#Can we have access to raw vocals from Neuro's covers?
1 messages ยท Page 1 of 1 (latest)
pls do
I don't think so sorry ๐

yes, here you go
i mean, it won't be as good, but you could just extract the vocals and process them to remove compression.
You can extract vocal and de-reveb with izotope RX. De-compression can be done with any transient shaper and expander plugin
You mean the extracted vocal stem? Try RX10 highest setting
That's the best algo on the market
Yeah that should work too. If RX can't give you the desired result then literally nothing else can lol
Use Music Rebalance with only Vocals on, try the highest quality and separation
Also you might have to deal with a little bit of artifact because of Neuro's TTS voice. It's never gonna be as smooth as extracting human voice. But if you're doing remixes or remasters I suppose it won't matter much
thats ozone no?
Do you have RX standard? RX's algo has better quality than Ozone's
Ozone's called Master Rebalance, it's not really meant for stem extraction
RX standalone editor is really advanced, I'd recommend get all stems processed there before dragging them into your daw
im not using rx to seperate vocals and instrumentals
instrumentals are whatever cuz i have the flacs
Ah ok, but still RX's vocal extraction is widely regarded as the best
@round plover has leveled up! (4 โ 5)
That's the best algorithm for ya
I'd personally preview like a 10 sec segment before committing
you said this does dereverb as well? or just separation?
i have a feeling its using htdemucs
similar stem format
right
when learning do i learn the reverb tail or just let it play?
Just let it play
The tail can be adjusted to your liking, like super dry or just reduction
oh wtf it doesnt train on the whole song in the daw
You're gonna need the standalone editor for best result lol
Also, if you have both RX and Ozone I suspect you have Neutron as well? Neutron's transient shaper can bring back some attack, and the compressor module can be used next as an expander. I personally won't worry too much about dynamics here because Neuro's TTS voice should already he pretty compressed raw. I have to log off soon so here you go. GL!
Even with RX it's still going to be awful - not only are you dealing with lossy audio in the first place (which RIP all the highs), you're also dealing with the inconsistent results of RX
| |I || |_ audio 
Sure, you can play with the audio all you want, you're still going to be spending half the time working on crappy audio with limited results, or even worse, artificial mids/highs with heavy background interference
RX shines at dealing with source quality audio that you have actual access to