE);
memcpy(vpar->extradata, payload.data(), payload.size());
_encoder is my instance of NvEncoder from SDK.
The wrapper is doing the same thing, however using deprecated struct AVCodecContext.
I’m trying to transcode a video on my iOS app using FFMpeg/LibAv.
What I’m trying to accomplish is to transcode a video in order to resize each frame and possibly lower the bitrate in order to save valuable MB in the device.
The resulting video must be playable on all iPhone5+ devices.
After reading the documentation I found out that:
I do not need to encode/decode the audio stream -> I’ll copy as-is to the output file
I need to encode the video using the h264 codec (LibX264) with a profile supported by iOS (baseline profile with level 3.0 - https://trac.ffmpeg.org/wiki/Encode/H.264#Compatibility)
I’m also setting the picture format to YUV planar since it’s the only one supported by iOS
For the sake of testing I’m not using any filter (I’m using a dummy/passthrough) at all or even trying to lower the bitrate, I’m just trying to decode the video stream and encode it again
Most of the code is based on the transcoding.c and filtering.c available on the FFMpeg examples directory
FFMpeg-wise what I’m trying to achieve with LibAv is:
ffmpeg -i INPUT.MOV -c:v libx264 -preset ultrafast -profile:v baseline -level 3.0 -c:a copy output.MOV
(the resulting file - which can be found below - is playable on QuickTime if it’s generated by FFMpeg through the command line)
The original video was generated with a regular iPhone using iOS 8.2 but the problem is not device specific or iOS specific, it occurs on all videos generated with LibAv.
Although both resulting files are playable by VideoLan (VLC) the one I generated through LibAv is not playable by QuickTime even though I can’t find anything wrong with it.
As you can see below, I create the video stream with the proper video codec on the call to avformat_new_stream:
AVStream *out_stream; // output stream
AVStream *in_stream; // input stream
AVCodecContext *dec_ctx, *enc_ctx; // codec context for the stream
AVCodec *encoder; // codec used
int ret;
unsigned int i;
ofmt_ctx = NULL;
// Allocate an AVFormatContext for an output format. This will be the file header (similar to avformat_open_input but with an zero'ed memory)
avformat_alloc_output_context2(&ofmt_ctx, NULL, NULL, filename);
if (!ofmt_ctx) {
av_log(NULL, AV_LOG_ERROR, "Could not create output context\n");
[self errorWith:kErrorCreatingOutputContext and:#"Could not create output context"];
return AVERROR_UNKNOWN;
}
// we must not use the AVCodecContext from the video stream directly! So we have to use avcodec_copy_context() to copy the context to a new location (after allocating memory for it, of course).
// iterate over all input streams
for (i = 0; i < ifmt_ctx->nb_streams; i++) {
in_stream = ifmt_ctx->streams[i]; // input stream
dec_ctx = in_stream->codec; // get the codec context for the decoder
if (dec_ctx->codec_type == AVMEDIA_TYPE_VIDEO) {
// lets use h264
encoder = avcodec_find_encoder(AV_CODEC_ID_H264);
if (!encoder) {
[self errorWith:kErrorCodecNotFound and:#"H264 Codec Not Found"];
return AVERROR_UNKNOWN;
}
out_stream = avformat_new_stream(ofmt_ctx, encoder); // create a new stream with h264 codec
if (!out_stream) {
av_log(NULL, AV_LOG_ERROR, "Failed allocating output stream\n");
[self errorWith:kErrorAllocateOutputStream and:#"Failed allocating output stream"];
return AVERROR_UNKNOWN;
}
enc_ctx = out_stream->codec; // pointer to the stream codec context
/* we transcode to same properties (picture size,
* sample rate etc.). These properties can be changed for output
* streams easily using filters */
if (dec_ctx->codec_type == AVMEDIA_TYPE_VIDEO) {
enc_ctx->width = dec_ctx->width;
enc_ctx->height = dec_ctx->height;
enc_ctx->sample_aspect_ratio = dec_ctx->sample_aspect_ratio;
enc_ctx->pix_fmt = AV_PIX_FMT_YUV420P;
enc_ctx->time_base = dec_ctx->time_base;
av_opt_set(enc_ctx->priv_data, "preset", "ultrafast", 0);
av_opt_set(enc_ctx->priv_data, "profile", "baseline", 0);
av_opt_set(enc_ctx->priv_data, "level", "3.0", 0);
}
out_stream->time_base = in_stream->time_base;
AVDictionaryEntry *tag = NULL;
while ((tag = av_dict_get(in_stream->metadata, "", tag, AV_DICT_IGNORE_SUFFIX))) {
printf("%s=%s\n", tag->key, tag->value);
char *k = av_strdup(tag->key); // if your strings are already allocated,
char *v = av_strdup(tag->value); // you can avoid copying them like this
av_dict_set(&out_stream->metadata, k, v, 0);
}
ret = avcodec_open2(enc_ctx, encoder, NULL);
if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Cannot open video encoder for stream #%u\n", i);
[self errorWith:kErrorCantOpenOutputFile and:[NSString stringWithFormat:#"Cannot open video encoder for stream #%u",i]];
return ret;
}
}
else if(dec_ctx->codec_type == AVMEDIA_TYPE_UNKNOWN) {
// if we cant figure out the stream type, fail
av_log(NULL, AV_LOG_FATAL, "Elementary stream #%d is of unknown type, cannot proceed\n", i);
[self errorWith:kErrorUnknownStream and:[NSString stringWithFormat:#"Elementary stream #%d is of unknown type, cannot proceed",i]];
return AVERROR_INVALIDDATA;
}
else {
out_stream = avformat_new_stream(ofmt_ctx, NULL);
if (!out_stream) {
av_log(NULL, AV_LOG_ERROR, "Failed allocating output stream\n");
[self errorWith:kErrorAllocateOutputStream and:#"Failed allocating output stream"];
return AVERROR_UNKNOWN;
}
enc_ctx = out_stream->codec;
/* this stream must be remuxed */
// copies ifmt_ctx->streams[i]->codec into ofmt_ctx->streams[i]->codec - Copy the settings of the source AVCodecContext into the destination AVCodecContext.
