Hi all I've been running some more tests using mjpegtools 1.6.2 and have a couple of questions regarding some odd effects I've seen when the result was played on a PAL DVD player.
Firstly the process. We have a (non-commercial) NTSC DVD which has been editted down using cinelerra 1.1.9. mpeg3toc was run on the DVD and the resulting file loaded into cinelerra and the required edits performed. The complete finished product was rendered to an mpeg1 layer 2 audio file at 320kbps (file.mp2) and a Quicktime/DV video file (file.mov). It would be nice to render to Quicktime/YUV from cinelerra to a fifo and have lav2yuv read that to eliminate the DV code/decode cycle but cinelerra doesn't like opening a pipe for output (possibly due to a need for seeking, but I haven't looked into this yet) and I don't have the disk space to store the YUV file as an intermediate. The video was converted to mpeg2 using lav2yuv file.mov | mpeg2enc -f 8 -q 6 -E -10 -s -c -R 2 -o file.m2v and the final multiplex program stream put together with mplex -V -f 8 file.mp2 file.m2v -o file.mpg This was then mastered using dvdauthor 0.6.9, mkisofs 2.0.3 and burned using dvd+rw-tools 5.19-1.4.9.7. Now to the issues. Note that we're playing back on a PAL DVD player, so what's being produced by the player is PAL-60 from the NTSC source. The first thing I noticed is that on my hardware player, doing still advance on this NTSC DVD advanced the video by exactly one frame. This is strange only because an earlier DVD I made with the exact same mpeg2enc options would advance by three frames - something I put down to the "-R 2" option at the time. However, this time around we get individual frames. The only difference I can see is that the earlier DVD was PAL whereas this one is NTSC. Note that the original DVD advances by 3 frames (or there abouts) so it's unlikely to be due to the NTSC format per se. Another thing I noticed about this DVD is that motion is sometimes noticeably jerky on hardware players. It's most noticable when the scene pans to follow a speaker (it's a conference DVD): the background doesn't flow smoothly but appears to jump. I think the effect happens with all movement though - it's just most noticeable when there's lots of it. We're not taking about excessive movement here - just someone walking in front of a background which was perhaps 10m behind. To my eye the jurky motion seems far more prevelant on hardware players - it is not really present when ogle or mplayer is used to play the stream. Perhaps we're dealing with another hardware player funny. I noticed that the player was reporting very high bitrates for most of the time - usually it's above 7.4. This might just be a maximum in a given time though so perhaps that's fine. However, at times it jumps to 8.0 and above - I've seen 8.4 for example. I found this odd due to the specification of "-f 8" to mpeg2enc and mplex. That should keep the video bitrate below 7500kbps and the audio is only 320k - the total here is therefore under 7900kbps. Either there's a lot of overhead bits I don't know about or there's something funny going on. Note that we're not joining streams or anything like that - the input to dvdauthor is a single mpeg file. Some time ago Andrew commented that he'd found some issues with motion estimation but as I understand it, that only applied when dual-prime was active. In the case above, dual prime isn't used due to "-R 2" (disabling dual-prime is the reason why I specify "-R 2"). I also thought of field order but this doesn't appear to be the issue. Individual fields extracted from the quicktime/DV stream are definitely in the correct order - there's no backwards motion occuring. I guess there might still be an issue with mpeg2enc but I think that's unlikely from what I've read. Any thoughts on either of both of these observations? Regards jonathan -- * Jonathan Woithe [EMAIL PROTECTED] * * http://www.physics.adelaide.edu.au/~jwoithe * ***-----------------------------------------------------------------------*** ** "Time is an illusion; lunchtime doubly so" ** * "...you wouldn't recognize a subtle plan if it painted itself purple and * * danced naked on a harpsichord singing 'subtle plans are here again'" * ------------------------------------------------------- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click _______________________________________________ Mjpeg-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/mjpeg-users
