On Fri, Sep 14, 2018 at 9:29 AM astian <[email protected]> wrote: > > > Well, I'm not sure what you mean. Pad doesn't seem very dynamic to me, I > know that it has an eval=frame option, but according to the manual there > is no frame-number variable that I could exploit. >
Pad the video beforehand, usually with a transparent colour, then use zoompan. Since your new frame is much larger, you can simulate a zoom-out of the image, since with the padding, your zoom value doesn't need to drop below 1. However, you can't output these frames of differing sizes, as ffmpeg will resize them to a fixed size prior to encoding. However, you can use them in overlay, where the transparent padding gets hidden and you get the effective result. Gyan _______________________________________________ ffmpeg-user mailing list [email protected] http://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email [email protected] with subject "unsubscribe".
