Hello,

and thanks for your answers so far. The truth is that I create the final video in two phases: One is to extract good takes from raw footage, which is in MJPEG (and adjusting brightness etc) and create intermediate footage, rendering with Cinelerra. Then I feed the intermediate videos into the final cut. Since there are different technical setting for different parts of the footage, this is the only sane way.


The 21-minute "footage" I mentioned previously is the intermediate footage. It's created by rendering to pipe as YUV4MPEG, piping it to mencoder, which can turn it to whatever I want. But what format will really work?


Now to your questions:


Martin Ellison wrote:

> Can you chop it up into smaller clips?
>


Of course I can. This is what I eventually did. And with mencoder at hand with its -ss and -endpos flags, any video can be chopped. But it's an ugly solution.


Ichthyostega wrote:

Question: what is the format of your original footage? DV?
As mentioned above, the original is Cinelerras YUV4MPEG. Or pick your choice.
Question: have you tried using the original footage (without
transcoding) in a quicktime container?
Maybe after transcoding the audio to PCM? for example
something like
Since some of you seem to have good experience with DV, I went for it, and created a DV file using the following command:

mencoder - -audiofile audio.wav -oac pcm -ovc libdv -of lavf -vf scale=720:480,harddup -lavfopts format=mov:i_certify_that_my_video_stream_does_not_use_b_frames -o "transcoded.mov"

Two remarks:

(1) Since this is actually YUV4MPEG rendering, the audio is taken from audio.wav. This is the way I work ed previously (and all the time), but spared this detail in my previous post. (2) DV is not a generic solution, since the mencoder codec (and possibly any codec) requires the output to be 720x480 or 720x576.

The bottom line is that I got a 4.5GB file (21 minutes, remember?) which was playable with (Windows) Quicktime. With Cinelerra something went wrong, causing the audio channels to appear really weird (maybe because the audio channels were detected as big endian?).

So I ended up with a file which is very hard to handle, and which Cinelerra itself doesn't import well. This is sort-of back to square one.

---------------------------------

Now you may ask why I'm bugging you all, if I have a workaround. The answer is, that since mencoder almost always manages to read any crazy video you give it, and transcode it consistently and cleanly to anything else, I'm looking for one, single, high-quality format to feed Cinelerra with, and make all video oddities, A/V sync in particular, a thing of the past.

As long as the video is short, both MPEG-2 and MJPEG do well. But when the video gets longer?

   Eli


--
Web: http://www.billauer.co.il


_______________________________________________
Cinelerra mailing list
[email protected]
https://init.linpro.no/mailman/skolelinux.no/listinfo/cinelerra

Reply via email to