I’ve been using UniFab Video Enhancer recently, and it works pretty well for upscaling and enhancing videos. The AI features like upscaing, denoising, and HDR conversion are quite useful, and the interface is pretty straightforward. Might be worth a try.
What do you think of VIDEO ENHANCE STUDIO AI?
Difficult to know without any trial version to test. Don’t really want to drop $300 without testing it.
And that’s per year pricing.
Ya, throw in the HDR-version and you’re looking at $750 per year. But hey, you get Davinci Resolve included for a year.
Davinci Resolve is a one time payment, not subscription and is not included.
I don’t think they (tekno3d) are an authorized Davinic Resolve Studio version reseller. Beware.
I reached out to the creators of Tekno3D and got 48h free trial license for both “Tekno3D Labs Video Enhance Studio AI” and “HDR-X Studio”. I don’t have any Davinci Resolve installed. The features seem to be part of the Tekno3D Software (i honestly don’t know as I’ve only used Davinci Resolve a handful of times and it looks nothing like this, but perhaps some of its features are included in hdr-x studio).
I am in no way affiliated with Tekno3D. I simply reached out earlier today and asked if it would be possible to get a trial license due to its high entry price, referencing this thread.
If anyone’s got any sample clip they want me to try. Compress them to a zip or 7-zip folder, upload to google drive and send me a link. Explain what you’re trying to accomplish with the clip and I’ll what I can manage with this software. Preferably no longer than 10-15s, as it’s just to test the software out.
Video Enhance Studio AI
The Video Enhance Studio AI has got several categories of AI Models. Upscaling, post-processing, interpolation and restoration models.
18x General Upscaling/Enhancement models.
15x Animation Upscaling/Enhancement models.
11x Restoration models.
12x post-processing models.
All standard RIFE models for interpolation + in-house tuned rife models. Enhanced motion AI to better handle scene changes, dark backgrounds, and/or high motion scenes.
It’s also got several ways of personally tuning each and every model. With enhancements to certain parameters such as Face Enhancement, Resolution Fixer, Detail Enhancement, High Quality Grain, Advanced Sharpen HQ(or Fast), HQ Temporal Spatial Denoise, KNL Denoise, MCDegrainSharp, Tekno3D Labs Denoise High/Medium, HQ Deband, Deband+Grain.
As well as reduce/increase saturation, intensity, AI Hue Fix, Sharpering, Exposure and “Fix Green Tint in SDR Videos”.
And the list goes on. I’d be here writing all night if i were to write down all models and tweaks/features.
HDR-X Studio:
Here are 2 screenshots of some of the options in HDR-X Studio:
2 Likes
OK, there you go.
Ohne Titel.mp4.zip (7.0 MB)
Of course this video really is a hard one with it’s low quality, low res and high compression artifacts.
Just stumbled upon this webpage:
Hundreds of free AI-models
Which refers to this webpage with FAQ and a list of several maintained GUI’s for AI enhancement, mostly open-source and free (Topaz Gigapixel is listed here as well - as paid though of course):
FAQ and Softwares
Found it by chance on this webpage which also have links to models one can download, other than the above.
You should join us on the discord server Enhance Everything!
Are there any new findings with Video Enhance Studio AI?
Did you get to try my „problem video“?
I’m quite curious what other AI software can achieve with such low-res heavily compressed sources.
Especially regarding the balance between cleaning up - preserving/reconstructing details - maintaining fidelity/staying true to the original.
hola! je voudrai bien que tu essaie pour moi ces quelques secondes de vidéo, issue de casette 8 mm en sd, je compte améliorer mes vidéos en hd au final. merci beaucoup de proposer!
début descente du rhône.zip (29.1 MB)
Sure thing!
I actually sent the video to the creator of the software and asked him to enhance it. Which he did, but he did for sure not put much time into it, and only ran 1 pass, upscaling it to 4k with a model + some post processing models.
Told him it wasn’t a very good representation if he’s looking to get customers. So he asked me, to ask you if you have the same video or similar but higher quality so he has a reference to what you’re trying to achieve.
I downscaled the first try he did to 1080p since 4k far from did it any good. Want me to link his first try either way?
I have shortly checked your sample. Deinterlace first, it’s “TopField” and do it without BOB
Well, if I had the video in higher quality I would not use this for upscaling 
And I do believe it’s a quite typical example for low-res vids that come from a VHS rip, so if it doesn’t fare well on this it probably won’t for others…
I guess I’ll have to stay with IrisLQ V1 with 2x upscaling - which does a quite decent job on this cleaning the (chroma) noise and compression / line artifacts quite well. Not much detail gained by that, but at least the video stays true to the original.
I’d still be interested in the result from Video Enhance studio, yes.
