New AI Virtual Vocalist Gets Into Uncanny Valley Territory

Eclipsed Sounds has launched an IndieGoGo campaign for SOLARIS, a new virtual vocalist powered by Synthesizer V.

Virtual vocalists differ from sample-based vocal libraries by letting you create custom vocal ‘performances’, based on text and notes, vs using ‘canned’ vocal phrases.

Here’s what the developers have to say about SOLARIS:

Similar to how text-to-speech is used to have your computer “read words aloud,” vocal synthesizers let you write music and have your computer “sing” it. This can be a way to draft songs before hiring a real person to sing them, but vocal synthesizers are more often used to feature on songs, giving the songwriter more creative freedom and flexibility not available in a real voice.

Synthesizer V Studio is a vocal synthesizer that excels in realism and expressiveness, especially with its AI capabilities, and offers unmatched opportunities for user control with a focus on ease of use that makes it unique among competitors.

Synthesizer V allows users to purchase vocals of their choosing in a variety of languages, tones, and vocal ranges, including both standard style and AI powered voices. Synthesizer V’s AI vocals are unique in how quickly they process changes, and how realistic their results can be while still maintaining a high level of user control and input.

Note: Crowd-sourced projects involve risk, so check the project site for details.

17 thoughts on “New AI Virtual Vocalist Gets Into Uncanny Valley Territory

  1. Though there is still a hint of artificiality, the pronunciation is much more natural than previous attempts.

    The goal of producing realistic vocals via artificial means will continue to evolve and improve. This seems like a pretty impressive step forward.

  2. Agreed that it reaches the Uncanny Valley. In this case, the artificialness could work in a creative context.
    Interesting that the timeline is longer than most crowdfunded software projects (and many hardware projects). And they specifically state that the dates may change. So they’re not overpromising in terms of timing.

    At 125USD for the digital version, it can be a pretty decent deal for such a tool. Anybody knows if that’s on top of a license for Synthesizer V? Sounds like that’s 80USD for a permanent license. So, 205USD for both?
    https://dreamtonics.com/synthesizerv-gen1/en/download/

    1. It sounds slightly better than the autotune ‘sound’ that became an aesthetic of its own a few years back before receding into blissful obscurity.

      1. receding into blissful obscurity? autotune is prominently in 4 of the top 10 billboard songs right now and like 12 of the top 20. just because you don’t hear it doesn’t mean it’s not pervasive.

  3. The name SOLARIS is trademarked, for use in both musical hardware and software, so they will have to change it before going into production. Any suggestions?

    -Michael Hester,
    proud owner of John Bowen Solaris Synthesizer #12

  4. It’s pretty close though, I still hear a few odd artifacts, also, it’s a bit “low-fi” sounding. Definitely better than that awful “AutoTune” effect that was so popular a while back (good gawd I hated that sound). I don’t know how close this is to the final sound it will output but, indeed, even in this current state, though it’s not perfect but, really not bad either. It would certainly work in backing vocals to be sure and, with the proper mixing tools and effects, I think it could be made to fit in a mix pretty nicely.

  5. The latest version is Synthesizer V Studio Basic (free) and Pro (paid+VST/AU). Basic will run paid vocals, but it’s limited. Scroll to the bottom of
    https://dreamtonics.com/en/synthesizerv/

    There’s a free beta English voice, Eleanor Lite, created with an alternate process–not what Solaris will use!

    You can buy Pro downloads at Dreamtonics’ partners
    http://www.ah-soft.com/product/download.html (multilanguage guest checkout)
    or anicute.com (English translations on site).

  6. Only people who never heard a real singer (without autotune and other effects) can think, that this sound “natural”….

  7. I think our perception of singers is changing. We expect in tune, digitally cut voices to front anodyne, weak songs, and it’s a little frightening to hear a real singer straining and breathing- it sounds unnatural. I have been using VoiceLive for years on my wheedley tenor voice, both performing and recording, and the electronics are industry standard now.

Leave a Reply

Your email address will not be published. Required fields are marked *