PG Music Home Forums General Discussion User Showcase Like Cinderella (Eleanor Forte AI)

Like Cinderella

Female:
The night is calling
I hear the wind that whispers through the trees
Somebody's out there
Somebody who's a dreamer just for me

I will appear
Like in some fairy tale
Waiting to be
Swept off my feet

Like Cinderella
But no fairy godmother
And no glass slipper
Well, maybe
Not much like her at all

Every night
I chase the same siren call
Looking forever for that damned ball

Male:
The night is falling
I hear the wind that's calling out to me (Where can he be?)

Guiding me somewhere
Somewhere that I'm supposed to be (How should I know?)
Will she appear like in the fairy tale? (I'm no princess)
Wish I could sweep her off her feet (My aching feet)

Like Cinderella
And I'll be her Prince Charming (My poor feet)
In shining armor
Not really,
I'm not like that at all (I hate these heels)

Every night
I chase the same siren call
I bet I'll never find that ball
I bet I'll never find that ball


Instruments
1569:Bass, Electric, Soul60sA-B Ev 110
1578:Piano, Acoustic, Rhythm Soul60sA-B Ev 110
1572:Guitar, Electric, Rhythm Soul60sHighA-B Ev 110
1575:Guitar, Electric, Rhythm Soul60sLowA-B Ev 110
3773:Guitar, Electric, Background FunkyGroove60sSoulRiff Ev16 095
~425:Organ, B3, Background Blues Sw 120
Soul60sEv^3-a:Snare, HiHat , b:Snare, Bass Drum, Ride

Signal Chain
Master: Lurssen Mastering Console, Finsher Retro (Vinyl History)
Female Vocal: Greg Wells VoiceCentric
Female Alt Vocal: Sunset Sound Studio Reverb (Plate 2)
Male Vocal: Greg Wells VoiceCentric
Male Alt Vocal: Sunset Sound Studio Reverb (Chamber/Studio 2)
Drums: Finisher Retro (Too Old Tube), Vulf Compressor (Old School Crunch)
Bass: RBass Mono
Piano: EZMix (Piano 2)
Rhythm Guitar 1: TrackSpacer, Retro Finisher (Soul Signal Chain), EZ Mix (Bright Spring Reverb)
Rhythm Guitar 2: TrackSpacer, Retro Finisher (Duo Vibe)
Rhythm Guitar 3: TrackSpacer, Retro Finisher (CEO Chorus)
Organ: EZMix (Organ 1)



The Back Story
For those unfamiliar, Eleanor Forte is a synthetic voice for the program Synthesizer V. The "AI" refers to the use of neural networks to generate the vocal. The harmonic resolution isn't as good as their concatenative synthesis method they have for some of the voices, but I think the synthetic vocal is quite good in other respects.

Fortunately, you don't actually have to understand how any of that works to use the software - you just type words in onto the piano roll notes, and it's sung back in the voice you choose.

I'd pre-ordered the Eleanor Forte AI voice when it was announced back in July, and the vendor sent a download link for people who had pre-ordered this week. I haven't done much songwriting lately, so I figured it was time to have a go at it.

The style _SOL60S2 is the backbone of the song, with rewrite of the chords and the addition of two more guitars. The style of the track reminded me a lot of The Eagles, although the song doesn't. I wanted to keep it short, so I kept it to two verses. I really like putting holds in songs, so I added a couple of those, along with a long hold to end the song. BiaB added an ending tag with an organ glissando, so I knew I had to work that in somewhere. (I ended up using it twice).

I've been intending to write something completely frivolous for a while, so I started with the idea of dancing, with the singer comparing herself to Cinderella, only to realize she had nothing in common. That got the first verse written, but having blown my only idea on the first verse, I wasn't sure where to go. I didn't want to extend the song with a third verse, so having the second verse echo the first verse from the male perspective seemed like a good way to solve the problem.

I don't have a male English voice for SynthV, so I pitched the voice down an octave and adjusted a number of parameters. I wasn't thrilled with the results, so I ended up doubling the vocal with a "Lite" version of Eleanor Forte (the non-AI version). I ended up doing that with the female vocal as well.

