- Mar 1, 2020
- 2,714
- 3,983
After review them again you are half correct. 33 & 34 I gotten from you after looked back in the thread. 35 & 36 I gather from this user:
Collection Video Shirakami collection [2022-12-06] [stukove/StukoveMMD]
- Jan 19, 2019
- 490
- 447
no offense but you're blind as fuck if you can't tell the difference
There's no good purpose, same with upscaling video or encoding to a new codec. You will never improve upon the original quality (barring AI which still is in its infancy). It's just a basic misunderstanding of video encoding at play. The only video which needs "restoration" are interlaced footage or video under 12-15fps.
- Jul 4, 2020
- 203
- 533
And you are misunderstanding the purpose of video processing and video encoding. It's never about improving "quality" (narrow sense).
It's just making videos more comfortable to watch or reducing file size, sometimes both.
Of course there will be some quality loss and effects are vary, but it depends, on the encoder's skill.
Possibly, but these objectives are not achieved when one interpolates 30fps to 60 or upscales 1080p to 2160p. The source is already comfortable to view and the size reduction could be achieved without introducing the many artifacts inherent in the guesswork of said operations.
When people see studios release remastered versions of old films in 4K they think they can achieve similar results. They don't understand the source is recorded on 35mm film and is roughly equivalent to 4K but was extracted from the source at a lower resolution due to the distribution constraints at the time. Now, the source 35mm gets resampled at a higher resolution - but the data is not being fabricated by an upscaling algorithm. The rest of what they do involves SFX and color grading. No video professional is upscaling using bilinear, bicubic or lanzcos and neither should we.
If you want a version with more appeal than fidelity, with a tiny file size: export as H265, medium, tune animation, CRF=26. This will take the file size down dramatically and people who don't care about original quality can rejoice.
- Jul 4, 2020
- 203
- 533
No actually, there are a bunch of "upscaled 4k" uhd blu-rays (which are upscaled from 2K DI) and if you look closely at them you'll see some characteristics of bilinear, bicubic or lanczos upscaling.
But I agree with your first sentense.
However, I would say interpolated 60fps is more comfortable to watch *if* the algorithm are good enough.
I guess my response would be that I was narrowly tailoring it to the application of remastering or improving upon a source. With regards to your example, they don't use such a workflow to improve the quality (remastering); nor are their sources low to medium bitrate YUV 4:2:0 when performing upscaling, which creates more gaps to fill in with artificial data. The 2k processing to 4k distribution for digital is simply based on time & budget constraints.
- Jan 19, 2019
- 490
- 447
I can see the difference in interpolated films. And it looks like crap. I can't say I've noticed much of a difference in the videos I downloaded from this thread. Maybe I accidentally downloaded the same video twice, I dunno.
- Oct 11, 2020
- 272
- 634
You must be registered to see the links
FINALLY HAPPENED!
- Jan 19, 2019
- 490
- 447
this video doesn't apply to repetitive motion (aka sex) w/o special effects to mitigate low framerate (like in between frames)
CGI porn doesn't have quality of pixar/disney or Japanese studio animation
there is no reason for them not to use 60 FPS outside of spending twice the time to draw/animate/render/etc
Bulging Senpai always does 60 and in Blender, so it's not as if it can't be done
CGI porn doesn't have quality of pixar/disney or Japanese studio animation
there is no reason for them not to use 60 FPS outside of spending twice the time to draw/animate/render/etc
Bulging Senpai always does 60 and in Blender, so it's not as if it can't be done
Last edited:
- Jan 19, 2019
- 490
- 447
Yeah, and if the original render is in 60 FPS, great! If not, that's great too. My point is I don't want to bother with interpolated stuff. I don't think it's a worthy improvement. That's what the video says, that interpolation tends to ruin a lot of stuff. Not that 60 FPS is bad. Or at least that's the point of the video I'm trying to reference. I remember people in the film industry were complain about smart TVs interpolating by default too
- Jul 4, 2020
- 203
- 533
Did you realize that the video you mentioned is takling about hand-drawn animations?
And the guy in the video does said that the "AI" algorithms are amazing for real-life videos (and I think it's not wrong to say amazing for these kind of "3d rendered piston motion videos" as well).
- Oct 11, 2020
- 272
- 634
Fuck I started an argument about tit size in one thread and interpolation in another.
- Nov 22, 2020
- 72
- 134
I find it funny you decided to link a yt video which, like other guy mentioned, doesn't apply much to these kinds of video, instead of just taking one of those interpolated videos and getting some screenshots to show us how bad the interpolation makes them look.
- Jan 19, 2019
- 490
- 447
I just happened to watch that video and remembered about this thread. I didn't specifically look for it. Didn't even plan to reply anymore until I saw it.