Synthesizer V - Vocaloid haters might want to check this

[BDT]

This was originally supposed to be my own test of the new audio to vocal conversion feature, but I was a little underwhelmed with the results, I guess you shouldn't force a woman to sing male vocals.
In the end it was MUCH easier to just let Solaria do her own thing.
She's definitely good enough for that
Enjoy:

 

[AceAudioHQ]

Things are moving so fast at the moment that I've halted buying anything for a while, Udio generates such realistic vocals for songs that I'm just waiting for more customization options or another program that would combine everything other programs do

 

[chapbot]

Things are moving so fast at the moment that I've halted buying anything for a while, Udio generates such realistic vocals for songs that I'm just waiting for more customization options or another program that would combine everything other programs do

I was thinking the same thing! Those Udio voices are so amazingly real that I've thought Synth V better up their game real fast.

 

[David Cuny]

I was thinking the same thing! Those Udio voices are so amazingly real that I've thought Synth V better up their game real fast.

Unlike Udio, Dreamtonics appears to be concerned about training their systems on copyright material.

This is a similar problem for other VIs. For example, Mike Greene just sunk a ton of money into paying singers for his choir VI. But someone using AI can just rip vocals from existing recordings to train their AI.

Literally every element in AI depends on copy someone else's work. A violin solo comes from copying and recombining prior performances. AI doesn't create new source material that it can use. When AI mines is its own materials, it creates a copy of a copy, like a snake devouring its own tail.

For AI to have value, it depends on the very thing it destroys.

 

[J. H. Smith]

@ AceAudio and chapbot: We already had a lengthy "discussion" about shiny AI software that does your songwriting for you a couple of months back.

IIRC, the conclusion was to take that subject to another thread, since this thread is about Dreamtonics Synth V, which is a fundamentally different kind of software. Thank you.

 

[madfloyd]

Quick question: Would it be accurate to say that Synth V still cannot handle tempo changes?

 

[madfloyd]

This was originally supposed to be my own test of the new audio to vocal conversion feature, but I was a little underwhelmed with the results, I guess you shouldn't force a woman to sing male vocals.
In the end it was MUCH easier to just let Solaria do her own thing.
She's definitely good enough for that
Enjoy:

Great job, I usually just listen to these covers to get a taste, and then move on, but I was totally drawn in with this. A lot of beauty in the singing here. May I ask what settings you used on Solaria?

 

[mothershout]

Quick question: Would it be accurate to say that Synth V still cannot handle tempo changes?

No… it’s always handled them, if you add them manually. The ARA plugin (though rather buggy) will handle them automatically.

 

[madfloyd]

No… it’s always handled them, if you add them manually. The ARA plugin (though rather buggy) will handle them automatically.

Hmmm, add them manually.... where? When I insert one in my DAW (Studio One) Synth V does not follow...

 

[mothershout]

Hmmm, add them manually.... where? When I insert one in my DAW (Studio One) Synth V does not follow...

You add tempo and/or time signatures like this: https://svdocs.dreamtonics.com/en/synthv/basic-usage/tempo

If your project has one fixed tempo, then alternatively you can use Project/Synchronize Tempo With Host.

 

[madfloyd]

You add tempo and/or time signatures like this: https://svdocs.dreamtonics.com/en/synthv/basic-usage/tempo

If your project has one fixed tempo, then alternatively you can use Project/Synchronize Tempo With Host.

Oh wow, and time signatures too. I was not aware of this, thank you!

 

[mothershout]

While we're on the subject of tempo... I've found the ARA plugin to be pretty much unusable in Logic. When the ARA Bridge is enabled, tempo and time signature changes can't be added in the SynthV plugin (which makes sense), but it regularly reverts back to 120 4/4 when I start playback, even when the bridge is active.

Turning off the ARA Bridge once it's been set up is difficult - it kept re-enabling itself when the project's re-opened. I was eventually able to turn it off completely by adding the plugin on a dummy audio track, turning off the ARA Bridge in SynthV and then deleting the dummy track...

(I spoke too soon - it's back to enabling the damn bridge on project open, even though there is no ARA Bridge plugin anywhere in the project).

Logic is supposed to be supported, but the notes in the manual say x86 macOS, Rosetta only. I'm on an M1 Macbook, running Rosetta, so it's possible that the issues are related to using Apple Silicon.

 

[Anthony]

I really wish Dreamtronics would enable their DAW plugin to automatically detect tempo and time signatures changes the way that the emvoice plugin does.

I regularly change both and having to manually enter them is a real drag.

 

[BDT]

Great job, I usually just listen to these covers to get a taste, and then move on, but I was totally drawn in with this. A lot of beauty in the singing here. May I ask what settings you used on Solaria?

Nothing fancy, see screenshot.
Most of the realism comes from the carefully chosen prosody, which syllables are emphasized, which are not, which are sung fast, etc. etc.
Prosody is a science in itself and can make or break a song.
I'm still learning.

Solaria settings for this song:

 

[madfloyd]

Nothing fancy, see screenshot.
Most of the realism comes from the carefully chosen prosody, which syllables are emphasized, which are not, which are sung fast, etc. etc.
Prosody is a science in itself and can make or break a song.
I'm still learning.

Solaria settings for this song:

Thank you. Indeed, prosody (a new word for me, had to look it up) is where most of magic happens.

 

[odod]

I am beta testing this contender .. I know it;s different, but it's quite interesting because it's realtime

 

[David Cuny]

I am beta testing this contender .. I know it;s different, but it's quite interesting because it's realtime

Vocaflex is also realtime. But I believe that it's a VSTi.

I can even change my gender and speak in a woman's voice.

Obligatory "I can't imagine anyone abusing that" comment. On the other hand, maybe finally woman gamers can use it to sound like teenage boys and finally avoid being treated like clueless noobs, so...

I can even sing in this voice on stage

Beatles cover bands rejoice!

All this AI stuff seems to be moving in the same direction:

 

[J. H. Smith]

Quote from a recenly mentioned AI songwriting site:

"AI has the potential to … enable anyone to create extraordinary music."

A great example of 'contradictio in adiecto'.

 

[richiebee]

Quote from a recenly mentioned AI songwriting site:

"AI has the potential to … enable anyone to create extraordinary music."

A great example of 'contradictio in adiecto'.

Yes, much in the same way that the radio has enabled so many air guitarists over the years.

Seriously though, it does have the potential. It just a question of whether people will use it that way, or whether they will take the easy route. I've made better songs since I started using Synthesizer-V. I've even finished some of them.

 

[SaRaBand]

I've made better songs since I started using Synthesizer-V. I've even finished some of them.

But Synthesizer V still doesn't write the songs for you. It just adds to your set of tools so that now you can do more than instrumental arrangements and complete your songs with a voice that sings your melodies...