A singing-app question

Dorico
1

I today received an email invitation form Cantai to send them $150 in return for a Lifetime License for a native-mode version of their app that will run on Dorico. Which they are ‘working hard on.’

It’s tempting, but . . .

I think to be safe it would be better to make sure the thing works, first. (By which time the lifetime license will undoubtedly no longer be available for such a price. So it’s a gamble.)

But there’s an even more important question. One which the development team may or not be willing to answer: Is there any plan to offer (‘someday’) a word-singing utility within Dorico itself, in some un-named future version, similar to the way the recent Etude piano works? If the answer is yes, then I will certainly put off a decision on Cantai. (If there is any answer at all, that is!)

Yes, I know there have been a lot of threads about Cantai and related questions recently, but I think what I’m asking is slightly different. (If there is an already-extant answer, perhaps someone would point me toward it. Otherwise the question stands.)

1 Like
2

[Spoiler alert: this is not the extant answer]

As the Dorico team do not tell us about any new features in any upcoming update, I suspect all they will say is they cannot rule it out in the future as they might be equally vague because they might not actually know themselves. (The future is a “continually moving feast” so to speak, or famine, we do not always get what we want (or cannot afford it).)

The Dorico team is very small as you know (even if it is Steinberg as a whole, meaning integrated into Cubase as well.) Developing it would take considerable time (languages, pronunciation, dialects, genres.)

Another question is, when will we be able to upload a musicxml file to an AI process and it be rendered with vocals (if not already), and would it be better than Cantai?

Cantai, as you have been following has a render process you can already try and it has improved over time. It is crafted by Richard and his team. Any AI process in the future may largely be automated (scanning scores then matching it with different performances with different singers). But any “errors” might take longer to be corrected because it is automated. With Cantai (currently at least) we can simply ask the Cantai team to look at a pronunciation of a word or syllable in that language (which is what has been happening if you are able to follow some of the Discord discussions with him and the team.) As you know, because you have been following its development, it is a long process of commitment and dedication (by a human team.)

Perhaps for you, it will come down to, how long you want to (or have to) wait, either for Cantai (which is already making progress and you can evaluate it now, we assume it will improve) or some kind of AI process (presumably many already in development) or perhaps Steinberg itself.

Thinking of Sample libraries, most of us are still searching for our “Ultimate” sample library, running to and from or between NotePerformer, Spitfire, VSL, Cinesamples … the same might be said for what might be eventually hundreds of AI vocal processes for music, perhaps including Steinberg/Yamaha in the future. At some point in the future, we might prefer one AI process for some kinds of music we are creating or arranging, but another for other kinds/genres of music, or perhaps like to mix the singers from two or three different AI programs for our ultimate rendering.

Aside: each performance of any (human) choral or vocal reading is nuanced, notwithstanding the conductor, players, instruments, studio/location all reacting at that time and place (then microphones, recording and mastering engineers…) We find we prefer one recording of a piece to another or like more than one and cannot decide which is the “best”.

What do you want to achieve with the vocals using Dorico—an acceptable demo or output rendering directly for a film score or a pop ballad for release on an album or just for entertainment?
How close does it have to be to something you wish to have heard as a recording … as humanity might have performed it?

3

If I knew the answer to that questions, I would be whole lot closer to knowing what to do about it!

I appreciate you taking the time and trouble to give such a detailed response. I think what I want comes closest to being definable as a ‘demo.’ When someone asks me about a song I wrote, I want them to be able to listen to a rendition of it, including the words. I don’t want them to have to imagine the song by listening to some ah-ah-ah rendition of the notes, and then, by reading a lyric sheet or a lead sheet, understanding how the song actually flows. (Most people aren’t really capable of doing that, even experienced singers.) I have no intention of incorporating it into a film or video nor releasing it on an album as a professionally sung production. I just need (or at least want) something equivalent to Note Performer for orchestral instruments or to Etude for piano. And Cantai, in my opinion, is a very long way yet from either of those.

4

Cubase 15 includes Yamaha Omnivocal, which is a virtual instrument in beta form capable of singing lyrics in a female or male voice:

It is reasonable to assume that this virtual instrument will improve over time and that Dorico may eventually be able to use it.

7 Likes
5

Wow, thank you so much johnkprice. I had never heard of this, but it’s good enough that I should probably learn how to use DAW technology just so I can use Cubase for this purpose. I’m VERY impressed.

Again, thank you.

1 Like
6

I paid Cantai well over a year ago, and even though Dorico has implemented the sufficient ‘plumbing’ to connect to Cantai, there is still no plug-in available that works with Dorico.

7

Maybe it’s also worthwhile to ask yourself the question of how much you like the audible outcome compared to other apps. After John’s demo of Syth V inside Dorico, I took advantage of the Black Fryday deal. And even though I am looking for good choir renditions a la voices 8 and the like, I am quite pleased with Synth V. I am following the progression of Cantai, and the moment I hear a reasonable choir sound that I want to use; I will add it to my arsenal.

Good luck with deciding.

Best, robert

8

Even Voces 8 requires a lot of audio post-production work to sound like Voces 8. :wink:

9

Not sure if you are joking ( I do see the emoji). I will say that having seen Voces8 live, they are remarkably fantastic. Yes, the recordings have post-production work (what doesn’t?), but the raw, unfiltered singing they produce is phenomenal. If you have the chance to hear them, run don’t walk!

10

My mention of Yamaha Omnivocal was not intended to be a recommendation on my part. I have Cubase 15 and have experimented with this virtual instrument, and in my opinion, the output of each instance of Omnivocal sounds more like a pop singer than a church singer. Here is a sample of SATB voices singing the first line of Hark! the herald angels sing. Other than specifying the gender of each singer and setting the singing style to straight rather than dynamic, all of the other controls were left at the default settings.

2 Likes
11

I agree that Omnivocal in the demo video and in your sample just now sound like pop rather than church. At the moment that will serve my needs fine. A church choir would be nice to have, too, but can wait a bit. If Omnivocal were included in the Elements level of Cubase 15 I would spring for it right now. But I have discovered that it isn’t. You have to go to at least the Artist level to get it, and I can’t justify doing that at this point. Still, I really appreciate your telling me about it and I hope it develops quickly, and that Dorico includes it or something like it at some point relatively soon.

As it is, I still can’t get my head around how DAWs work, so would have to learn that first, before being able to use Omnivocal in that setting. (I’m very much a notation-oriented person, which is why I asked in my original post if there was any knowledge of a plan for Dorico to include a word-pronouncing vocal instrument, as their new Etude provides a realistic piano.)

12

The webpage to compare the editions of Cubase indicates that Omnivocal is included in every edition of Cubase.

13

I stand corrected. I was trusting Google, which was wrong, as it often is.

–L3B

14

Thanks @benwiggy :slight_smile: I am not afraid of doing manual work at all… and even if it becomes Voces 6, that would make me very happy already. :wink:

Best, robert

15

As if by providence, I received an email from Cantai, saying that they were working on the VST for Dorico (& Sibelius).

They claim that it will work locally, without passing the data to their server, with a “quick response”.

16

Google search or Gemini? :thinking:

17

Interesting…. did not receive that mail.

18

Me neither…

19

In my young and foolish days, I bought ‘lifetime subscriptions’ to several companies that died - and it’s their lifetime they’re talking about, not yours or mine. My rule of thumb on this: if the product needs to live in the cloud, the lifetime promise is tenuous at best. If it’s something you run on your own equipment, then it’s lifetime of the equipment.

1 Like
20

Google Search.

1 Like