August 2007 - ffmpeg-user - lists.ffmpeg.org

File size bigger??
by lukemessenger＠gmail.com 31 Aug '07

31 Aug '07

Hello, When I took a video that had smaller frames and increased the frame size the file size got bigger... I kind of expected that so I thought nothing of it. This time I took a video and decreased it's size to 220x160. A 9MB .mov with 320x240 resolution went to 12 MB flv with 220x160 resolution, and that's with qscale 3.5. At qscale 0.1, trying to keep a good quality, it went up to 24 MB. Here is a command line I used: ffmpeg -i /tgo_special_edition_another_perfect_season_trailer.mov -f flv -ar 22050 -ab 32 -s 220x166 -qscale 3.5/tgo_special_edition_another_perfect_season_trailer.mov.flv Here is the original file: http://www.mxsponsor.com/tgo_special_edition_another_perfect_season_trailer… The flash video encoder, that came with flash 8, produced it at 8.3 MB with the Sorenson Spark at high quality, and 32 kbps on sound. The flash video encoder was also able to do it at 220x165, which ffmpeg gave me a "multiples of 2" error. -Luke

3 4

Encoding grey pattern to WMV causes green square artefacts in Windows Media Player
by vanista＠gmail.com 31 Aug '07

31 Aug '07

Uploaded files to : ftp://upload.mplayerhq.hu/MPlayer/incoming/wmv_green_squares_wmp Subject : Transcoding grey static video to WMV causes green square artefacts in Windows Media Player I transcoded a simple GIF animation to WMV. Part of the picture displayed as green square artefacts during playback in Windows Media Player. I've extracted and expanded this part of the picture. It features grey pixels with somewhat high frequencies (noise). This new samples easily produces the problem in WMP. The same output is decoded fine with libavcodec and all players using it. Using Windows Media Encoder on the same input doesn't produce the problem. I also figured that filtering out the high frequencies would fix the problem. For example, using GStreamer's postproc_lowpass5 filter with scope=2 quality=0. ffmpeg -i grey.avi -vcodec wmv2 ffout.wmv FFmpeg version SVN-r9376, Copyright (c) 2000-2007 Fabrice Bellard, et al. configuration: --prefix=/opt/sherdev/MPlayer/1.0rc1 --enable-gpl --disable-strip --disable-mmx libavutil version: 49.4.0 libavcodec version: 51.40.4 libavformat version: 51.12.1 built on Jun 20 2007 14:58:35, gcc: 3.4.6 Input #0, avi, from 'grey.avi': Duration: 00:00:02.0, start: 0.000000, bitrate: 93 kb/s Stream #0.0: Video: rawvideo, yuv420p, 96x80, 1.00 fps(r) Output #0, asf, to 'ffout.wmv': Stream #0.0: Video: wmv2, yuv420p, 96x80, q=2-31, 200 kb/s, 1.00 fps(c) Stream mapping: Stream #0.0 -> #0.0 Compiler did not align stack variables. Libavcodec has been miscompiled and may be very slow or crash. This is not a bug in libavcodec, but in the compiler. Do not report crashes to FFmpeg developers. Press [q] to stop encoding frame= 2 fps= 0 q=2.0 Lsize= 4kB time=2.0 bitrate= 14.5kbits/s video:2kB audio:0kB global headers:0kB muxing overhead 120.742092% Note, for some reason, the output by ffmpeg doesn't play as is in WMP, I had to remux it by fixing timestamps using GStreamer. I provided both outputs for analysis. In either case, the container isn't the issue, it realy is the encoded stream. Files : grey.avi : the input sample ffout.wmv : the output from ffmpeg (doesn't play in WMP) greyout.wmv : passthrough remuxed output (plays in WMP) wmp_ss.jpg : screen shot of Windows Media Player depicting the problem report.txt : this bug report The remuxed output is produced using this pipeline : gst-launch-0.10 \ filesrc location=ffout.wmv ! \ asfdemux ! \ asfmux force-duration=10000 input-video-buffer-duration=2000 ! \ filesink location=greyout.wmv

2 6

FLV vp6
by jshupert＠pps-inc.com 31 Aug '07

31 Aug '07

Friends, I am thinking that ffmpeg cannot encode to an flv with the VP6 codec- but rather only h263. Is that still true ? Does anyone know of a means to cmdline encode to flv with the vp6 codec? Does anyone have any idea of when VP6 codec will be supported by ffmpeg? thnaks j

