I recently ran a series of one-on-one tests of some of the new ai-based upscalers against the old guard like Davinci Resolve and After Effects. As the other article goes into more detail, when Topaz Labs and others talk about artificial intelligence, it’s best to think of it as machine learning or shape recognition training software, giving thousands and thousands of before-and-after pairs of low-and high-resolution images or videos. In this way, the algorithm receives a real world training on how to restore clips. With up to 8K on the menu, it’s a brilliant idea that could get better year after year.
One of the most important AI options is Video Enhance AI from Topaz Labs, which offers a range of products designed to enhance your photos and videos when they are soft or loud. Video Enhance AI is your video-specific upscaler, with additional features recently added – as they say, “” upscaling, noise reduction, deinterlacing and world-class frame rate conversion on your desktop.”
I thought I would do more in-depth testing on video enhancement of AI and talk about how they are used in the real world and why you might want it.
First of all, if you are interested, use the 30-day trial. It is in use that you see if your computer can handle it and if you can get out enough to seduce your wallet.
Video Enhance AI as I write a little Google search allows you to easily find 15% discount by finding the coupon codes of one of their partners (tip: google “Topaz video enhance coupon”) and is also on sale at various points throughout the year. Topaz Labs notes that with it, you only get one year of updates (without maintenance updates) and offer you an extra year for 99 which, which I didn’t take. The pricing page suggests that you can buy it after, for example, in two years or every time you release version 3 to then update it. So far this year, Topaz Labs has released a number of different updates, with some big changes and new models, so updates are clearly something you might want to pay for at some point.
First, this is a standalone application and not yet a Premiere or OFX plugin, still available in the cloud, although I imagine that could eventually change. It can run on Windows and Mac, and also lists support for M1 chips. The official system requirements are listed as quite low – 2 GB of VRAM, 8 GB of RAM, Intel processor of 2015, etc.in the real world, I could not run it on my laptop at all, which is far above these specifications.
There is no doubt that it is one of the most hardware-intensive and slow programs you can find and that works best on a desktop machine with a good GFX card. Again, make sure you use the test before buying.
When you open the standalone application, you are greeted with a fairly simple GUI, which, at least for me, did not fill with confidence. Three tutorials are offered below and these are worth your time.
Once you’ve inserted one or more clips, you can set input and output points as needed, and then go to the Right column-choose a model, choose your output size, optional grain settings (grain can make it a bit more natural), and the video output format. I use ProRes 422 HQ, but there are advantages to working with image sequences, such as partially rendering a clip and continuing where you left off.
There are different views to look at, your clip-only view, split View, look side by side and compare, but it is important to note that there is no live preview, as you would do if applying a filter in Photoshop in all display modes, you do not see the result of the model, as long as you did not press the blue Preview button.