nvidia rtx voice Would RTX Voice work on Pre-Recorded recordings?

[Mr.Stork]

It would be good, to be able to do that especially for making Youtube Videos or similar work.

I think it might be possible if...

If, I take the input from windows not mic, play the voice, and record the output. I think. Or, I could just install a app which can add a virtual mic which will only play what I need to be played and same with the output.

Idk, is there a different that uses AI to do this without so much hassle because it really shocked me with the results I got!

 

I saw the video on Short Circuit and it was very good, so I had to try it. Quick note, I don't have a RTX GPU, only a GTX 1660Ti. I used the Guru3d's forum's method.

[The Flying Sloth]

Yes, in theory there's no reason it shouldn't work, perhaps there are software limitations but I see little reason it wouldn't work.

Sloth

[Kdog]

I actually just did some testing on this so I can answer fairly well. Firstly, I did this two ways, firstly by using a physical loopback cable from the output of my interface back into the input which seemed to be the best way to do so. With that being said, I also tried using VB-CABLE which simulates an input and output. With this, I used the input and output as an insert in Adobe Audition routing some intentionally messy audio and got good results but it was a hassle to setup. 

 

If you want more information/have any questions let me know!

[Mr.Stork]

19 hours ago, The Flying Sloth said:

Yes, in theory there's no reason it shouldn't work, perhaps there are software limitations but I see little reason it wouldn't work.

Sloth

Are you flying rn or playing audio game.

Yeah I think that too.

[The Flying Sloth]

5 minutes ago, Mr.Stork said:

Are you flying rn or playing audio game.

[Mr.Stork]

19 hours ago, Kdog said:

I actually just did some testing on this so I can answer fairly well. Firstly, I did this two ways, firstly by using a physical loopback cable from the output of my interface back into the input which seemed to be the best way to do so. With that being said, I also tried using VB-CABLE which simulates an input and output. With this, I used the input and output as an insert in Adobe Audition routing some intentionally messy audio and got good results but it was a hassle to setup. 

 

If you want more information/have any questions let me know!

K so I don't have an audio interface (yet xD), but I can get the voice meters' vb cable app which you mentioned. So I'll try those if possible. Tho they should really make a dedicated thing for pre recorded stuff,cuz it's truly a hassle.

[RejZoR]

I think you could fool the software into inputing from internal source. Creative soundcards have "WhatUHear" input module which essentially tricks Windows into thinking it is getting external input, but is really just sound you hear at this very moment (from media player in Windows for example). Realtek has it too but is named differently. Just set that as input/microphone if you're allowed to and anything you play should loop through that and be processed by this thing. You will have to process it in real-time though which means filtering 5 minutes long clip will take 5 minutes. Coz it's a hack. Would be nice if NVIDIA also offered offline filter where you'd just stick sound file into it and it would process it in a moment.

[Froody129]

I would imagine you could just do a screen recording or a sound recording or anything like that. Turn on RTX Voice on your output device and play whatever video or audio you want. Then you'll have a recording with RTX Voice. Likely not so great for pure quality but it should work. 

 

 

[Kilrah]

If you use Voicemeeter mixers it's easy to tell RTX Voice to use one of the virtual output tracks, assign the RTX Voice mic to one of the real inputs, set your default audio out on Windows to one of the virtual input tracks and your real audio output to one of the real output tracks.Send the virtual input (default audio out) to the virtual output (RTX voice), and the real input (RTX voice) to the real audio out (speakers). 

 

If you use Voicemeeter Potato you get 3 virtual input tracks so you can use Windows' per app output selection to use either of those 3, so you can send some straight to speakers and some through RTX Voice.

 

It's funny to launch a movie and get nothing but speech out 

I notice I prefer RTX voice noise suppression about half way, not full...

[DJSAiNtC]

I have done it this way and it is fairly easy... First download a virtual wire program like VB-Audio Virtual Cable (you actually don't need Voicemeter to do this either). 1. Setup to play a video in a normal video player like VLC where you can select the audio output to goto VB-Audio Virtual Cable. You can also select audio output in windows settings if using windows craptastic media player. 2. Set the mic input on Nvidia RTX voice to VB-Audio Virtual Cable. 3. Set your recording programs mic to Nvidia RTX Voice (easy to do in Audacity just up at the top next to the mic icon). 4. hit record on Audacity (or sound edit program that records), hit play on VLC (or whatever video player you use to play original video), hit stop on both when done. Ta Da then move new audio into whatever video editing software you use, replace audio and line up, sync etc.

[The Flying Sloth]

I will note however that there are much better sounding options, iZotope RX7, Klevgrand Brusfri, there's even an open source RTX voice alternative on Github. 
Just use one of those, will be much less buggy.