Best Hybrid settings for encoding broadcast vhs captures

[waxstone]

I've captured lots of raw Japanese broadcast tapes and have found a simple but effective tool called Brovicon, which converts the video into (mainly) everything I need. However it's still a bit buggy, and because it's a GUI for avisynth it's quite slow when compared to other tools out there. This is the general setup:


[Attachment 86010 - Click to enlarge]

I've decided to use Hybrid because it can do batch jobs at a decent encoding speed, but I'm getting very stuck with getting the correct PAR and SAR

This is what I'm trying to achieve, similar to Brovicon:
Noise reduction
Crop the dud regions and dot crawl
Add Borders to maintain 720x576
29.97fps -> 59.94fps, QTGMC deinterlace "fast" TFF
720 x 480p -> 1440 x 1080p (whatever scaling method works best for YouTube)
h264 at -crf 18, aac at 320kbps, .mp4 container

Please see sample video attached.

I'd really appreciate if anyone would be able to recommend the correct settings in Hybrid based on this video input.

[Selur]

Source has no PAR signaling, setting Input PAR to 8:9 (MPEG NTSC 4:3) seems fine, I would stick with 1:1 for output PAR when uploading to YouTube.
For deinterlacing: Are you sure the source is interlaced and not telecined?
From the sample I would have suspected telecine or another pull-down. So I would have used TIVTC not QTGMC.
For filtering I would start with SpotLess and add some other denoisers, some deblocking, dehaloing,..


Playing around, I attached and exampl of conventional filtering (script: https://pastebin.com/n7qLw6vG), probably nowhere near compared to what one could do with enough effort, scene-by-scene filtering, proper color adjustments,... encoding such stuff in batch is usually a dumb idea, since the filtering should be adjusted per source. (especially when aiming for 'best' or at least 'decent')

Cu Selur