I had written the melody so that a counter-line could be added, which I did on the second verse.

A bunch of rewrites later, and the vocals were done. Everything went into my DAW, where I removed chunks of instruments so everything didn't happen at once. I also used TrackSpacer to perform ducking on the guitars so the vocals would be more clear.

Normally, I spend a lot of time manually balancing the vocal, but I wanted to get the first version done quickly, so I didn't sweat it that much.

Since I was going for a '70s sort of vibe, I chose a number of low-fi effects from UJAM's Retro Finisher so I could justify having purchased it.

I'm likely to go back and adjust the mix and even the vocals, so as usual, feedback of any sort is appreciated.

I hope you enjoy it!

Hi David ...

Very interesting! I really enjoyed my listen.

There aren't many songs written specifically for duets these days. That was a nice surprise! AI voices have really come a long way the past few years and you seem to have a mastery of them. Ah ... if only I could have such good pitch! I like the song a lot. I particularly like these lines:

Every night
I chase the same siren call
Looking forever for that damned ball

You are very creative! I read the back story. Although I didn't understand everything, I got a sense of how you approached and structured the vocal elements. Ya did good!

Thanks for something new and different that was enjoyable. best to you.

Alan

Hello David
Good Song
I think the singing is very well done
But it seems to be very complicated
I really have no idea about it
didn't know it works so well
maybe this way I will get a good hard rock singer one day ;-))))
Really cool made song
Many greetings
Andi ;-)

Tight, pleasant and very successful at catching the 70s vibe! I can hear Bowie doing this in his hammy, retro way.

When you say Eleanor "arrived", do you mean the vendor had to create something for you? I was looking at Synth V but was frustrated by the difficulty figuring what was available and how to buy it.

My first impression of Eleanor is that she sounds a little congested, as in, has a cold (hopefully not COVID!) I don't know how to translate that into a useful observation, it's just something I kind of feel in my own chest when listening, you probably understand what I mean better than I do. I want to say she needs more "mouth", but I do not understand the levers here.

Hi Dave,

listening to the song without having read the info I was somewaht irritated by the female voice. But then I read that it was artificial - wow, I had nnever heard of Eleanor Forte AI or the like before, simply amazing! After I had read your info I thought that the software must cost a fortune, but on the Taewanese website they sell it for about 110 bucks - well worth considering a purchase.

Whatever: Thanks a lot for your extensive descriptiom.

Ah, yes, the song : Sounds really good to me, and I think it was the perfect choice for your first go at ER AI.

Thanks a lot for everything,

Stefan

I'm glad to hear that!


Well, it was a surprise to me, as well.

On a forum like this where singer/songwriters do all their own work, there are only a couple people who are able to do duets.


Pitch correction software has come a long was, too. There's a lot more effort being made to make it more transparent and subtle, although people generally don't use it that way. It's really tempting to overdo it!

As vocal synthesis improves, people will first think "That's a over-corrected vocal" instead of "That's a robot singing".


Thanks!

I initially edited that line out because of the (mild) profanity, but eventually put it back in, because it's key to her character. That, and it made me laugh.

Thanks for stopping by to comment! I'm glad you enjoyed the song.

Hi, Andi.

Thanks for the kind words!

It's not that complicated, although Emvoice is easier to use. There are a lot of controls in SynthV. Most of them can be ignored, but they're good to have when you need them. Singing isn't that simple!

There's no English male voice for SynthV yet, although I know they're working on it. From the little I've heard, it's more "pop" than "rock". There used to be an option for adding vocal "growl" as well, but that was removed because it was hard to use and didn't sound that great.

I'm looking forward to hearing what they come up with.

Thanks again for stopping and commenting!

Thanks!


Sorry, I've updated my post to (hopefully) better explain it.

Dreamtonics, the developer of SynthV, have a really awful website.

You can download the Basic (free) program from there, as well as the Lite (free) voices. Just change the language to English and scroll to the bottom of the page to find the download links. In super-tiny letters under the big DOWNLOAD buttons for the Basic version of the studio, there's a link titled Download Lite version Voices. Click that to download the voices.

Once you've installed the Basic Studio, you can install a voice by dragging and dropping it onto the Studio. The Eleanor Forte (Lite) voice is the only free native English speaker.

