Hello Guys,
I cut my mpeg2-material framebased with this algorithm at the beginning of a
cut:
<snip>
int bytesRemaining = mCodecPointer;
int bytesDecoded = 0;
int frameFinished = 0;
int bufPointer = 0;
AVFrame* frame = avcodec_alloc_frame();
// Temporaerer VideoBuffer
uint8_t* video_outbuf = (uint8_t*) av_malloc(SMALL_BUFFER);
int pts = 120;
int frameNr = 0;
int out_size = -1;
while (mCoDecFramePointer != mCoDecFrameSizes.end())
{
bytesRemaining = (int) *mCoDecFramePointer;
cout << "Dekodiere ..." << bytesRemaining << endl;
bytesDecoded += avcodec_decode_video(mPCodecCtx, frame, &frameFinished,
&mCoDecBuffer[bufPointer], bytesRemaining);
bufPointer += bytesRemaining;
pts += (40 * frameNr) * 90;
if (frameFinished)
{
frame->pts += pts; // <= Hier liegt das Problem
//frame->key_frame = PKT_FLAG_KEY;
out_size = avcodec_encode_video(mPOutputCodecCtx, video_outbuf,
SMALL_BUFFER,
frame);
frameNr++;
cout << "Encoder: OutSize ist " << out_size << endl;
if (out_size>0)
{
cout << "Frame kodiert" << endl;
AVPacket pkt;
av_init_packet(&pkt);
pkt.size = out_size;
pkt.data = video_outbuf;
packetToCutBuffer(&pkt);
}
}
mCoDecFramePointer++;
}
int i = 0;
// Buffer leeren
for (; out_size; i++)
{
out_size = avcodec_encode_video(mPOutputCodecCtx, video_outbuf,
SMALL_BUFFER,
NULL);
if (out_size>0)
{
cout << "Letztes Frame kodiert" << endl;
AVPacket pkt;
av_init_packet(&pkt);
pkt.size = out_size;
pkt.data = video_outbuf;
packetToCutBuffer(&pkt);
}
}
av_free(video_outbuf);
<snap>
This code is placed in a method called encodeVideo and the background is to
encode the beginning of a cut, if I cut framebased and if the cutin is not an
I-Frame. My mPOutputCodecCtx looks like:
<snip>
mPOutputCodecCtx = mAVStream->codec;
mPOutputCodecCtx->codec_id = mPCodecCtx->codec_id;
mPOutputCodecCtx->codec_type = CODEC_TYPE_VIDEO;
/* resolution must be a multiple of two */
mPOutputCodecCtx->width = mPCodecCtx->width; //mWidth;
mPOutputCodecCtx->height = mPCodecCtx->height; //mHeight;
/* time base: 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
identically 1. */
mPOutputCodecCtx->time_base.den = mFrameRate; //mPCodecCtx->time_base.den;
mPOutputCodecCtx->time_base.num = 1; //mPCodecCtx->time_base.num;
//c->gop_size = 12; /* emit one intra frame every twelve frames at most */
mPOutputCodecCtx->pix_fmt = mPCodecCtx->pix_fmt;
mPOutputCodecCtx->sample_aspect_ratio = mPCodecCtx->sample_aspect_ratio;
mPOutputCodecCtx->gop_size = mPCodecCtx->gop_size;
mPOutputCodecCtx->max_b_frames = 5;
mPOutputCodecCtx->bit_rate = mPCodecCtx->bit_rate;
mPOutputCodecCtx->time_base= mPCodecCtx->time_base;
mPOutputCodecCtx->palctrl = mPCodecCtx->palctrl;
mAVStream->sample_aspect_ratio = mPCodecCtx->sample_aspect_ratio;
mPOutputCodec = avcodec_find_encoder(mPOutputCodecCtx->codec_id);
if (avcodec_open(mPOutputCodecCtx, mPOutputCodec) < 0)
{
cout << "[VIDEOTRACK]: Could not open OutputCodec!" << endl;
}
else
{
cout << "[VIDEOTRACK]: Ready for new encoding!" << endl;
}
<snap>
After the encoding I copy the next GOP's beginning with an I-Frame into the
output stream and mux it into an mpg-file. This works, but I get a dirty
noise if I watch my video. Some pictures get noisy and I see some dirty
pixel-errors. What is the Problem? Could anyone give me an hint how I make it
better?
Many thanks,
Sven
_______________________________________________
libav-user mailing list
[email protected]
https://lists.mplayerhq.hu/mailman/listinfo/libav-user