ret = avcodec_copy_context(ofmt_ctx->streams[i]->codec,
ifmt_ctx->streams[i]->codec);
if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Copying stream context failed\n");
[self errorWith:kErrorCopyStreamFailed and:#"Copying stream context failed"];
return ret;
}
}
// dunno what this is for
if (ofmt_ctx->oformat->flags & AVFMT_GLOBALHEADER)
enc_ctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
}
if (!(ofmt_ctx->oformat->flags & AVFMT_NOFILE)) {
// Create and initialize a AVIOContext for accessing the
// resource indicated by url.
ret = avio_open(&ofmt_ctx->pb, filename, AVIO_FLAG_WRITE);
if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Could not open output file '%s'", filename);
[self errorWith:kErrorCantOpenOutputFile and:[NSString stringWithFormat:#"Could not open output file '%s'", filename]];
return ret;
}
}
/* init muxer, write output file header */
// Allocate the stream private data and write the stream header to an output media file.
ret = avformat_write_header(ofmt_ctx, NULL);
if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Error occurred when opening output file\n");
[self errorWith:kErrorOutFileCantWriteHeader and:#"Error occurred when opening output file"];
return ret;
}
return 0;
You can find the files here:
Original final: https://www.dropbox.com/s/2jjs1uy2pu2veyy/IMG_5705.MOV?dl=0
File generated with FFMpeg - https://www.dropbox.com/s/9hfmq3fcifgpfqc/local-ffmpeg.MOV?dl=0
File generated by code - https://www.dropbox.com/s/rttvny39rj7ejpf/generated-by-Ze.MOV?dl=0
Thank you so much,
Ze
I use ffmpeg's MPEG4 decoder. The decoder has CODEC_CAP_DELAY capability among others. It means the decoder will give me decoded frames with latency of 1 frame.
I have a set of MPEG4 (I- & P- )frames from AVI file and feed ffmpeg decoder with these frames. For the very first I-frame decoder gives me nothing, but decodes the frames successfully. I can force the decoder to get the decoded frame with the second call of avcodec_decode_video2 and providing nulls (flush it), but if I do so for each frame I get artifacts for the first group of pictures (e.g. second decoded P-frame is of gray color).
If I do not force ffmpeg decoder to give me decoded frame right now, then it works flawlessly and without artifacts.
Question: But is it possible to get decoded frame without giving the decoder next frame and without artifacts?
Small example of how decoding is implemented for each frame:
// decode
int got_frame = 0;
int err = 0;
int tries = 5;
do
{
err = avcodec_decode_video2(m_CodecContext, m_Frame, &got_frame, &m_Packet);
/* some codecs, such as MPEG, transmit the I and P frame with a
latency of one frame. You must do the following to have a
chance to get the last frame of the video */
m_Packet.data = NULL;
m_Packet.size = 0;
--tries;
}
while (err >= 0 && got_frame == 0 && tries > 0);
But as I said that gave me artifacts for the first gop.
Use the "-flags +low_delay" option (or in code, set AVCodecContext.flags |= CODEC_FLAG_LOW_DELAY).
I tested several options and "-flags low_delay" and "-probesize 32" is more important than others. bellow code worked for me.
AVDictionary* avDic = nullptr;
av_dict_set(&avDic, "flags", "low_delay", 0);
av_dict_set(&avDic, "probesize", "32", 0);
const int errorCode = avformat_open_input(&pFormatCtx, mUrl.c_str(), nullptr, &avDic);
I want to realize an application that firstly decode a multi-media file(such as test.mp4 file, video codec id is H264), get a video stream and an audio stream, then make some different in the audio stream, at last encode the video stream(use libx264) and audio stream into a result file(result.mp4). To promote the efficiency, i omitted the decode and encode of video stream, i get the video packet via function "av_read_frame", then output it directly into the result file via function "av_write_frame". But there is no picture in the output file, and the size of output file is fairly small.
