Topaz Video AI 5.5 + 5.5.1

Updates:

We’re working on finding the cause of VFR outputs and will release a patch as soon as we have one.

Cloud credit pricing is currently being updated.

4 Likes
2 Likes

Here are some screenshots of my machine., within the benchmark, and also processing speed with Proteus 4, which is more than 2 times faster than yours. I know the processing speed is different from one input to another, but that’s a pretty significant difference.

Ah yes, but I’m taking about upscaling to 4k. :slight_smile:

Topaz Video AI  v5.4.0
System Information
OS: Windows v11.23
CPU: 12th Gen Intel(R) Core(TM) i9-12900K  63.779 GB
GPU: NVIDIA GeForce RTX 4090  23.576 GB
Processing Settings
device: 0 vram: 1 instances: 1
Input Resolution: 1920x1080
Benchmark Results
Artemis		1X: 	30.77 fps 	2X: 	12.87 fps 	4X: 	03.72 fps 	
Iris		1X: 	36.67 fps 	2X: 	17.39 fps 	4X: 	03.72 fps 	
Proteus		1X: 	38.69 fps 	2X: 	16.82 fps 	4X: 	04.03 fps 	
Gaia		1X: 	15.49 fps 	2X: 	10.55 fps 	4X: 	03.60 fps 	
Nyx		1X: 	16.62 fps 	2X: 	13.32 fps 	
Nyx Fast		1X: 	29.33 fps 	
Rhea		4X: 	03.38 fps 	
4X Slowmo		Apollo: 	32.59 fps 	APFast: 	57.07 fps 	Chronos: 	30.32 fps 	CHFast: 	26.80 fps 	
16X Slowmo		Aion: 	28.98 fps 	

Well, that’s a different story then. I was just trying to provide my input.

Personally I don’t need to have everything in 4K, 60 fps. If it’s something that you can fall asleep to, then I am good. And that usually applies to old tv series where the dialogue is enough to keep you awake when you are dozing off.

Personally I don’t need to have everything in 4K, 60 fps.

Me neither. At original fps, things are usually good for me. Besides, chosing higher fps rates, you eventully have to intruce some hiccups because of missed frames (unless you precisely doubled your rate).

I’m surprised. Like you, I assumed they had just ignored the ffmpeg license. Good surprised, mind you. :slight_smile:

No one pointed out the export video contains “_2” since 5.5? anyone know why this underscore with 2 start appearing on export video on all models since 5.5? is it something to do with cloud exporting?

bandicam 2024-11-28 07-25-43-155

I see Log out here… Do I need to do that before every close? When, why and what idea is for that?

My guess, yes. _1 will likely be their default cloud export option (at $151,485 dollars for a 14 minutes render); and _2 the option for when you manage to find the export option for local renders.

:rofl: :rofl: :rofl: for that much I’ll just build my own datacenter I think

5 Likes

I kid you not! Topaz Video AI 5.5 - #256 by diego.peixoto

I had expected a dev to come out and say, “No, no, the program just misplaced a decimal point!” Alas, that didn’t happen.

1 Like

They train their own models.

1 Like

Satire
Don’t worry only platin credit card owners can get 150.000$ charged for 15min cloud render. So it’s fine :laughing: because this function is for rich peoples only who finance us TVAI development and our renewal price drops. We look forward to a wonderful TVAI in the future, long live cloud rendering :sweat_smile:

2 Likes

All kidding aside, can you update us on the new prices? And is that $150,000 for 1 movie, mentioned above, true? If so, then kindly remove cloud rendering from this home user product. Thank you.

3 Likes

Or is it? Just asked the dev for clarification.

1 Like

Yes, different technique. I know nothing ab rendering video in the cloud.

The individual frames can each be broken up into as many pieces as needed and rendered on any machine in the farm.

Yes, you’re right. Distributed rendering I think is the exact term for that and only makes sense for hires stills. Since Redshift I don’t have to use that anymore. Rendering animation in the cloud, each node takes 1 frame, or more of you like. Rendering video is much faster though. Using multiple machines to render a frame would only slow it down…

So, why this weird detour then?

Please don’t ask why lol. I’ve worked with A/V & 3D applications for 20 years. Then there is Topaz. At no point in time have I ever understood pretty much any of Topaz’ decisions when it comes to the development of their Video AI app.
Just opened again after months, and again the whole freaking thing has changed. Not better, just… changed. Still weird (wrong) terminology and icons, still overly complicated ways for the most simple things, still unnecessary stuff and still important things missing. The application never gets better, things still make zero sense to me from a UX/UI perspective. Never have I seen such mess as with Topaz Video AI app.

If this was a purely professional app, performing like this… I don’t think it would’ve survived…

2 Likes

the position of the application windows is not being reminded correctly. it always starts a little bit off next start.

1 Like

Is there is a problem with cloud render quotes in the app (5.5.0), according to your website a 3 minute video rendered at 1080p should cost 3 x 6 = 18 credits (Output file size - 24fps/1 min @ 1080p x Credit cost/video - 6), altho it does say ‘according to model’ (what does that mean, you just gave me a price but your saying it’s not the price?), however I loaded a 3:19m (yes 3 1/2 minutes not hours) 640x480 video upscaled to 1080p, and for Iris it quoted 1792 credits and Rhea 7272 credits, with both taking 36 minutes to complete, that would cost £1818 for 1 Rhea video without subscription, and is way higher than what your website alludes it would cost me, plus I did it locally in 13 minutes on a 3yr old desktop. This cannot surely, be a serious pricing strategy.

1 Like