Joel Hruska said:
I will get you some samples made today. Do you have a Microsoft OneDrive account or do you prefer Google Drive?
Thank you, but please try to keep it SMALL. I live on the flattest country 😉
A person who paints paintings is not the same as the person who restores them, but you cannot restore art if you do not understand something of the artist.
No question. You have to know a lot and if you do your best, it’s not easier than to create something like a tv-series, where they also do their best.
The creativity in restoration comes in choosing which techniques to apply, and maybe in finding / inventing new techniques that handle something better than a previous solution did.
Yes, you are right, this is a creative, sometimes VERY creative process, too.
I am preparing some clips to show you final output of the method. I want to make sure I am clear about what I expect you to see:
1). Recovery of the same progressive frames as in the 23.976 fps Rio Grande version.
2). Interpolated frames that provide the smoothness required to make the 59.94 version move smoothly in places where the 23.976 fps version fails to recover ideal motion.
Ok, ok, thanks, but: This is a possible way to go, BUT:
If you go to 59.94fks, then:
- you get smoothest motion in the parts, that were 29,97fps originally
- you get somehow smooth motion but with slightly variing speed for all the parts that were pulldowned from 23.976 (originally 24)fps.
The parts with 2. are MUCH larger than 1.
So it’s the only way to go to do FIRST the best you can do for these large parts, and that means:
If you do this right, you get 100% smooth, progressive frames 1:1 of the original film material.
THEN you can think about what to do with the original 29.97 parts. You have to slow them down to 23.976
This is only possible with motion flow. If you decimate somehow you get big jerky junk. As I said above: We did this with Alchemist, that did a quite good job, not perfect though. In Avisynth there are also several ways, one is a larger script called “framereatconverter”, another, that is used by the first, is using MVTools.
There are also AI possibilities, f. e. Resolve, which makes use of GPU / CUDA - but I am not very good in this.
I tried it several months ago, but was not very pleased with the results.
Bringing everything to 120 (119.88)fps is not good, too. Ok, you can take the smooth, ready IVTCed 23.976 and “enlarge” it with a ratio 4:1, that is, repeat each frame 5 times. For the 29.97 portions, also smooth, repeat each frame 4 times. Everything 100% smooth, ok, BUT: The speed ratio of both won’t be correct, the latter ones (29.97) will play 4/5 slower than the others.
The 60 fps motion of Orinoco is smoother than the 119.88 fps conversions that I’ve tested (I tested 119.88 as a way to equalize the 23.976 fps and 29.97 fps content). The motion is less irregular. There are still absolutely interpolated frames being used to smooth out the presentation and I won’t claim there aren’t, but the subtle, hitching jerking that I spent so many months trying to repair has vanished. I’ve messed around with changing the frame rate from 23.976 to 29.97, 47.952, 59.94, even 71.9128 (a little) and 119.88 (a lot).
These experiments are not new to me. I did similar things some years ago, and I fear, everyone who has to work with these things had to do so once. I also can remember how frustrating this can be - and how many fallacies you make…
The encode method I call Orinoco is the best 59.94 conversion method I have found.
I believe you, and also that the method is intelligent! But the results cannot be “correct”… That is
-always have the same speed
Just because you can’t divide 60 through 24… 😉
I’d prefer not to use it. It takes a very long time to upscale 59.94 content. But it does solve problems in the places where Rio Grande leaves bad motion – and it does so without causing visible frame rate stutter. Are the interpolated frames still there? Absolutely. Do I want them there? No. I’m just willing to tolerate them if I have to.
As I said, you have to decimate for the pulldowned sections and then take care of the real 29.97 material in a different way.