The Choir Project Development/ Help

[sorohanro]

Hi everyone
As you can see, I'm trying to push forward the usability of Mu.Lab, from rock to jazz and "soundtrackish" music. Now I try the next step, a choir, suitable for classical or film.
Don't think of it to be an "epic" EWQL choir, but more as a vocal ensemble.

There are several good choirs for free, my favorite is PapelMedia "Ahh Choir", but as you can see, almost all choirs have just one option... the Ahhh.

What I try is to have different vowels that can be changed by:
- velocity
or ...
and here comes your imagination/ help.
One idea can be the modulation wheel but I don't know how to do that

--------------------------------------
So, for now I will upload just few voices and just two vowels (A and E).
In the upcoming pack will be A, E, I, O, U and more voices, but in parallel I would like to see some ideas of how you think would be best.


--------------------------------------
1 - I would like a solution to make the transition between vowels smoother.
2 - All samples were recorded at 32 bit float. Should I make the samples at 32 bit ? 24 bit ? 16 bit for saving space ?
3 - Should the patch contain a reverb ? (in MUX) or should I let the choice to each user ?

[robenestobenz]

the papelmedia ones are fairly Lo-fi sounding as it is, so I wouldn't use anything higher than 16 bit. Reverb, I wouldn't use it, but it's probably best to include it as standard - it's eaay enough to disable anyways.

re smoothing vowel transitions, I can't think of any non-plugin solution except for crossfading, which might have CPU implications.

[sorohanro]

I use quite often Papelmedia choir. The only thing I dislike is the small choice (actually just one choice) of vowels.

So, maybe there is a possibility to put several multisamplers and crossfade between them ?
Soon I'll upload the samples and one patch.

[sorohanro]

http://www.mediafire.com/?anuyv11086bqttq

Here, a multisampla patch demo:
- 16 bit wavs (thefinal version will be on 32 bit maybe)
- just 2 vowels (A and E)
- velocity triggers different vowels
- no looping points at this moment
- some minor mistakes in mapping
- not all the voices at this moment (will be more altos, baritones and tenors... dunno yet about sopranos, maybe...)

[robenestobenz]

Cool, thanks for this -- downloading now.

One point though -- I just checked inside the Papelmedia soundfont and in the version I have (20.4mb) all wavs are 16-bit 44.1khz, so 24/32 bit would just be a bigger file.

Also, might be cool to rig up the vowel change to the modwheel, if that was possible.

[sorohanro]

I was originally thinking to make some attacks (like LA, LE, LI, HA, HE, HI, KA, Ke, KI... etc) but that would "eat" too much space and work.
Still, linking the attacks to velocity and vowels to mod wheel, you would easily make some words even.
So... maybe in the future...

[sorohanro]

So... I see 6 downloads.
Any thoughts ?

[robenestobenz]

So... I see 6 downloads.
Any thoughts ?

Didn't get a chance to try them sorry! Will do later.

[TheGuysanIdiot]

Downloading now.

Will look at it tonight and I will comment sometime after.

OZ

[sorohanro]

Thanks guys.
Let's try to find a solution to make a smooth fade between vowels and we'll have quite an interesting thing

[TheGuysanIdiot]

I have seen a few choir type VSTi.

Most seem to have one preset that plays the "Ahhs" spread across the keyboard.

Another preset will have the "Ohhs" spread across the keyboard.

And another preset with "Ohhs" at bottom of keyboard and "Ahhs" at the top of the keyboard.

And so on.

So if you want "Ahhs and Ohhs" you need two copies of the VSTi / Multisampla loaded etc.

If you increase the "Release Time" in Multisampla's AMP ENV to say 600ms the note sounds after the key is lifted so the voice is not cut off and blends into the next voice more naturally.

I cannot see how the modulation wheel can do what you want. It is normally linked to a LFO to create wobbles.

The above is a simple way to do choirs.

I believe you are trying to do more of a speech thing. Almost a talking choir. You probably would need to do this within a specifically programmed VSTi.

Jo may have ideas on how MU.LAB could do this but we would have to wait for his return from holiday.

Anyway the above is my thoughts so far.

Sorry to be of not much help on speech stuff.

OZ

[TheGuysanIdiot]

A quick 10 minute mess about with your test choir using both A and E samples.

http://www.mediafire.com/file/qlpdfe8ua8egu0f/Choir.mp3

OZ

[sorohanro]

Thanks OZ for the feedback.
It will be also a Kontakt patch and that will have the vowels changes linked to mod wheel. I don't know how to do thin in Mu.Lab, or maybe using other controls.
You're right, I try to make sort of talking choir.

Interesting track, like the Vangelis sort of feel.

[robenestobenz]

Cool patch, but the change from Ahh to Ehh is too sudden.

I can't work out how to do a crossfade with the in-built modules though. With an external MIDI plugin (either NicFit's MIDIMunger, or InsertPizHere's midiConverter2 (multiplatform) would do) setting up either velocity crossfading or the ability to scroll between various vowel sounds with the modwheel wouldn't be a problem, but this loses the convenience of having everything self-contained.

[sorohanro]

yep, you nailed my problem better than I could.
So... any other ideas ?
Maybe a 4 separate multisamplers in a MUX and a crossfader with 4 positions that mix between them ?
That could be possible but kind of eats a lot of CPU probably.