3 5

mpeg4 regulating bitrate.
by rcortina＠orb.com 31 Aug '07

31 Aug '07

I am encoding a file using my own application developed using ffmpeg. The format is 3gp with mpeg4 and aac. The encoded video frames vary quite a bit. Even though I am requesting ~320kpbs, I get a stream that fluctuates between ~170 to ~800kbps. This is because of very large video frames. This is unacceptable since I want to stream this file to a cell phone. What can I do to make sure I stay under a resonable boundary? Rafael Cortina

3 2

Decoding problems
by tchristensen＠nordija.com 31 Aug '07

31 Aug '07

Hi I'm building a transcoder that receives an mpeg2 video element stream and decodes it for later encoding in a different codec. The part I'm having trouble with is currently the decoding. The logic is Read input While more input available Decode buffer as frame If complete frame Encode frame Stream encoded frame Read input In the following example I'm just counting the frames I was able to decode. I know that I could use avformat specific file reading operations but in the end the stream is received over sockets. I'm getting to few frames decoded and see lots of errors. Any hints? Thomas Log: --starting transcode 622080 bytes read [mpeg2video @ 0x231a250]ac-tex damaged at 6 21 [mpeg2video @ 0x231a250]Warning MVs not available [mpeg2video @ 0x231a250]concealing 660 DC, 660 AC, 660 MV errors 622080 bytes read [mpeg2video @ 0x231a250]ac-tex damaged at 31 16 [mpeg2video @ 0x231a250]Warning MVs not available [mpeg2video @ 0x231a250]concealing 880 DC, 880 AC, 880 MV errors 622080 bytes read ... Source code: ....initialisation.... avcodec_register_all(); inputCodec = avcodec_find_decoder(CODEC_ID_MPEG2VIDEO); inputContext = avcodec_alloc_context(); inputContext->width=720; inputContext->height=576; if(avcodec_open(inputContext, inputCodec)<0) throw "Input codec failed to open"; // Init output outputCodec = avcodec_find_encoder(CODEC_ID_MPEG1VIDEO); outputContext = avcodec_alloc_context(); outputContext->bit_rate = 400000; outputContext->width=720; outputContext->height=576; outputContext->time_base= (AVRational){1,25}; outputContext->gop_size = 10; /* emit one intra frame every ten frames */ outputContext->max_b_frames=1; outputContext->pix_fmt = PIX_FMT_YUV420P; if(avcodec_open(outputContext, outputCodec)<0) throw "Output codec failed to open"; inputFrame = avcodec_alloc_frame(); outputFrame = avcodec_alloc_frame(); inputBufferSize=avpicture_get_size(PIX_FMT_YUV420P, inputContext- >width, inputContext->height); inputBuffer=new uint8_t[inputBufferSize]; outputBufferSize=avpicture_get_size(PIX_FMT_YUV420P, outputContext- >width, outputContext->height)*8; outputBuffer=new uint8_t[outputBufferSize]; .....transcoding loop.... // read input int got_frame=0; framesDecoded=0; int dec_bytes=0; int bytesRead = fread(inputBuffer,sizeof(uint8_t), inputBufferSize,inputFile); std::clog << bytesRead << " bytes read" << std::endl; // while has more input while(bytesRead>0) { // decode input into mpeg2 frame // Decoding dec_bytes = avcodec_decode_video(inputContext, inputFrame, &got_frame, inputBuffer, bytesRead); if(got_frame) //decoded one complete frame { framesDecoded++; // if complete frame // encode frame into mpeg1 frame // save frame } // read input bytesRead = fread(inputBuffer,sizeof(uint8_t), inputBufferSize,inputFile); std::clog << bytesRead << " bytes read" << std::endl; }

2 2

How can I know an h264 media's default level and profile encoded
by tandangdang＠gmail.com 31 Aug '07

31 Aug '07

thanks -- "Thank you, Lord, for making all necessary things simple, and all complicated things unnecessary"

1 0

How can I know an h264 media's default level and profile encoded by FFmpeg?
by tandangdang＠gmail.com 31 Aug '07

31 Aug '07

thanks -- "Thank you, Lord, for making all necessary things simple, and all complicated things unnecessary"

