Using choir libraries

[Guy Rowland]

East West's original Symphonic Choirs broke the mould. For the first time, you could dictate what a sampled choir actually SAID. It used its own language - Voxos - but there was a sort of English to Voxos translation. The world was our oyster!

People quickly realised that the world, in fact, was not entirely our oyster. It was fiddly to use - it came with a timeline editor (did that come later?) - and with enough patience you got results that edged towards passable. But what I quickly realised was that the chances of actually discerning the words was near-impossible unless you already knew beforehand what the words were supposed to be.

Then choirs like Tonehammer's Requiem came long. They reasoned that the goal of writing your own verse was a fairly impractical one, since all anyone seemed to hear was some vague Latin type stuff. They concentrated on well defined latin word elements. Most people - including I - found that total gibberish was absolutely fine, it was the illusion of movement that was the thing. It worked great for staccato chants, it was quick to get good results, but it very much had its limits. There were some poly-sustains which went through a few pre-baked phrases, but for flowing lines, only the vowels worked really.

As the years marched by, things got more sophisticated with legato transitions and all. I recently bought Audiobro's celebrated Genesis library in an too-tempting offer, and it does sound lovely, They have legato and Melisma modes - with the former you have one syllable until you release all keys, with the latter each new overlapping note advances to the next syllable. You have Latin Syllables, or alternatively you can make up your own using a combination of prefix, vowel and suffix.

For some simple lines, it works well and the tone is gorgeous. But I found myself very disappointed with the actual specifics - suffixes seemed to get buried no matter what I did. It became mushy, and neither legato and melisma modes offered a natural movement within a phrase. Although it can sound lovely in specific circumstances, it's much easier to make it sound lumpy and unnatural.

In some ways, it is a step back from the original Symphonic Choirs where you could program a syllable to last for a period of time before crossing to another. But with its clumsy and fiddly timeline editor, and less than brilliant end results, going back isn't appealing. So you'd be forgiven for thinking that we're stuck. Big staccato Latin chants - great. Smooth oohs and ahhs - wonderful. But natural movement within longer well-defined complex phrases - forget it.

Or perhaps not. Because I do have one library that can actually do this brilliantly, and it has a totally different philosophy.

Fluffy Audio's Dominus Choir is, imo, a work of genius. On the face of it, it's all familiar stuff - there's a bunch of syllables and a word editor. But it works in a totally different way. First, you start with a syllable, and are then offered only choices that will naturally blend with the previous one. Start with a Do, and you only get offered syllables that start with an O, and so on. That's a promising start, but the other half of the equation is that this then plays through synced to the DAW - in other words the notes you play do not trigger a next syllable necessarily, nor do they just hold until the next key is pressed. They advance all on their own, the syllables naturally flowing one to the next. You can change the notes within the phrase and they keep singing the same relative part of the phrase - which is exactly what a real choir would do. All the clever stuff like the legato transitions are entirely handled under the hood without you thinking about it. It's actually quite hard to make it sound bad.

This video explains how it works and super-cleverly uses the analogy of dominos to illustrate Dominus.



The results can be so good that I'd be happy leaving them exposed in a final version. It's such a clever system, so well implemented that nothing else I'm aware of comes close.

Where AI takes us in the future, who knows. Vocaloid has been around for years with increasingly impressive results for solo singers. But if you've been eternally frustrated by trying to have smooth and well defined syllables gracefully transitioning from one to another and sung by a gorgeous choir, there is very good news. We're often bemoaning a tsunami of new products that offer little more than variations on a theme - here is a library that genuinely stands apart.

[Luciano Storti]

Very much agree with this, Guy. I've only recently upgraded to the Pro version of Dominus after years of owning the first entry. You can immediately tell when you play it that it is special. There are still limits I find, and if I seek a quick workflow, it's not always my friend. But it is way better than Symphonic Choirs engine and sounds just lovely. To even have the capability to approach the process in the spirit of notation is great, and easier to follow.

I keep waiting for a proper project to use it extensively, so far it's been in bits and pieces.

