Video Enhance AI v1.0.2 -- Initial Release

With your GPU you would have 10x faster processing speed than on CPU.

Processing time highly depends on the resolution of source video and also on scaling factor you select. When trying, I noticed that the same video was processed a bit slower when pushed thru 3x scaling, comparing processing this video with 2x ratio. All on RTX 2080 (non SUPER, non Ti).

This program worked for me on Windows 7, if this OS isn’t supported by this program, then it was just luck. :slight_smile:

1 Like

I was waiting the Mac version of Video Enhance Al with the greatest hopes, but I am very disappointed.

I have used the version for old still photos which works terrific, but with this one that is made for videos, after one hour of different tests on footage (Black & White, as well as colour) I can see NO DIFFERENCE

Is it supposed not working at all with old footage? Your superb teaser using the old arrival of the train in the french station was so astonishing. I can’t imagine you released a version that doesn’t change anything. I mean : no improvement on the details, nothing better on the blur parts…

I tried every settings…

Hi vidivana, You had my heartbeat raised pretty high for a moment there. I had made an assumption, for the millionth time in my life, that when I saw “use CPU” after I dropped a video file into Video Enhance, that the software had chosen that setting as the best. It was the first time I’d used the software and wasn’t very familiar with it. I just put another file in and it now shows 2 options, CPU and my graphics card and it shows GPU selected. It looks like they have changed the user interface a bit to maybe make it more understandable.

Hi majatt, I made a wrong assumption. My machine was using GPU. It looks like they have change the UI a bit and I now see 2 options, CPU, and then my graphics card. And it selected GPU as the best option. Sorry I misled you, but I wasn’t very familiar with the software at that time.

Give Topaz team some time - it’s REALLY hard and time consuming to train AI the desired way. It should only get better as the AI data will be more accurate and will include more “cases”. Notice that the team will have to incorporate almost every AI algorithm - Denoise, Sharpener, Gigapixel and probably something from “JPEG to RAW” AI into this software. That, and also cropping, aspect ratio setting, possibly audio support to mux everything into the final stream without any external tools.
It HAS to take some time.

Yes mikmod1, I understand it tales some time, but if I want to make the decision of buying it (with 100 dollars off) I only have 12 days to decide.
My post was made to know if other mac users tested it and succeed to get good results, because once again, every setting I try makes no difference. The movie stays the same. Maybe more pixels but no extra definition or sharpen job

I think this is the same issue as with Gigapixel AI. Your video footage detail density does not match the pixel density. I suggest converting the video file to a smaller pixel size. For example, convert a 640 x 480 video file into a 320 x 240 video file and then give Video Enhance another try!

I downloaded the update yesterday and tested so far some footage, which I also tested in the version before.
Overall: Video Enhance AI v. 1.1.0 renders remarkable faster using my 6GB graphic card.
I recognize so far a real improvement on face recognition. Faces get with the version before often damaged. This seems to be nearlly solved with v. 1.1.0: faces hold lines and structures much more better.
Also hair structures are now improved.
On the other hand: in some footage I see now in comparison to the version before remarkable differences on deblocking and denoising. With the old version I get remarkable results on deblocking improving lines/edges and denoising. Now with version 1.1.0. the results also remarkable, but not as good as in the version before.
I guess, the price, if you would save/preserve faces and e.g. hair/grass structures, is, that the denoising and deblocking has to decrease. But, this depends on footage to footage.
Therefore I ask again for a next update, to give users manual adjustments slighters for deblocking, sharpening, denoising. Perhaps AI wouldn’t be able, to deal with each footage automatically.

Topaz: the new version brings remarkable improvements in speed and preserve of faces and hair structures. Great.Thank you that! On the other hand, the deblocking and denoising and sharpening are not as good as in the version before. I guess, you made the decision like a “middleway”.
I’m sure, if you would implement manual adjustments, that you nearlly have it.

I tried on a old noisy VHS footage and from my experience to get better results it is good to apply some denoise filter before processing in Video Enhance. Otherwise it creates some weird artifact on human faces.

1 Like

Imo thanks so much!!! Your suggestion was the good one. Strange that it is nowhere mentionned by Topaz. If I didn’t knew I could never had a difference with the original! Well I am speaking of the MAC version that just went out. It is a bit odd to reduce the size before increasing it, but anyway the result is there. Thanks IMO, you’re a master!

@ alreverend: Glad I could help. :smile:

Presumably this means that files enhanced now, will be of inferior quality compared to files enhanced in 6 months? If that’s the case, what’s the point in enhancing extreme time consuming files now. Why not wait until the program can perform better? If I can only realistically enhance one 30 minute file par day, then the prospect or re doing an entire series 6 months from now is not appealing.

Same Issue with X2 xeon E5 2667 V2 and Geforce GTX Titan X with 128Gb DDR3.

about 1.5 FPS. Only CPU encoding no GPU

That’s totally up to you, when you buy a software, right? If you NEED it now, buy it and be patient with conversion times or delegate another machine to this task.

I’d not thought about that.

Can anyone tell me if the denoise/deblock feature deintelaces video? I did a few clips and the comb edge effect seemed to reduce considerably. Is that just a side affect?

No support here then.

I can only guess that it was a side effect. There are other ways to deinterlace video footage. Maybe try VirtualDub or VLC media player? I suggest deinterlace before using Video Enhance.

Good quality SW! My problem is only that when I upgraded my GPU from GTX 1050 Ti (which worked) to GTX 980: only CPU will work (even when GTX 980 selected).
Will this GPU be supported? When?