Game:

Spore (2008)

Duration: 7:03

187 views

5

I used ffmpeg scene detection with a threshold of 0.001 to keep only the parts that had actual movement going on in the footage I captured. There was a lot of pauses to look at reference material and such that you no longer have to bear witness to. Still turned out to be quite long of a video. Hopefully it's a little more palatable though. :/

ChatGPT helped me compose a python script for it that makes use of ffmpeg-python, as I really don't know what I'm doing:
```
import ffmpeg

input_file = 'input.mp4' # Replace with your input file
output_file = 'output.mp4' # Replace with your desired output file

# Set the scene change detection threshold to a lower value for higher sensitivity
scene_threshold = 0.001 # Lower value for higher sensitivity

# Run ffmpeg with the select filter to detect scenes and keep the transitions
(
ffmpeg
.input(input_file)
.filter('select', f'gt(scene,{scene_threshold})')
.filter('setpts', 'N/FRAME_RATE/TB')
.output(output_file, vcodec='libx264', acodec='copy') # Adjust codecs if needed
.run(overwrite_output=True)
)

print(f'Scene transitions have been extracted to {output_file}')
```

Other Statistics