How to crop and merge .ogv videos?

Question

I am presently creating a video tutorial for a software tool. I have recorded a few short videos (between 30 seconds and 4 minutes) with RecordMyDesktop that I now need to merge into a single video. The task pipeline looks like the following:

1. Convert the files from ogv to avi (or another workable format).

2. Crop unwanted sections from the videos (such as the interaction with RecordMyDesktop itself).

3. Merge the various cropped videos into a single file.

Sounds simple, but I have lived a whole saga the past few weeks trying to do it without success. Below I detail the steps I have taken so far and the problems I found.

Converting ogv to avi

There are various threads here about this subject, pointing to multiple tools. Many of these threads seem to be outdated but I eventually succeeded using avconv, e.g.:

avconv -i "input.ogv" -vcodec mjpeg -acodec mp2 -qscale 10 "output.avi"

Previously I tried with ffmpeg, Avidemux, Lives and others, but all of them fail with errors or crash on Ubuntu 14.04.

With avconv there is a visible loss of image quality and a good deal of noise is added to the audio, but none of it is catastrophic.

Cropping

For this task the only tool I am aware of that allows cropping at the frame level is Avidemux. Others such as Lives also advertise this functionality but are not able to perform it.

I had no errors performing this task with Avidemux but the resulting videos started to yield problems. Totem is not able to play one of them, issuing an exception; OpenShot also issues a warning when opening some of these videos cropped with Avidemux.

Merging

So far I have tried this task with five different tools, all of which failed:

• OpenShot - filled up 4 Gb of RAM and went on to fill 4 Gb of swap, at which point the system became unresponsive, forcing a reboot. In total the cropped avi videos add up to little over 200 Mb.

• avimerge - produces a merged file without issuing warning or error messages but the image and sound are out of sync. There are moments where the sound played is from the parts cropped out in the previous step.

• Kino - also produces a merged file without issuing errors but completely ravishes the image, it constantly flickers and jumps up and down, making it impossible to understand.

• cat + mencoder - the end result has similar problems as with avimerge.

• Avidemux - eventually found out it is also able to merge files, although the process is not obvious. The resulting videos yield sound and image out of sync; when opening a merged file with a player such as Totem, the length is about the double of what is reported by Avidemux.

The question

At this stage I strongly suspect something is going wrong upstream of the merging step, perhaps with the cropping (and thus the detailed story).

My question comes down to: how to perform these three apparently simple tasks with videos produced with RecordMyDesktop? I would appreciate suggestions either for a completely new processing pipeline or to changes in one of the steps. Comments from experienced RecordMyDesktop users are also welcome.

Update I: included results of merge with cat and mencoder.

Update II: included results of merge with Avidemux.

Answer 1

Perhaps you are looking for different workflow.

You can first cut out the parts that you want with oggz-chop. Simply type in the terminal:

oggz-chop -o output.ogv -s start -e end input.ogv

This should give you good files that you want to keep. Make sure you name them in some sequential order, to be able to deal with them later.

After that:

There is an oggCat command, which is able to concatenate two or more ogg files together producing one ogg file of the length equal to the sum of all provided.

The command line that will help you is:

oggCat output.ogv infile1.ogv infile2.ogv infile3.ogv ...

Do not confuse it with another command from the same package titled oggJoin, which does something quite different, creating a single file with several streams that will run in parallel.

Now, for cropping, i know that there is a way to do that losslessly, by editing the header of Theora stream, but i was unable to figure out how to do that as of yet. I will update the answer if i'll find the way.