logo Sign In

nightstalkerpoet

User Group
Members
Join date
25-Oct-2010
Last activity
29-Jun-2025
Posts
1,108

Post History

Post
#1087374
Topic
STAR WARS: EP V &quot;REVISITED EDITION&quot;<strong>ADYWAN</strong> - <strong>12GB 1080p MP4 VERSION AVAILABLE NOW</strong>
Time

Does anyone else get excited when you see 005’s avatar under Star Wars Fanedits on the index page? It tells me 1.There’s probably something new in the Revisited thread and 2.It’s actually on topic.

.

When I see Frink’s, I immediately picture the ? Power Up Block from Mario Kart 64, complete with the sound it plays while choosing.

Post
#1086711
Topic
Return of the GOUT - Preservation &amp; Restoration (Released)
Time

Lol, didn’t catch you in time with my post edit:

Otherwise I’d at least resize to 768x326 instead of 720x326 to achieve the correct AR without needing to worry about PAR.

Pal 720x326 is still slightly anamorphic and stretches to 768x326. This is equal to a 2.35:1 AR. The equivalent at 720p is 1280x545. So nothing should get screwed up in the process 😃

Post
#1086708
Topic
Return of the GOUT - Preservation &amp; Restoration (Released)
Time

Slightly confused why you’re scaling to anamorphic PAL, or even downscaling to get to correct AR. I would just resize straight to 720p, which I believe is 1280x545 with add borders of 88 and 87. That should keep the increased resolution from EEDI2 better. Two Lanczos4 calls is going to reintroduce edge enhancement if I’m not mistaken.

Post
#1085677
Topic
Automated color grading and color matching with a Machine Learning Algorithm
Time

Is this limited to color, or could you change other attributes? Like running say 5 different shots through PFclean, then letting the learning algorithm try and degrain based on the results.

Or using it to combine different sources- say you have two different sources, but one has more picture on the left and right or top and bottom (Fellowship of the Rings theatrical vs extended)

You manually combine a variety of the frames, adding as much extra picture as possible without causing uneven letterbox borders. You train the algorithm to accept the 2 source images, and a reference image with them manually combined.