pbsds / hatenatools Goto Github PK

In my testing (PPM.py), I found that because of the way ffmpeg is used with the -shortest flag when combining the BGM, if the BGM is shorter than the flipnote, then the video is cut off when the BGM ends, rather than when the flipnote ends. If the BGM is longer or the same length, then this works fine.

Simply removing the -shortest flag creates the opposite problem, where it works fine if the BGM is shorter or the same length as the flipnote, but if the BGM track is longer than the flipnote, then the video doesn't get appropriately trimmed and the end becomes somewhat glitched.

I think the ffmpeg encode duration for this step needs to be locked to the duration of the flipnote, rather than dependent on the length of the BGM.

It would be really nice to have a Py3 version of these files!

Right now, I was trying to use 2to3 on UGO.py but I always end up on this issue:
TypeError: sequence item 2: expected str instance, bytes found

Guess reverse engineering work doesn't like being automatically translated.

I did try to fix these, but with new issues appearing everytime, I just gave up.

my compile.bat not working

This isn't really an issue, but I modified PPM.py to allow for frames to be upscaled 6x from their native resolution of 256x192 to 1536x1152, which is triggered by command line flags. It's a bit slow due to multiple nested for loops, processing a little over 1 frame per second, but it allows exporting flipnotes that can be viewed much larger than the original size without excessive blurriness.
Is this worth adding to the project?

native frame:

native frame zoomed to fit 1080p screen:

upscaled frame: