#newtone

[artstationable]

Mark Hretskyi

Concept artist/Illustrator [email protected]

artstation youtube linkedin twitter instagram

More from «Artstation» here

[kermakatti]

Nintendo Swooce

[notokra]

newtone.inst: https://www.instagram.com/p/CcedsHHPM8a/

Art Direction & Design @newtone.inst

Photography YUKI KUMAGAI @yuuukikumagai

Styling NATSUKO KANEKO @natsukokanekoop

Hair & Make-up SHINJI KONISHI @shinji_konishi

Paint Art KOICHI YAIRI @koichi_yairi

[punkrockmixtapes]

Listen/purchase: Brain by BAD OPERATION

[deadpamvllo]

https://soundcloud.com/superwiibros08/i-discovered-the-newtone-effect-of-fl-studio

[kermakatti]

Huomenta poika

[minindo]

Any advice on dealing with sources that don’t work that well in melodyne? (something like dat boi for example where a lot of the words are identified as sibilants in melodyne)

It's case-by-case. Some sources might genuinely not be tonal, but other sources might need to be processed (or de-processed) first to work better. For example, if you're pitch-shifting something monophonically you should almost always remove reverb (like with an algorithm on mvsep.com). If there's backing vocals, try the lead/back vocal seperator model, or for other vocal harmonies you can use the polyphonic mode in Melodyne to separate those, see my video doing this with guitar chords.

You can also use note assignment mode to reassign what notes it detected a source at, which can sometimes help fix a note wrongly detected as percussive. There's also a method to "force" Melodyne to ignore the issues it's detected with the syllable and let it be pitched anyway, here's a gif showing how to do it (@melvinheboyi made this wonderful gif, I just reuploaded it to my website to link it here.) From my experience, it works 50% of the time and the other 50% of the time it just makes it silent. It can also lead to a different "true" pitch sound than what it visually shows, as you're forcing it into an assignment that may not be correct.

Sometimes though, if all else fails, maybe the vocals are just too raspy (like Lady (sometimes)) or robotic (like Dat Boi) to work in Melodyne's melodic algorithm. Probably just use other tools to pitch-shift it, honestly. Though I remember for my very first rip on SiIva I used a combination of Melodyne (which I was very unfamiliar with at the time) and FL Studio's Newtone. I'd say that while it's doable in Melodyne with some reassignment, Dat Boi probably works better in Newtone, and the "robotic" sound Newtone can sometimes produce is perfectly fine for a TTS voice.