I tracked the ffmpeg code and found that in the function "av_write_frame->mov_write_packet->ff_mov_write_packet", it will call function "ff_avc_parse_nal_units" to obtain the size of nal unit, but the return value is very small(such as 208 bytes).
I find that the H264 stream in the MP4 file is not stored in Annex-B format, so it can't find start code(0x000001), now my problem is how can I change the H264 stream to Annex-B format, and make it work?
I added start code at the beginning of every frame manually, but it still not work.
Anyone can give me any hint?Thanks very much.
Following is the codes similar with my:
// write the stream header, if any
av_write_header(pFormatCtxEnc);
.........
/**
* Init of Encoder and Decoder
*/
bool KeyFlag = false;
bool KeyFlagEx = false;
// Read frames and save frames to disk
int iPts = 1;
av_init_packet(&packet);
while(av_read_frame(pFormatCtxDec, &packet)>=0)
{
if (packet.flags == 1)
KeyFlag = true;
if (!KeyFlag)
continue;
if (m_bStop)
{
break;
}
// Is this a packet from the video stream?
if(packet.stream_index == videoStream)
{
currentframeNum ++;
if (progressCB != NULL && currentframeNum%20 == 0)
{
float fpercent = (float)currentframeNum/frameNum;
progressCB(fpercent,m_pUser);
}
if (currentframeNum >= beginFrame && currentframeNum <= endFrane)
{
if (packet.flags == 1)
KeyFlagEx = true;
if (!KeyFlagEx)
continue;
packet.dts = iPts ++;
av_write_frame(pFormatCtxEnc, &packet);
}
}
// Free the packet that was allocated by av_read_frame
}
// write the trailer, if any
av_write_trailer(pFormatCtxEnc);
/**
* Release of encoder and decoder
*/
return true;
You might try this: libavcodec/h264_mp4toannexb_bsf.c. It converts bitstream without start codes to bitstream with start codes.
Using your source file, does ffmpeg -i src.mp4 -vcodec copy -an dst.mp4 work? Does it work if you add -bsf h264_mp4toannexb? (all using the same version/build of ffmpeg as you are trying to use programmatically of course)
Converting 3gp (amr) to mp3 using ffmpeg api calls
I try to use libavformat (ffmpeg) to build my own function that converts 3gp audio files (recorded with an android mobile device) into mp3 files.
I use av_read_frame() to read a frame from the input file and use avcodec_decode_audio3() to decode the data
into a buffer and use this buffer to encode the data into mp3 with avcodec_encode_audio.
This seems to give me a correct result for converting wav to mp3 and mp3 to wav (Or decode one mp3 and encode to another mp3) but not for amr to mp3.
My resulting mp3 file seems to has the right length but only consists of noise.
In another post I read that amr-decoder does not use the same sample format than mp3 does.
AMR uses FLT and mp3 S16 or S32 und that I have to do resampling.
So I call av_audio_resample_init() and audio_resample for each frame that has been decoded.
But that does not solve my problem completely. Now I can hear my recorded voice and unsterstand what I was saying, but the quality is very low and there is still a lot of noise.
I am not sure if I set the parameters of av_audio_resample correctly, especially the last 4 parameters (I think not) or if I miss something else.
ReSampleContext* reSampleContext = av_audio_resample_init(1, 1, 44100, 8000, AV_SAMPLE_FMT_S32, AV_SAMPLE_FMT_FLT, 0, 0, 0, 0.0);
while(1)
{
if(av_read_frame(ic, &avpkt) < 0)
{
break;
}
out_size = AVCODEC_MAX_AUDIO_FRAME_SIZE;
int count;
count = avcodec_decode_audio3(audio_stream->codec, (short *)decodedBuffer, &out_size, &avpkt);
if(count < 0)
{
break;
}
if((audio_resample(reSampleContext, (short *)resampledBuffer, (short *)decodedBuffer, out_size / 4)) < 0)
{
fprintf(stderr, "Error\n");
exit(1);
}
out_size = AVCODEC_MAX_AUDIO_FRAME_SIZE;
pktOut.size = avcodec_encode_audio(c, outbuf, out_size, (short *)resampledBuffer);
if(c->coded_frame && c->coded_frame->pts != AV_NOPTS_VALUE)
{
pktOut.pts = av_rescale_q(c->coded_frame->pts, c->time_base, outStream->time_base);
//av_res
}
pktOut.pts = AV_NOPTS_VALUE;
pktOut.dts = AV_NOPTS_VALUE;
pktOut.flags |= AV_PKT_FLAG_KEY;
pktOut.stream_index = audio_stream->index;
pktOut.data = outbuf;
if(av_write_frame(oc, &pktOut) != 0)
{
fprintf(stderr, "Error while writing audio frame\n");
exit(1);
}
}