AI vocals, my first steps with Synth V

Dorico
,
1

I’m a recent convert to Synth V for creating vocal mock-ups, and I have to admit that I’m genuinely impressed (and just a little alarmed) by what it can achieve in such a short time. While it doesn’t replace the magic of a live singer breathing life into a phrase, tools like Synth V fill a surprisingly valuable niche. They give composers and arrangers an expressive, cost-effective way to hear their work come alive, without having to beg favours from singer friends or sell a kidney for studio time.

Of course, there are valid concerns about AI replacing professional musicians. And I completely agree: nothing can match the artistry, intuition, and spontaneity of a real performer. But I’d argue that, used thoughtfully, AI isn’t here to steal jobs, it’s here to support the creative process. It helps bring music to life earlier, allowing ideas to be shaped and refined before they’re handed over to live musicians.

For the curious, here’s the setup I used to produce the attached song, “Catch My Mother’s Smile”:

  • Solo Vocal : Hayden 2 through Synth V Pro 2
  • Flute, Piano & Strings : Cinematic Studio Series via NPPE
  • Percussion : BBSO via NPPE
  • Drums : Superior Drummer 3 (huge thanks to Marc for the excellent percussion map!)
  • Bass Guitar: Electric Bass (Steinberg)
  • All other instruments : NotePerformer
  • Software : the incomparable Dorico, and Logic Pro (to bounce stems from Dorico and Synth V)

This workflow has opened up new creative possibilities for me, and dramatically shortened my list of excuses for missing deadlines. I’d highly recommend it to anyone looking to inject realism, nuance, and emotional depth into their demos or mock-ups. It’s such a pity NotePerformer Playback engines have been discontinued, as I love the results that can be achieved, especially with the Cinematic Studio Series Strings.

For info, the song is part of a new musical I’m developing with lyricist Rob Messik, titled The Ballad of Raggedy Alice.

PS I know the vocal shaping could be refined further, but the real joy of tools like Synth V (and NotePerformer) is the immediacy of results. I’m doing my best to resist the gravitational pull of endless tweaking, because there’s still plenty of composing to be done and only so many hours in each day.

(The file is too large to upload, and the DropBox sharing facility produces an error when embedded in a message here. This link will allow you to download the file.)

https://www.dropbox.com/t/Fb71WcocX8X87JQn

4 Likes
Viable AI to convert Lyric to AI singer - for initial scores?
2

Same. Extremely impressed and it’s only been a few days.

3

I was surprised to learn that some of the voices I had ordered a few years ago were only subscription and had expired.

1 Like
4

Interesting. Nothing during ordering nor in the confirmation said anything about a subscription. At ~ $$75/voice, one assumes this is an outright purchase.

5

Must have been an error. When I tried to load Synth V (v2) into Cubase this morning, I started getting messages about voices that were marked “unsubscribed” with a date a couple of years ago. Checking the Dreamtonics website this evening, I did not find any info on that or on any of my voices having lapsed. So whatever I saw was a mystery.

6

was curious about this but the link has already expired. is it possible to renew it?

7

Yes, same here… Would like to hear it too.
I am interested in either using this , or Cantai to get a somewhat better feedback from what I’m writing.

8

Here’s a quick mock-up with Natalie 2, Sheena 2 and Liam.

Guitar is Ample Taylor using the riffer with some touch up in Studio One. I didn’t do much with the voices except some obviously glaring pronunciation errors. I tried to use Dorico’s MIDI export with lyrics but it didn’t quite work in some spots (timing and placement mostly) so I just typed the lyrics into Synth V pro 2.

EDIT: Sorry, I didn’t use the Ample riffer, it’s Dorico’s notation with lengths touched up in Studio One.

2 Likes
9

Apologies — the previous link expired. This one should work now:

Catch My Mother’s Smile (MP3)

I’ve been gradually becoming more comfortable with Synth V, and with time and careful tweaking, it’s possible to create natural, expressive performances.

That said, I do wish the software handled English phoneme timing a bit more intuitively. I often find myself manually adjusting certain sounds, for example, shifting “th” slightly ahead of the beat, or delaying final “s” sounds to land just after it, to better reflect how a human singer might phrase them. Another recurring issue is the overlap of consecutive “s” sounds when one word ends in “s” and the next begins with it. Without editing, this can easily sound artificial.

Synth V has impressive potential, but it still needs a human touch to fully capture the phrasing and nuance of sung English. While I don’t expect AI to match the depth, colour, or expressive range of a human voice anytime soon, what Synth V already offers is both remarkable and creatively inspiring.

10

Impressive indeed. Thank you for sharing this. :slight_smile: It seems to me that Synth V is great for pop and other popular styles, and I’m surely interested in using it for that. But at the moment I am working on a choir piece (satb) and am looking for something suitable for that. Think of Voices 8. Its unclear to me what’s happening with Cantai, so maybe AceStudio might be an option. Best, robert

1 Like
11

I’m very impressed.

Btw. We have a dedicated forum for AI and Music Notation:

1 Like