I’m a backer!
post must be at least 10 characters
Such a vivid voice!
I’m also in.
C’mon, guys. So many of you have been begging for some diversity in the english lineup; here’s your chance.
I’d need an english AI voice asap, which one will be released first : Eleanore Forte AI or Solaris ? Arghh timing is not good. February 2022… I’d need it now ^^ (as a composition tool, to create the lead singer and backing vocals before singing it in a shiny studio)
Since the AI is trained in the style of the Voice Bank provider, it will apply the style of the voice provider.
When Solaris is released (assuming it reaches the monetary goals), that AI will apply Emma Rowley’s style to the track, since that’s what it’s been trained on.
But there’s no need to wait until then.
- The AI code doesn’t use phonemes to make decisions - only the pitch and duration of the note;
- The AI code only writes to the Pitch Deviation Parameter track;
- If you change Voice Banks, the only thing that changes is the interpretations of phonemes.
This means that you can use the AI feature to apply Pitch Deviation style choices of other Voice Banks. There are currently two free Voice Banks that have AI:
- Saki AI (Lite)
- Tsurimaki Maki AI (ENG Lite)
- Koharu Rikka AI (Lite) - limited time free
So download one (or both) of those voices, and use Eleanor Forte (Lite).
- Once you’re ready to apply AI, change the Voice Bank to Tsurimaki Maki AI (ENG Lite) or Saki AI (Lite). Don’t worry if the Voice Bank doesn’t support English phonemes, since they aren’t used by the AI.
- Select all the notes n the track you want to apply AI to.
- Choose Parameters | Auto Pitch Tuning. You’ll be given the option to modify the level of Improvisation and the Random Seed.
- SynthV will write the changes to the Pitch Deviation Parameter track.
- Change the Voice Bank back to Eleanor Forte (Lite). The Pitch Deviation parameters will remain in place, and be used.
- Rinse, lather, repeat.
Keep in mind - AI isn’t magic. Making the track sound more “human” often means making it “worse” in some ways - for example, less on pitch. It won’t make timing changes, and it won’t change other parameters such as Loudness and Tension, which are also important for expressiveness.
What works for one song may not work for another. So use your ear, and adjust the results by re-applying AI where it didn’t work, or make manual changes.
That’s absolutely brilliant ! I wouldn’t have found it without your explanation, thank you dcuny !
Neither. Maki Tsurumaki english is going to hit the market in two weeks.
Have you listened to Tsurumaki Maki English?
Here’s a link to the official demo for the upcoming Voice Bank - I think it speaks for itself.
It’s certainly suitable for a Japanese production that’s looking for some English lyrics, I wouldn’t choose it for use as a lead vocal on an English production instead of Eleanor Forte if the primary audience was native English speakers.
And - back on topic - I’m sure Solaris will also be a go to for “native” English vocals.
That demo (Jingle Bells) was made entirely with an in-development version of Maki JP Standard.
To be extremely clear, that is a demo made with an unfinished Japanese voice bank and in no way represents the English voice bank.
These are two examples using Tsurumaki Maki English, the first is Standard Lite and the second is AI lite:
Thanks for the clarification. I should have read the translation on the YouTube video more carefully.
She may be a little wonky, but I prefer her to Ellie (at least for the time being). Also, she’s available, and phonemes have some options for substitution. Solaris is… starting to concern me.
I think the problem with Solaris is not only the hurdle of how small the market for English vocal synths has unfortunately become in recent years, as well as hesitancy to back a project that might not even reach the goal necessary to produce results.
I’d also say not bringing in a producer for an original song/demo might be contributing to the issue (I can’tsay for sure, would just be speculation), because while “Miles” is great for showing Solaris’ capabilities (as well as giving a good comparrison to the voice provider), and could bring in some of Emma Rowley’s fans, I think it might not completely be reaching an audience who would be more inclined towards purchasing vocal synths. I could just as easily be incorrect about that, however… I can say though, if I weren’t on this forum, I don’t know if I’d even know about Solaris being a thing, which means there’s definitely a limit or something that’s not being overcome to allow information saturation outside of smaller bubbles.
Sorry if I’m just sorta saying obvious things or if I’m off base about the issue…
With the full version of SynthV, substitute phonemes are available for Eleanor Forte as well. I use them all the time.
What surprised me–and this could very well be my being new to this–is the emphasis in the offer (and funds spent on):
- Merchandise: Is that really so important?
- Artwork: Likewise, is that vital at this stage?
- Physical distribution: Maybe because I’m Mac-centric I haven’t bought software in physical form in 10 years and so I’m surprised there’s a critical mass of people who have DVD drives on their computer.
All the above seems to unnecessarily complicate a rather small, fragile venture.
All of that also creates funding tiers that are unattractive to me, and $125 feels pretty expensive to risk on software that might not materialize. What I’d like to see is the ability to fund them any amount (e.g. $50) that would be applied to a coupon of equal amount when the software is published.
Ah good points!
I’m one of those people that likes having physical copies of software if possible, but yeah, it’s definitely not everyone’s cup of tea. Maybe it would have been better for a physical edition to end up as stretch goal?
As for the funding, apparently if they don’t reach the goal, they’re giving themselves 5 months to make up the required difference for the project, and if they can’t by then, they’ll do refunds then. That’s my understanding, at least. Even then though, I think it’s still too risky for some people, and it just leaves them without the chunk of money they put towards it, and then being unsure for quite a while if it was even worth it.
The coupon idea is really clever though!
There are a lot of people who buy physical versions of vocal synths for novelty reasons. It’s not uncommon for people to import boxes from Japan just to have the physical version. Additionally, given the price difference between the digital and physical tiers, they’re definitely not losing money on a $50 DVD.
It’s also worth noting that the merch is rather low-cost items like stickers and buttons, outside of the top tiers where again, even after adding a t-shirt and tote bag (two items which share a production method) they’re definitely still making money from the higher pledge amount. And people are pledging for those things, so just because you aren’t interested in that doesn’t mean nobody else is.
If you have ideas for new tiers, they have actively been asking people what they’d like to see. If you want a “coupon” tier you should make a comment on the campaign page.
I think the problem is that - if you’re really just interested in getting the Solaris Voice Bank - there’s not a whole lot else that can be offered.
Given how small the Eclipsed Sounds team is, I highly doubt that they are doing much on the technical side of Solaris other than the recording of the vocals.
That would explain why Dreamtonics is likely to charge a hefty (but completely undisclosed) sum of money. It would make sense for Dreamtonics to keep a tight lock on their IP, while also maintaining quality control of the voicebanks released for SynthV.
If that’s the case, there’s not much else Eclipsed Sounds can offer that has value to someone who’s only really interested in the vocals.
So I think they’re leveraging the part of the market that is willing to pay extra for something.
On yesterday’s livestream, the Eclipsed Sounds team basically said that even if they fail to reach the target amount, they are close enough to the target that they’re willing to make up the difference out of their pockets to get Solaris completed.
While I’m not sure that’s a prudent choice - because it might indicate a weak market for English SynthV voices - I appreciate the team’s enthusiasm.