LordCrash wankz remasters are badly stretched for me. It feels lacking a lot and 0 immersion. Every body and object is very unnatural. Totally can't watch.
For some reason original videos were fine. No idea how did they manage to break it completely. At least for me.
Never understood you guys screaming for such remasters. Not the first time users are totally wrong.
doublevr think you are misinterpreting comments here coz i cant see anyone screaming for remasters
can't really make sense of many things people are screaming for. glad it works for some of you guys seeing magic where there is none while we are sticking to proof based stuff.
once again, the "gains" you see with upscales posted here are not necessarily improving perception. many times the image feel very unnatural with that "AI" crap
wankz remasters are badly stretched for me. It feels lacking a lot and 0 immersion. Every body and object is very unnatural. Totally can't watch.
They are bad compared to modern 6k/7k videos, but they are an improvement to their original videos.
That said there also sometimes quite huge differences how much they improve. I was pretty blown away how much better their first remaster was, I still have the original and the remastered file, and the difference is huge. (Gina Valentina and Bailey Brooke in Easter Bunnies).
The improvement of some newer scenes was quite a bit less, I am wondering if they cut back quality, to push out more remasters, as they said I think that it needs multiple days to upscale just one video.
I think the problem is more the expectation. Is an AI Upscaled video better then a video that is natively shot in the higher resolution? Of course not. Is it better then the original, well usually yes. If it is worth all the additional processing, well at least questionable, certainly for new things shooting with better cameras is always the superior solution. But for older things, where you either have the choice of a crappy video or a somewhat less crappy video most will prefer the latter.
ibins I was pretty blown away how much better their first remaster was
In the gearvr the remastered are not as noticeable, but since I got a quest 2 recently 
...
the remastered is very noticeable 
Just met Topaz creators. They confirmed its useless to upscale 4K to 8K. Also it's intended for flat content and significant work is required on their end to make it work with vr because of projection.
We arranged they will take one video and do their best to upscale the very center of the video to see if it is even worth doing any further research.
doublevr source: trust me bro
No visual improvements? Are the improvements worth it should be a better question. Without a doubt there are visual improvements. Again, the source material must have good exposure and noise/block levels. Nothing will be able to fix that. If those old really blocky scenes were being upscaled, I can see why they would not work. I will work on getting some samples up for people to compare on their own.
Here are a couple of screenshots of an upscale from 4k to 8k that I am working. Would love to hear back from other users to see if they notice any difference. First is the original 4k version, second is the upscaled 8k version, and the third is the upscaled 8k with a little color tweaking for my preference.
Original 4k
Upscaled to 8k (No color grade)
Upscaled to 8k (color grade)
I recall a discussion about this topic from some months ago.
AI upscaling for VR has potential. However, for there to be visible improvement one would have to improve the textual models generated specifically for VR -- from what I've been told, and what seems logical.
One would have to begin with the studio file or the original recording in highest resolution, so the film doesn't suffer from iterative downgrading (through compression from the various encoding formats, most of which are lossy, not lossless).
A company, like Topaz, would then have to re-train their AI working from images captured with a fisheye lens.
You could try this and use an in-house model (developed for VR). It's free.
https://towardsdatascience.com/deep-learning-based-super-resolution-with-opencv-4fd736678066
The difficulty would be training it for your needs. But you have a fantastic library to draw textures from!
You know what you could also do... Ask Meta. Of course, they are really creepy to deal with, and they say they HATE porn (because secretly they just LOVE it)... But if you claim to be a researcher you could probably get help and ideas...
With better algorithms and better screens there should be more of the difference. The only thing it is only relevant for old videos while we are looking bright into the future
April 2024 Update: we've been doing some further tests with upscaling recently. The tech has come quite a long way in the past year or so.
Here's a sample video on the SLR Labs channel with a few random snippets of some older SLRO scenes, both the original versions and the upscale samples. If you look up towards the very top of the screen you'll see whether what's currently playing is the "Original" or "Enhanced" version.
We would be interested in hearing any thoughts & feedback! I've noticed that the visual improvements are more obvious w/ the downloaded versions vs streaming (this is something we're also looking at for the "enhanced" versions).
Is it worth considering putting more time/effort into this effort for older SLR Originals releases?
justsomedude101 Why did you pixelate the image? The original doesn't look like that.
Hairsational Why did you pixelate the image? The original doesn't look like that.
It's the cover pic for the SLR Labs test scene. I assume it was used for dramatic effect.
Edit: removed the cover pic from the thread 
. Be sure to watch the actual scene!
justsomedude101 Putting it in a thread where you are asking for feedback will mislead people into thinking the improvement is much better than it really is. Just sayin'.
hello, didn't see any difference (deo streaming then download with heresphere)
I didn't see any difference either
