Read more.Beta AI-based software supports many popular broadcasting and conferencing apps.
Read more.Beta AI-based software supports many popular broadcasting and conferencing apps.
I guess they had to find a use for the silicon they are charging everyone for.
Interested to see if it actually works and also how much, if any, it takes away from game performance if the game is using the cores for something already like DLSS.
Tensor Cores, a solution looking for a problem in the consumer space.
CAT-THE-FIFTH (20-04-2020),Output (20-04-2020)
Interesting, I rarely stream with RTX on so this would be an interesting use case. OBS has a decent built in noise gate but keyboard clicks still get through at times, so i'd be curious to give it a try.
When not streaming I can continue to have RTX & DLSS enabled and enjoy raytracing
The question now is: how much Latency does this add to the audio signal.
Can't say I've ever had much of a problem with anyones' background noise, apart from one guy who has parrots in the same room as his PC. Occasionally one of those will yelp or squawk or whistle or do something else that fires over the airwaves like a bullet and cuts through your skull at the most inappropriate moment... I bet this tech still does nothing to shut the little cretins up, though!!
_______________________________________________________________________
Originally Posted by Mark Tyson
Nvidia integrate tensor tech on Microsoft Word so that they can write reports for me.
It seems to work: https://www.youtube.com/watch?v=rd7c7FVofOE
_______________________________________________________________________
Originally Posted by Mark Tyson
The problem of Noise cancelling is not difficult at all for hardware. Most mid tier headphones have it. And if you want proof it works on CPU alone, here you go KRISP software https://krisp.ai/ that does exactly the same without requiring GPU of any kind...
But kids going to like it xD
A Guru3D forum users says that RTX Voice works on GTX 10 and 16 cards if you fiddle with the extracted installer file. Indicates that the app must use a CUDA-based GPGPU codepath rather than Tensor cores, and understandably Nvidia was quiet about that. https://forums.guru3d.com/threads/nv...es-how.431781/
I haven't tested this.
So, i've used this for calls all morning and it's actually really impressive - works really, really well.
It works both ways - i.e. for your microphone & speakers - and the difference is stark.
After my testing i've decided to only enable it for speakers (inbound sound) as I don't have any real background noise at home personally (And it does some odd things for keyboard click removal - so i'll keep my noise gates for that) but for incoming sound, the difference is night and day.
Someone joined one of our morning calls today and had a nasty buzz in his microphone - very annoying whenever he spoke. Enable this though and the buzz completely vanished and his voice came through loud and clear.
After 4 weeks now of constant conference/video calls over teams i've encountered so many different qualities of microphone/background noise and so on, and it can be super distracting (particularly if it's the echo type as someone doesn't want to use a headset!) - this will make my life much more comfortable whilst stuck at home
Good job Nvidia.
There are currently 1 users browsing this thread. (0 members and 1 guests)