So, I want to make a death metal vocal library - any advice?
-
- KVRAF
- 2448 posts since 12 Sep, 2004
Oh interesting… vocaloid uses some form of resynthesis, IIRC… perhaps the metal voice is just another parameterization of their voice model, and there's no post-processing per se. That would kinda make sense. The hard engineering question would be then: how do you sample an arbitrary voice, break it down into a spectral model, apply your death metal transforms (a la vocaloid???), and resynthesize. And if it's that easy, why doesn't vocaloid just operate on arbitrary voice samples, rather than a synthesized voice? With all the voice R&D over the years, I bet the necessary techniques are already out there (and appropriately patented)…
You need to limit that rez, bro.
-
- KVRAF
- Topic Starter
- 2134 posts since 20 Sep, 2013 from Poland
I think getting Hatsune Miku to do death metal is mostly a matter of processing Vocaloid's cleanly singing output afterwards, but I'm trying to find out for sure.
I guess that's what makes it really interesting. There are different approaches to sequencing singing. No idea which will work best and which will work worst for death metal. At worst I'll end up with a library of growly vowels.
I guess that's what makes it really interesting. There are different approaches to sequencing singing. No idea which will work best and which will work worst for death metal. At worst I'll end up with a library of growly vowels.
-
- KVRer
- 6 posts since 13 Jul, 2015
check this video out for some processing ideas for metal vocals
https://www.youtube.com/watch?v=HGvN3DX ... e=youtu.be
https://www.youtube.com/watch?v=HGvN3DX ... e=youtu.be
-
- KVRAF
- Topic Starter
- 2134 posts since 20 Sep, 2013 from Poland
Hi! This project is still on, we're now going to do it for Plogue Alter/Ego (another piece of software I wrote a big long BPB article about), and hopefully will record Friday or Saturday. Would have recorded a few weeks ago but my vocalist caught a cold, had a gig the day after we were going to record, and we didn't want to wear out her already sore throat. It's going to be interesting to see how Alter/Ego's pitch detection algorithm handles growls.
In the meantime, we did release this simple pack of phoneme and pre-drop samples, and recorded raw material for a voice bank for UTAU but I never got around to actually installing UTAU - the whole Japanese locale thing and its general awkwardness for English vocals made me put it off long enough that Alter/Ego got released.
In the meantime, we did release this simple pack of phoneme and pre-drop samples, and recorded raw material for a voice bank for UTAU but I never got around to actually installing UTAU - the whole Japanese locale thing and its general awkwardness for English vocals made me put it off long enough that Alter/Ego got released.
-
- KVRian
- 1207 posts since 16 Sep, 2006
Big fan from way back (Suffocation, Dying Fetus, Death, Nile, Deicide, Corpsegrinder-era Cannibal Corpse) and this is a very interesting idea to me.
Make a lot of sub vocals (Effigy of the Forgotten, Killing on Adrenaline) please!
Make a lot of sub vocals (Effigy of the Forgotten, Killing on Adrenaline) please!
Ha ha suck it!
-
- KVRian
- 1367 posts since 30 Jul, 2013
-
- KVRAF
- Topic Starter
- 2134 posts since 20 Sep, 2013 from Poland
A few months and four recording sessions later, we now have all the necessary raw material recorded. I'm cleaning it up now and getting the volumes consistent, then the formant-tagging work starts.
Edit: no, not really. It turns out we'll need to record one more round of corrections. Heh. Not surprised.
I wonder what commercial uses this might have outside of music. Monster voice sound design, certainly.
Edit: no, not really. It turns out we'll need to record one more round of corrections. Heh. Not surprised.
I wonder what commercial uses this might have outside of music. Monster voice sound design, certainly.
-
- KVRAF
- Topic Starter
- 2134 posts since 20 Sep, 2013 from Poland
Got the first working prototype working. Sounds scary and growly, though high notes are actually kind of cute, in a cute demon way.
- KVRian
- 943 posts since 11 Dec, 2014 from one rocky mountain to another
Cute demons are where it's at. Looking forward to this lib.
Instead of exhaling to let out a scream. Try inhaling instead. If done right you can get a wicked pig-type squeal.
Instead of exhaling to let out a scream. Try inhaling instead. If done right you can get a wicked pig-type squeal.
-
- KVRAF
- Topic Starter
- 2134 posts since 20 Sep, 2013 from Poland
Oh, I'm not using my voice - I got a girl who's a real death metal vocalist to record all the material. Even though she knows what she's doing, we've still run into some weird things, like how do you growl "ffffff"? You can't really growl it, you can sort of go "ffffff" a little meaner and more aggressive than normal...
It's interesting, that's for sure.
It's interesting, that's for sure.
- KVRian
- 943 posts since 11 Dec, 2014 from one rocky mountain to another
-
- KVRAF
- Topic Starter
- 2134 posts since 20 Sep, 2013 from Poland
There's a lot of stuff like that for transitioning between "f" and other sounds, but for a long sustained f that can be looped...
You just discover funny stuff along the way. Like we have a whole lot of equally harshly recorded vowels, but some of them end up cleaner than others when resynthesized, as if the engine is better able to find and follow a pitch in them. I need to figure out why and get everything more consistent, hopefully without rerecording extra sustained vowels.
You just discover funny stuff along the way. Like we have a whole lot of equally harshly recorded vowels, but some of them end up cleaner than others when resynthesized, as if the engine is better able to find and follow a pitch in them. I need to figure out why and get everything more consistent, hopefully without rerecording extra sustained vowels.
-
- KVRAF
- Topic Starter
- 2134 posts since 20 Sep, 2013 from Poland
Sure! She was finished for the KVR DC last December, and can do black metal reasonably well with some parameter changes.gorgorsan wrote:Hi DSmolken, any news on this as I am very interested, also will black metal style be included as well?
http://www.karoryfer.com/karoryfer-samp ... /marie-ork