I need to compress some videos about laboratory experiments which contain motions of some objects.

Since in analysis time intervalls on the video are needed as accurate as the camera captured them, it is important that the conversation doesn't introduce any time stretching, shifting or any other inaccuraries.

I am using ffmpeg (or avconv) for conversion and vp8 or h264 as video codec. Audo is irrelevant in my case.

Since I am not a video expert, my first question is, if there are any problems regarding time accuracy to expect at all.

If so, my second question is, what I should have in mind to bypass those problems when compressing a video.

To make it more concrete, my first attempts where:

avconv -i input{.MP4,webm}


ffmpeg -vcodec h264 -i input{.MP4,mp4}

  • As timing is crucial in your experiment, I suggest that you keep the initial frame rate of the captured source video. Then I would suggest to use a video codec that only uses Intra frames (no prediction). You don't mention which is the original codec/format of your video.
    – audionuma
    Jul 27, 2015 at 9:09
  • The original codec was h264 (mp4).
    – student
    Jul 27, 2015 at 11:25
  • I wouldn't try to transcode a video that is already in a lossy format as your original source is. Are you having a file size issue that requires transcoding ?
    – audionuma
    Jul 28, 2015 at 9:15
  • @audionuma Yes, it's about filesize.
    – student
    Jul 28, 2015 at 10:29


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.