Not everyone wants to buy a system just for TVAI, I’ve bought a Win11 system in the last 3 months, which isn’t suitable for TVAI as I already had one for that, and really have no interest in buying yet another (large) computer just to run it, alas.Topaz Video AI focuses solely on completing a few video enhancement tasks really well: deinterlacing, upscaling, and motion interpolation. Which is probably most annoying for those who tested it on their non-AVX2 system (BTW I did this before purchase), found it worked and a little later they found the remaining 10 months (or whatever) of their free upgrades were no use to them and annoying issues remain… I only have a little upgrade time left, so probably won’t make it to 4.1.x and this doesn’t really apply to me, although I have a licence for 6 versions of v4 that I can’t use. I guess the good news is that they can keep improving it.Īlas as I can’t install 4.0.4 on my non-AVX2 system, and my questions about whether it would work without AVX2 if you had a good graphics card (so would use that instead of AVX2) went unanswered, I decided what-the-heck let’s give it a go, since I have a licence which supports the latest version and it has been suggested it will work this way…Īnyway, here’s the Benchmarks run on my AVX2 system (for amusement value, and why I’d like to run it on my faster system - BTW I checked and the BIOS is … But than with the same model I used it on a German Shepard and it looked amazing. I remember trying to make some sand falling 2x in size, but its such a unique footage, probably never was in its database so it didn’t look acceptable. I guess it also depends on the model, not all have same level of success rate. But when it works, its really impressive, like with this example of yours. There were some posts in this forums with alien looking teeth and car sign looking like something it should not. One thing that is a bit of a hit and miss with AI is that if it has enough references it can do something like this and salvage what would otherwise be a impossible, while at other times when its not sufficient trained enough it can create some unpredictable results. considering it shows the same image frame before and after. In the original on the leeft there were all details gone, but video a.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |