video.io | video conferencing from webrtc-experiments | Video Utils library

The solution is passing inputdict={'-framerate':str(FRAMES_PER_SECOND)} to FFmpegWriter instead of outputdict with -r.

Neither inputdict nor outputdict are documented in skvideo (no mention of ffmpeg documentation, no examples), nor is -framerate properly documented in the ffmpeg documentation (it speaks about grabbing rather than saving, and claims that the default is 30000/1001 even though it is 25), nor is -framerate even mentioned when calling ffmpeg --help.

Shout out to @Goury for not deleting this thread (which helped here a lot) despite having received a negative score of -3 from the StackOverflow community in that thread.

You need to add the year date (year, month, day) to datetime, else this will default to 1 January 1900.

What you do is this:

Weirdly, it does work if I read in the video with video capture and then iterate over the frames.

So one solution to above is

Using processes I reduced time only by 15-20 seconds because ffmpeg even in single process was using almost full CPU power and my computer didn't have power to run other processes faster.

First I reduced code to make it shorter.

Code in try and except had similar elements so I moved them ouside try/except.

Next I used

Getting the same hash requires when writing and reading a video file requires careful attention.

Before comparing the hash, try to look at the video first.

Executing your code gave me the following output (first frame of video_2):

When the input (first frame of video) is:

I suggest the following modifications:

• Use AVI container (instead of MKV) for storing test_2 video in raw video format.
  AVI video container is originally designed for storing raw video.
  There could be a way for storing raw, or lossless RGB video in MKV container, but I am not aware of such option.
• Set the input pixel format of test_2 video.
  Add an argument: pixel_format='rgb24'.
  Note: I modified it to pixel_format='bgr24', because AVI supports bgr24 and not rgb24.
• Select video a lossless codec for test_2 video.
  You may select vcodec='rawvideo' (rawvideo codec is supported by AVI but not supported by MKV).

Note:
For getting equal hash, you need to look for lossless video codec that supports rgb24 (or bgr24) pixel format.
Most of the lossless codecs, converts the pixel format from RGB to YUV.
The RGB to YUV conversion has rounding errors that prevents equal hash.
(I suppose there are ways to get around it, but it's a bit complicated).

Here is your complete code with few modifications:

The latest version in the documentation (the one you are looking at) corresponds to a dev version 2.x, which is not released to PyPI yet. The version you have installed through pip is most likely 1.0.3, which is the latest on PyPI, and it doesn't allow an encoding parameter.

From the PR where the feature was introduced, you can see that it's only been tagged for release in 2.x versions.

Copying only that file to your source code will most likely not work, because it will depend on changes that happened in between the two versions. However, if you feel adventurous, you can install the dev version of the package, by following the Method by hand section in moviepy's docs.

opencv itself cant deal with auto. it is called OPEN Computer Vision not open audio.

If this is just for some actual conversion use, i would suggest

ffmpeg -i input.mp4 -vf scale=640x360,setdar=16:9 -c:v libx264 \ -preset veryslow -profile:v main -crf 18 -c:a copy output.mp4

change the input.mpr to your video and output as any name

reference. https://superuser.com/questions/837032/ffmpeg-resize-video-but-maintain-audio-quality

If you really need opencv. then it might be hard you can follow this Audio output with video processing with opencv

Hmm...I'd merge this answer and the actual implementation of ffmpeg_extract_subclip(https://zulko.github.io/moviepy/_modules/moviepy/video/io/ffmpeg_tools.html#ffmpeg_extract_subclip):

I figured it out thanks to some hidden comment in the official scikit-video repo. Apparently you have to set the path and then import the package again: