The assumption is that the GPU can do it better. Using Discord's built-in version last weekend and Skype for Business every day, I can tell you that there is definitely room for improvement from the software based versions.
I'd love to see gpu's used for processing audio. Maybe some kind of DAW intergration for running vst/s/i's. It'd only take for Waves to support it and the whole industry would be scrambling to do the same.
The assumption is that the GPU can do it better. Using Discord's built-in version last weekend and Skype for Business every day, I can tell you that there is definitely room for improvement from the software based versions.
Using good microphone, no noise, no distant sounds.
And when one uses lower quality microphones, there is simple but very efficient way of canceling entire room. System has 2 microphones. One is meant to capture person talking, other is wide angle/area room microphone.
You can imagine what each catches and what is filtered out.
hmmm a company boss once gave me one of those "ear pieces" for representatives a simple one ear+mic thing, you can phone while driving, get out of the car unload stuff, load stuff enter your house and the only thing the other person hears is your voice so clearly it's not new tech and you don't need AI to do it
plantronics voyager legend with 12k ratings on amazon https://www.amazon.com/Plantronics-Voyager-Wireless-Bluetooth-Headset/dp/B00DQ5NU76?th=1
I don't use it often but in the rare case of a conf call or a serious phone call I always use it
Using good microphone, no noise, no distant sounds.
And when one uses lower quality microphones, there is simple but very efficient way of canceling entire room. System has 2 microphones. One is meant to capture person talking, other is wide angle/area room microphone.
You can imagine what each catches and what is filtered out.
That software takes CPU cycles though and requires extra equipment. The idea here is no additional hardware or software (as long as you have a RTX card already).
I wonder what your response to this would have been has AMD developed this? I remember you praising the audio tech in Hawaii.
That software takes CPU cycles though and requires extra equipment. The idea here is no additional hardware or software (as long as you have a RTX card already).
I wonder what your response to this would have been has AMD developed this? I remember you praising the audio tech in Hawaii.
There is additional software here though. And there is quite a performance hit. According to that epos video, it's roughly 10 percent drop in FPS per audio source, not insignificant. Could be worth it if you're a steamer though.
It automatically drops the max frequency by 200mhz. Plus I'm sure it hits other parts of the GPU.. on the subreddit the actual impact seems to vary based on the game from basically nothing to 10-15%.
Idk it's a free tool, that seems to work really well, and presumably be supported going forward. I don't see why anyone would be against it. I would imagine newer series with improved/more tensors would lessen the impact.
This is seriously great and i cannot wait to check it out. I bought a nice microphone (yeti usb thingy, sounds great) but it's so sensitive that it captures too much.
On OBS you can setup filters/noise cancellation but it's such a brute force method that you're losing a lot of 'texture' in the sound, to compensate you end up having to use compressor otherwise the voice is extremely boring and flat.
That software takes CPU cycles though and requires extra equipment. The idea here is no additional hardware or software (as long as you have a RTX card already).
I wonder what your response to this would have been has AMD developed this? I remember you praising the audio tech in Hawaii.
TrueAudio is great technology that did not catch ear of mass consumer. But good spacial audio is not dead, it is back with "raytracing-like" methods.
Honestly, a lot of people have problems with microphones in gaming. Not because super bad software, but because of super cheap microphones with usually not so good input on sound cards. (And this often can't be fixed by buying more expensive microphone because it still uses analog jack, has unstable resistivity and sound card is not exactly good with given impedance.)
But everyone I know who bought quality USB microphone (mic itself + electronics made for it and PC gets just data) has no problem at all.
And what kind of logic it is to buy $400 GPU and then be like: "I can now buy $2 microphone and it only sounds half horribly."
And then again as you wrote: "That software takes CPU cycles" != "no additional hardware or software (as long as you have a RTX card already)."
Apparently it will be processed by driver and it eats CPU cycles and it eats GPU cycles.
Do I want to eat some cycles on CPU where I have extra cores that just sits idle? Sure, as long as it makes things better.
Do I want to eat some cycles from GPUs budget when it is already 100% utilized?
Let's just hope that people will not find out that it runs on non-RTX cards too because this processing ends in driver and is handled cheaply by CPU anyway.
TrueAudio is great technology that did not catch ear of mass consumer. But good spacial audio is not dead, it is back with "raytracing-like" methods.
Honestly, a lot of people have problems with microphones in gaming. Not because super bad software, but because of super cheap microphones with usually not so good input on sound cards. (And this often can't be fixed by buying more expensive microphone because it still uses analog jack, has unstable resistivity and sound card is not exactly good with given impedance.)
