Which AI model do i use to upscale anime content? Lets say the video is 720p, ok i want to select it to upscale to 1080p, lets say the quality of my vid is medium, and its blurry, and now whats left is AI model itself?
Please provide actual documentation for the models, how they different, pros and cons, and even things like relative processing requirements. This video just says "use the Wizard" and the Wizard itself barely explains its decision making process. The best documentation is the is tooltips when mousing over each model name.
I just tried it with NTSC video... the most amazing feature is dehaloing. I didn't expect it to ever be addressed, but Topaz shut me right up 😂 The quality and control is incredible, instead of dedicating hours of mediocre restoration, this software does way superior results in seconds!!!!
Not really. It takes a long time to render whether it's upscaling or anything really. Just a simple 5min video upscaling to 4k takes one hour. Once you introduce frame interpolation it's even worse. A movie takes days to render. And this on a decent specs PC as well. 13900k + RTX 4090.
I'm hoping one day we'll see a model that makes Enhancement and Framerate Interpolation become a 1-step process Artemis + Chronos in a single click, especially if Chronos evolves to improve rendering times
In version 3.0 it allows for filter stacking, I’m told by TopazLabs, so you can upscale and interpolate together. I’m yet to use Topaz as my current MacBook Air is 8 years or so and will most certainly be to slow to cope with it.
@@VintageBoxingTM hey my friend quick question I’m new to this stuff. Does it matter which order I do the “stacking” should I Denoise then upscale or vice versa
@@razorbackroar honestly mate I wouldn’t know myself just yet as although I’ve just purchased the software, I won’t be getting my new MacBook Pro M1 until this weekend to try it out. Mentioned Mac deliberately in case anyone wants to chip in with their experiences of it using Topaz.
Is there a setting to see the comparison on the one frame still? Instead of having to play a 30 frame clip. In your video above, the one frame looks all the same. It would be great to see a preview of the preview. The 30 frame clip doesn't really show much of a difference between the 3 versions. Or can you extend the 30 frames to something longer? Like 5 seconds.
Hey hello i have a problem when i try to render the ai models to compare them only one of them are completly renderd and on the other ones it says :not renderd could you pls help me
Hello Topaz. Please help! If i want to compare ai models just one is rendered in the preview. The others arent. This way i cant compare before processing yhe footage. Whats the problem?
Hi! If this is happening in comparison view, would you mind writing into our support team so we can find a solution? In general, it does have to render each separately which can take some time but should automatically start rendering the next model once the first has finished. You can find our support page at support.topazlabs.com 🙂
So I really wanna be able to upscale classics shows like sabrina the witch or married with children The least artifacts is dehalo but the skin looks to sharp and the mouth moves weird So i have to use proteus I wish we could get a version of dehalo where we can adjust dehalo like proteus Its all I want cause I have other shows I wanna upgrade to im using 2.6
Hi Jacek! This product is currently on sale and we offer an educational discount with proof of enrollment (just send us a request through our support). Alternately, we give out free 3 month licenses to top contributors at the end of each testing period as a thank you. If you're interested, you can apply at community.topazlabs.com/g/Video-Beta-Testers/
I'm really concerned that every time I use this, it connects to your servers. What happens if your servers go down or you close your company? Also, are you uploading content from my computer?
Hey there, Mat. To clarify, Video Enhance AI only connects to download optimized models and you can turn this off in Preferences. As far as uploading, the only thing we collect is anonymized data, and you can turn that off in Preferences as well. We do not collect *any* video content processed in Video Enhance AI.
@Fernando Martinez We'd love to work with you to figure this out. Please contact our Support team so that we can help: support.topazlabs.com/. Click on the blue chat bubble on the lower right corner of the screen to chat with us!
Hi! Each AI model is trained on hundreds of thousands of videos and takes several months to train which unfortunately is not feasible for most users. I have noted your interest in the ability to train or provide your own models or model input. We do have a feature in the image apps to give feedback on model quality and are looking at some ways we can implement this in Video Enhance AI. If you'd like you can also submit this and any other feature requests/feedback directly in the app through the Help menu. 🙂
Hi! Speed will depend on a few variables including your hardware, settings under File -> Preferences, the model chosen, and input/output size. In general, we recommend using a GPU (one that is considered a "gaming" GPU is typically best) with 4GB or more VRAM. If you'd like, you can write into our support team with a copy of your system specs and we can give you some additional information about how we may be able to get it to run faster or general recommendations on hardware/eGPU. Our support page is at support.topazlabs.com
@@kenlaw1253 Yikes! If you'd like to write into support we're happy to help you with this as well and see if we can get it running any faster for you :)
