About a Video Enhance AI snapshot function

Hi everybody,

I wonder why there is not a Video Enhance AI snapshot function to let us quicky export one specific frame as JPG or TIF.

Actually, the only way to proceed is to set the first frame to render to the snapshot position we want and add 1 to the last frame. This way it renders two frames.

When I try to render only one frame, it does not work this way and it renders the whole footage.

Snapshots help us a lot to see picture quality, noise, bad colour rendering and so on.

Kind regards.



I agree with this. Sometimes I just want to get a frame from a video to produce a cool looking still. Enhancing with the current model and settings and just save a snapshot would help a lot with this.

I created a script on my Mac using Automator. It allows to right click on a set of files and pick which frame you want to output (it asks time index). I create one file per snapshot so I can compare. you may have to change the path of ffmpeg. Hope this helps. Dom.

on run {input, parameters}
set theTime to the text returned of (display dialog “Enter time in mm:ss” default answer “”)
if theTime ≠ “” then
repeat with f in input
set videofile to POSIX path of f
set outputfile to POSIX path of (videofile & " " & theTime & “-%02d.png”)
set res to do shell script "/opt/homebrew/bin/ffmpeg -ss " & theTime & " -i " & quoted form of videofile & " -frames 1 " & quoted form of outputfile
end repeat
end if
return input
end run