I love working with Cinelerra, I really do, but sometimes I find myself having to make a few compromises when it comes to getting it to work seamlessly with newly recorded footage. Normally I would have to convert the original video to a format that Cinelerra can edit. From the various cameras I’ve worked with, the original footage has already been compressed, converting it to another format before editing it is unpleasant because it can degrade the quality further.
Canon 7D uses the .MOV container format to hold a h264 encoded video and a 2 channel pcm_s16le encoded audio track together. Works fine when you play it with VLC or run it through some Transcoding tools like Avidemux. With Cinelerra though, it’s somewhat problematic. Although you can import the video file just fine and run it through the timeline, Cinelerra is unable to play the audio portion of the file. Cinelerra is able to read h264 encoded videos separately just fine , likewise for pcm_s15le encoded audio but the problem lies with the container format. Run it through ffmpeg and you’ll get a message like this:
“Seems stream 0 codec frame rate differs from container frame rate: 50000.00 (50000/1) -> 25.00 (25/1)”
At the moment Cinelerra 4.1 only decides to work with the video stream and ignores the audio stream, normally in this case I would extract the audio portion of the file, import it separately into the project and manually attach it to the video file in the timeline but that’s far too tedious to make it worthwhile. Instead, repackaging the file into another container will do the trick. My choice is the Maktroska format, an open source container format similar to MOV, AVI or MP4 but much more flexible in terms of how much media content it can hold and what codecs it supports. You can read more about the format from this Wikipedia article. The important thing to note is that Cinelerra 4.1 does support it.
Changing the container takes only a few seconds with very negligible difference in the size of the original video file and the recently produced MKV file. Most importantly the audio and video content remains untouched. The quickest way of changing the container is through the command line using ffmpeg as the transcoding tool, this command will do the trick:
“ffmpeg -i mvi_8927.mov -vcodec copy -acodec copy mvi_8927.mkv”
Explanation: So here the “mvi_8947.mov” is used as the input file, the video codec ( vcodec ) is to set to copy the input instead of converting to another format, same goes for the audio codec ( acodec ) and finally “mvi_8947.mkv” is the file output with the Matroska video extension. Ffmpeg then goes ahead copies the audio and video portions into a Maktroska file which you can then import into Cinelerra 4.1 . Now we’re able to work with Canon 7D files with both audio and video portions made playable in Cinelerra.

Canon 7D footage repackaged into MKV format, now with audio and weird artefacts in the video preview
But it’s not without its oddities, version of Cinelerra 4.1 has some artefacting issues in the Compositor when the video is paused. Despite this, the artefacting never shows up when played or rendered in Cinelerra. It’s nothing to be too worried about.
Note: This tutorial won’t work with community version of Cinelerra, for the time being Cinelerra 2.1 ( 22nd February 2010 build version ) is unable to read MKV properly for me.
Thanks for sharing your experiences with this setup!
– Have you had any success in rendering the edited 7D footage back to H264 but with AAC sound?
/Johan (fellow 7D + Cinelerra user)
Not much I’m afraid, I’ve always have had rendering issues with most of the codecs provided by Cinelerra. I think your best bet is to render the film using Motion JPEG ( crank it up to highest quality ) instead. You’ll end up with some fairly large video files but you can transcode it to h.264 after using either ffmpeg or Avidemux.
At the moment I’m having a much better editing experience with the 7D footage with the latest version of Kdenlive. I’ll write a post about it later.
Hey! How’s that post about 7D + kdenlive going? I’ve just started with that combo and would love to read your experience 😉
KDEnlive works great with it actually, I now do the primary timeline editing with it, then I render it out as a high quality h264 video and import it into Cinelerra to colour adjustments and whatnot. The problem with the 7D video footage is that I haven’t managed to get KDEnlive to play the audio channel properly so I have to transcode the .mov files again to convert the audio channel to another format supported by the container ( in this case I just got lazy and stuck with mp3 ).
I have started work on a python script that converts .kdenlive projects into a basic Cinelerra project file but it’s proving to be quite tedious.
Hmmm… I’m not having big problems with the native 7D video footage when recorded at 720 (sound worked fine, as far as I can remember). With 1080 kdenlive simply suffers too much.
I’m having quite decent workflow ease transcoding native footage to dnxhd codec with ffmpeg. But not sure yet if I can get it better… Will keep trying.
Best regards!
Just realised that the version of Kdenlive stored on the official Ubuntu repository only had version 0.7.5. Updated to 7.7, sounds works fine now as does 1080p playback ^_^ . Version 7.8 will be released next week, looking forward to the colour correction feature http://www.kdenlive.org/users/granjow/introducing-color-scopes-histogram
Hi,
Thanks for your post, it helped me get cinellera working for my 500D movies. I played around and what you suggested didn’t work for me, however, I found a way to change the audio codec only (which was the one causing the problem):
ffmpeg -i MVI_6638.MOV -vcodec copy -acodec libmp3lame MVI_6638a.MOV