I have several questions regarding your FFmpeg library: 1. Please answer if it is possible to refactor your libavformat source code to enable av_open_input_stream function call with the null fmt argument in the same way as it is possible calling av_open_input_file function performing various input files probing? In my opinion this addition could be useful since ByteIOContext may be opened from the memory buffer using url_open_buf function to get AVFormatContext then.
2. If I undersand properly that swscale currently can't accept pal8 as destination pixel format? I created destination picture using avcodec_alloc_frame and avpicture_alloc functions, but got an exception trying to convert using sws_scale function. 3. Could you clear how to extract a frame at the specified timestamp position? I'm seeking to the key frame using function av_seek_frame with AVSEEK_FLAG_ANY | AVSEEK_FLAG_BACKWARD flags, then avcodec_flush_buffers function is called to flush buffers. However, after reading packet using av_read_frame function I found avcodec_decode_video returns argument got_picture = 0 and I need to read several following packets to obtain video frame finally. 4. Could you add some exportable function to detect bits per sample for a given pixel format? As I understand it is impossible currently since field bits_per_coded_sample in AVCodecContext structure is not filled by some decoders. Best regards, Mike Gratsas _____________________________________________________________ Are you a Techie? Get Your Free Tech Email Address Now! Visit http://www.TechEmail.com _______________________________________________ libav-user mailing list [email protected] https://lists.mplayerhq.hu/mailman/listinfo/libav-user
