Aspect ratio of video sliced with FFmpeg unintentionally changes

Question

I am trying to extract a small part from a larger video without changing the codecs, resolution or something else. Therefore, I use FFmpeg with the following settings.

ffmpeg -vcodec copy -acodec copy -ss 01:00:00 -t 00:00:30 -metadata title="Custom title" -metadata artist="Artist Name" -metadata language="English" -i input_file_h264.mp4 output_file.mp4

I am processing the video on Ubuntu 10.10. with FFmpeg version 0.6-4:0.6-2ubuntu6.2.

The output file comes with the same codec information as the input file. Though, it does open up in VLC and other media player as if the resolution of the video changed. The input source comes with a resolution of 720x576. The codec used is H264 - MPEG-4 AVC (part 10) (avc1). When the output video opens up it appears wider (16:9) although the video settings did not change.
I could not find anything like aspect ratio misconfigured in VLC also.

Do you know of a problem with H264 on Ubuntu? As far as I understand libx264 is used. But is that relevant when I simple slice a video?

• Here is the input video file (876MB).
• Here is the FFmpeg output.
• Here is the output video file (1.4MB).

---

As suggested by slhck I changed the order of arguments.

ffmpeg -ss 01:00:00 -i 28c3-4906-en-lightning_talks_day_3_pecha_kucha_h264.mp4 -vcodec copy -acodec copy -t 00:00:30 output_file_slhck.mp4

• Here is the corresponding FFmpeg output.

Answer 1

You're not really copying the video stream.

Supply the acodec and vcodec options after the i option. Just think about how you would process them. First read the file, then decide how to encode, then output. If you don't do this, the options will be parsed differently¹, which would explain the difference in input and output according to your log:

Input:  Stream #0.0(und): Video: h264, yuv420p, 720x576 [PAR 64:45 DAR 16:9]
Output: Stream #0.0(und): Video: libx264, yuv420p, 720x576 [PAR 89:44 DAR 445:176]

Note the difference in Pixel Aspect Ratio (PAR)? Currently, you're trying to decode the video with the copy codec, which obviously does not exist. Placement of options in FFmpeg really matters sometimes.²

This is also what the manual says:

Select an encoder (when used before an output file) or a decoder (when used before an input file) for one or more streams

So, to summarize — metadata left out for readability:

ffmpeg -ss 01:00:00 -i input_file_h264.mp4 -vcodec copy -acodec copy -t 00:06:00 output_file.mp4

---

_{1 – With newer versions of FFmpeg, the command would have failed because a copy decoder does not exist.}

_{2 – Also, place the t option after the encoding process, and leave ss at the beginning. Putting ss at the beginning ensures that the file is read up until the time, and then begins encoding. If you were to place ss after i, that would encode the file from the beginning on, which could take a long time.}