So there is this new plugin that I just found out about
It’s very interesting and powerful, it practically is an AI contained in a plugin format that does just one thing, looking for any form of feedback reverb and noise that doesn’t match a human voice and removes it with mind blowing quality. Each new plugin instance is another AI, so things are very cpu consuming.
There’s a few down sides to that. For one, it’s damn expensive. Second, it’s very finicky, as it doesn’t run in every computer, they tested a bunch of configs and only a few are really working, it mostly needs a separate computer.
I tested on my very beefy AMD 9900X3D/64GB RAM and besides all my Waves plugins for voices and kontakt/nexus for keys, I can squeeze 4 instances, one for every voice, but had to up the latency to 128 samples instead of 96, or else cracking appears. Haven’t bought yet the plugin, am just trying it, as they let you use it freely for testing, the audio drops out every few seconds.
Is there anything Deskew can do to accommodate this plugin, to work better? Just asking, cause I have no idea if/how it’s possible, but a lot of huge sound engineers are praising it. Mariah Carey’s sound guy heard about it and the next day bought it for the tour, untested, it’s that good.