1 0

help compiling
by jesus＠alvento.com 30 Aug '07

30 Aug '07

Hi, I'm compiling ffmpeg on RHEL5 with faad support. Hi have installed faad2 and faad2-devel rpm's (2.5-2). I have this error: [root at lerez ffmpeg]# ./configure --enable-libmp3lame --enable-libvorbis --disable-mmx --enable-libogg --enable-libamr-nb --enable-libamr-wb --enable-libfaad --enable-libfaac --enable-gpl --enable-libxvid --enable-libx264 ERROR: libfaad not found If you think configure made a mistake, make sure you are using the latest version from SVN. If the latest version fails, report the problem to the ffmpeg-devel at mplayerhq.hu mailing list or IRC #ffmpeg on irc.freenode.net. Include the log file "config.err" produced by configure as this will help solving the problem. At the end of config.err: check_lib2 faad.h faacDecOpen -lfaad check_func2 faad.h faacDecOpen check_ld check_cc BEGIN /tmp/ffmpeg-conf-8886-22902-27900.c 1 2 #include <faad.h> 3 int main(){ 4 (void) faacDecOpen; 5 return 0; 6 } END /tmp/ffmpeg-conf-8886-22902-27900.c gcc -fomit-frame-pointer -c -o /tmp/ffmpeg-conf-17328-22902-3562.o /tmp/ffmpeg-conf-8886-22902-27900.c /tmp/ffmpeg-conf-8886-22902-27900.c: In function ?main?: /tmp/ffmpeg-conf-8886-22902-27900.c:4: error: ?faacDecOpen? undeclared (first use in this function) /tmp/ffmpeg-conf-8886-22902-27900.c:4: error: (Each undeclared identifier is reported only once /tmp/ffmpeg-conf-8886-22902-27900.c:4: error: for each function it appears in.) ERROR: libfaad not found And I have these librarys: /usr/include/faad.h /usr/include/neaacdec.h /usr/lib/libfaad.so Where is the problem? I think I have done all fine, but something is wrong. If I compile without faad2 all works fine. Thanks. Jesus

1 0

yamaha MIDRADIO player can not play the mmf file encoded by ffmpeg?
by angelsky2005＠gmail.com 30 Aug '07

30 Aug '07

Dear Sir: I use ffmpeg to encode mmf file from mp3. The conversion is ok. But I want to use yamaha midradio player to play it , the yamaha player report unsupport_file_format error. Would you please tell me what's wrong with it? Thank u all. I use ffmpeg at winxp sp2 (mingw + gcc ) The ffmpeg output is below: ffmpeg -i input.MP3 -acodec adpcm_yamaha -f mmf -ar 8000 -ab 32k -ac 1 -y out.mmf FFmpeg version SVN-rUNKNOWN, Copyright (c) 2000-2007 Fabrice Bellard, et al. configuration: --enable-gpl --disable-shared --enable-static --enable-libmp3la me --enable-libamr-nb --enable-libamr-wb --enable-libogg --enable-libvorbis --en able-libxvid --enable-liba52 --enable-pp --enable-libfaad --enable-libfaac --ena ble-libx264 --enable-avisynth --disable-ffserver --disable-ffplay --enable-memal ign-hack --disable-debug --disable-vhook --disable-network --enable-small --extr a-cflags=-I/usr/local/include --extra-ldflags=-L/usr/local/lib --disable-ipv6 -- extra-ldflags=-static libavutil version: 49.5.0 libavcodec version: 51.40.4 libavformat version: 51.12.2 built on Aug 24 2007 10:29:42, gcc: 4.3.0 20061204 (experimental) Input #0, mp3, from 'input.MP3': Duration: 00:06:49.9, start: 0.000000, bitrate: 128 kb/s Stream #0.0: Audio: mp3, 44100 Hz, stereo, 128 kb/s Output #0, mmf, to 'out.mmf': Stream #0.0: Audio: adpcm_yamaha, 8000 Hz, mono, 32 kb/s Stream mapping: Stream #0.0 -> #0.0 Press [q] to stop encoding size= 1458kB time=373.1 bitrate= 32.0kbits/s video:0kB audio:1458kB global headers:0kB muxing overhead 0.005427% Best regards

1 0

Error concealment in ffmpeg
by xiaolin.cheng＠gmail.com 29 Aug '07

29 Aug '07

Hello all, I used ffmpeg to encode a raw yuv file into a mpeg4 file and stream it over network. However some frames got lost during the transmission. So the number of frames in the reconstructed yuv file at the receiver end is less than the original yuv file, which makes it difficult to compute the PSNR values (the frame number is different). I am wondering if there is any option which enables ffmpeg to produce the same number of frames in the yuv file after network transmission. Thank you all.

1 0