[Linos]

How do you write for SATB with Dominus? Is it possible? If I understand correctly, it is only divided into women and men, not soprano, alto, tenor, basso. If you can't write for individual voices, that rules Dominus out for me.

[Luciano Storti]

You are correct, Linos. You can go as far as muting either the female or male choirs and of course double the instance to create 4 parts, albeit with limited ranges. A great way to use the library is to go with polyphonic legato, where the ranges aren't extreme. It blends very well. As mentioned above, I'm still discovering all the ins and outs, so it's possible I missed something.

[Guy Rowland]

All true on all points.

Personally I'd take all those limitations 100 fold to have phrases that actually sound natural and join up. But is there anything else out there I've missed? Is there any other library that can convincingly do this?

[My name is Nobody]

All true on all points.

Personally I'd take all those limitations 100 fold to have phrases that actually sound natural and join up. But is there anything else out there I've missed? Is there any other library that can convincingly do this?

I don’t know who has experience with Strezov’s Arva, Freya and/or Wotan. But whenever I listen to the walkthrough of these libs I find the legatos wonderful. As to it’s word builder capabilities and natural end result……. Is there some one with experiences?

[Thomas Mavian]

Wow. This was on another level for sure, been using Symphonic Choirs quite a bit but as you say it's "fiddly" to say the least. Requiem Light took it's place after a while. Not great but "doable".

Don't know why I haven't looked into this one, since I use and really like their solo woodwinds so the shame is on me.

[Linos]

Ah, tough. I think Symphonic Choirs is the only one with a word builder and separate voices. But it's definitely fiddly and the results are not always very good. If it's important to you what words a choir is singing in your track, why shouldn't it be important what each voice is doing musically? Apparently that's a minority view I have. But I just can't warm up to choir VSTs that don't allow you to write for all the voices.

[Guy Rowland]

Ah, tough. I think Symphonic Choirs is the only one with a word builder and separate voices. But it's definitely fiddly and the results are not always very good. If it's important to you what words a choir is singing in your track, why shouldn't it be important what each voice is doing musically? Apparently that's a minority view I have. But I just can't warm up to choir VSTs that don't allow you to write for all the voices.

I find the polyphonic legato so good (it usually isn't in libraries) that I find writing four parts at once actually works. As Luke says, you can at least break it down to SA and TB. But I do understand it's still not the ideal for us.

Hearing the specificity of the words is actually less important to me than having it sound natural. That's my problem with every other virtual choir library when doing lyrical lines - it always sounds Frankensteinian to me. I play Dominus and it sounds like a choir.

[Linos]

I find the polyphonic legato so good (it usually isn't in libraries) that I find writing four parts at once actually works. As Luke says, you can at least break it down to SA and TB. But I do understand it's still not the ideal for us.

Hearing the specificity of the words is actually less important to me than having it sound natural. That's my problem with every other virtual choir library when doing lyrical lines - it always sounds Frankensteinian to me. I play Dominus and it sounds like a choir.

Wish that we could have both. Alas. Polyphonic legato is not important to me because I don't work by laying down several parts at once on the keyboard. Rather, I write a separate part for each instrument/voice and then create the mockup from that. Current choir VSTs don't offer a solution for this, with the exception of symphonic choirs. If there is a good discount on Dominus, I might get it. But I am not keen on having mixed voices singing each part.

[Lawrence]

Depending on the piece, I still use EWQLSO choirs with the WB. I’ve given up on intelligibility and try to make sure the choir is singing “something authentic sounding” in some language.

Considering that it doesn’t have legato I think it still sounds pretty good, especially the boys choir. I’ve been able to use that a few times to good effect. I’ve recently purchased Genesis during the AudioBro sale, though I have yet to tuck it into a piece.

[Blake Ewing]

Admittedly, I don’t use choir libraries an awful lot, but I wrote a demo for Dominus called “Harbor” and it’s still my favorite choir library to date, especially for those lush close harmonies a la Lauridsen, Durufle, Whitacre, etc.