But everyone I know who bought quality USB microphone (mic itself + electronics made for it and PC gets just data) has no problem at all.
And what kind of logic it is to buy $400 GPU and then be like: "I can now buy $2 microphone and it only sounds half horribly."
And then again as you wrote: "That software takes CPU cycles" != "no additional hardware or software (as long as you have a RTX card already)."
Apparently it will be processed by driver and it eats CPU cycles and it eats GPU cycles.
Do I want to eat some cycles on CPU where I have extra cores that just sits idle? Sure, as long as it makes things better.
Do I want to eat some cycles from GPUs budget when it is already 100% utilized?
Let's just hope that people will not find out that it runs on non-RTX cards too because this processing ends in driver and is handled cheaply by CPU anyway.
Few things I want to point out:
I watch tons of podcasts/twitch streams that feature 2-12 sometimes more people chatting on microphones - most of these streamers/podcasters, the good ones, already employ various noise cancellation/compression software on the stream to improve the quality. They typically aren't playing games when the podcast/stream is going so the idea that it even uses any GPU % is pretty irrelevant for them. I know a few of them that use Krisp, which is a paid software and as demonstrated this seems to do the job better, for free (as long as you have an RTX card).
This cancels noise on outgoing and incoming audio streams. So it's not just improving your $x microphone but in use cases like the one I mentioned above (content creators) it's improving everyone's $2 microphone, including people that don't have RTX cards.
You keep saying "get a better mic" but in the video above we see a guy who has a Shure SM7B, a $400 microphone, who not only demonstrates that it has value even with that microphone but says before this released he used various other noise suppressing software/hardware. So clearly even with expensive microphones (arguably the best microphone there is) it's still useful.
It was trained via cuDNN and we already know the inference from cuDNN can be ported to various other hardware. Obviously Nvidia is probably not going to offer that because they are essentially selling this as a value-add for RTX owners but I don't see the problem when they are offering as a free plugin.
I've tested it myself by listening to the recording device and it works amazing well. No background noise and I was clapping right next to my mic and it didn't pick it up. Really pleased with this.
I use a Blue Snowball ICE, the even cheaper version of the cheap and popular USB microphone. I've got a great EqualizerAPO setup for this thing. This app puts it on another level. It sounds fantastic with what I already had setup.
Though I did notice a lot of memory usage. Maybe that's normal?
It goes to that memory usage immediately, not over time. Thought that was curious.
EDIT: Had another strange issue. I was messing with settings, making recordings. I closed the RTX Voice window (which minimizes to the system tray and stays active). When I reopened Audacity to try another recording, all I got was VERY loud fuzz. Uncheaking my mic from the RTX Voice app made it worse. I had to fully close RTX Voice and relaunch, then it's fine.
EDIT: Here's my 3 way comparison at the moment.
1st is just the 'raw' $40 Blue Snowball ICE.
2nd is with EqualizerAPO boosted settings, along with ReaFIR (classic free VST) to clean up noise... badly. ReaFIR can be turned up, but it makes my voice sound terrible.
3rd is with RTX Voice at 75%, EqualizerAPO stays on with the same settings, but the ReaFIR VST is disabled. IMHO, it sound fantastic, completely distinguishable from the first sample, to the point it sounds like a different microphone.
https://www.mediafire.com/file/huq2snbshqwpaiw/RTX_Voice_Comparison.mp3/file
It goes to that memory usage immediately, not over time. Thought that was curious. EDIT: Had another strange issue. I was messing with settings, making recordings. I closed the RTX Voice window (which minimizes to the system tray and stays active). When I reopened Audacity to try another recording, all I got was VERY loud fuzz. Uncheaking my mic from the RTX Voice app made it worse. I had to fully close RTX Voice and relaunch, then it's fine. EDIT: Here's my 3 way comparison at the moment. 1st is just the 'raw' $40 Blue Snowball ICE. 2nd is with EqualizerAPO boosted settings, along with ReaFIR (classic free VST) to clean up noise... badly. ReaFIR can be turned up, but it makes my voice sound terrible. 3rd is with RTX Voice at 75%, EqualizerAPO stays on with the same settings, but the ReaFIR VST is disabled. IMHO, it sound fantastic, completely distinguishable from the first sample, to the point it sounds like a different microphone. https://www.mediafire.com/file/huq2snbshqwpaiw/RTX_Voice_Comparison.mp3/file