There are a couple Japanese voices that support English, but they've got heavy accents. They're aimed at Japanese speakers who typically only want to be able to sing short English lines in mostly Japanese songs, and don't care too much about pronunciation.

You can also find the same free program and voices on the ANiCUTE website.

If you're interested in buying SynthV, you'll need to do it from the ANiCUTE website. It's a bit confusing because they have multiple copies of the same item for sale, because the listings are in English, Japanese, and Chinese.

The Eleanor Forte AI voice is still in "pre-release", and I don't know when it'll be "officially" released. At that point, the price will go from $88 to $97.


I'm guessing you're referring to the "male" version of the voice, which does a lot of tweaks to the original voice. The position of the formants have been moved down, and in addition to giving the impression of a different sized oral cavity, it's shifted the position of the nasal anti-resonances, which are usually at a fixed position. So that's going to sound pretty unnatural.

I really wish they'd hurry up and release a native English male vocalist.

Thanks for giving this retro song a listen!

Hi, Stefan.

Thanks!

Since your first impression was there was something wrong with the voice, and not to immediately say it was because it was a robot, I'll call that a success.

I think of these as "demos" anyway - I try to do the best I can with them, but as long as the artificial voices don't detract too much from the song, it's all good.

The voice is pitched a bit higher than I'd usually take a female voice, partly because that was the initial key (although it's trivial to change that in BiaB). But it also seemed that a light, higher voice was a better fit for this style of music.


I'd suggest trying out the free Basic version and voices first.

The Pro Studio is currently $89. Eleanor Forte AI is at the pre-order price of US$ 88, and will go up to US$ 97. There's also an Anri AI voice for pre-order, but it appears it can only be purchased with Eleanor or the Pro Studio.

Keep in mind, if you buy only the Pro Studio, you'll still be able to do additional things like use alternate takes of phonemes in the non-AI Lite voices. The AI voices use a different technology, so that's not an option.

There's also a native American English AI voice coming out in January 2022 called SOLARIA, so that'll be cool.

Now they just need an English male voice, so I don't have to keep morphing the female voices into male vocals...


I'm glad to hear that you liked the song!

Thanks for commenting, and good luck with SynthV if you decide to go that route!

No, I meant the female. She sounds... I don't know, she sounds... plump. She sounds to me like an alto singing a soprano part.



Arggh. Now I remember, this kind of cryptic product status was part of what settled me on Emvoice, even after I figured out how to download the free Synth V voices and played with it a bit. The idea of selling "pre-release" software is a contradiction, unless we're talking about black market leaks. I do want Synth V, but I think I'll wait until Synth V decides it's really for sale before investing the time in learning it.

Thanks for all this info.

It's not that they're selling beta software, it's that the software is being sold before it's available - that it, "pre-release" means that you're buying something before it's been released.

It's pretty common in the virtual instrument world to do this, as it gives developers some cash to work with, and typically rewards buyers with a modest discount.

In this case, they're also rewarding early adopters with copies of the software before it's generally available. That's not a big incentive for me, but apparently it's enough of a driver that they've made it a point of there pre-sales campaign.

If that doesn't float your boat, that's fine. I've also purchased the "pre-release" of Anri and SOLARIA, but that's just because I'll get them anyway, and who knows if I'll have the spare cash available when they are released? Plus, I want to support the development of more/better SynthV voicebanks.

Your preference is to pay for things when they're ready, and that's sensible. Plus, you've already got a pair of voices with Emvoice, so you're only only talking about adding yet another synthetic vocalist to your stable of singers.

Fair enough, David, I don't mean to imply anything shady. I guess I was looking for a sign to tell me what singing synth to get, more as a matter of conservation of mental resources than money, and Dreamtonics' confusing offerings made it harder to choose them.

I liked the song and its structure. The lyrics were good, clever in a few spots. The male voice sounded more natural than the female voice to my ears. But the female voice wasn't bad. So you did a great job reworking the female voice to get the male voice!

David

These voices are getting better. As a backing vocalist they would I assume sound very realistic in a mix.

Interesting technology - the song was a pleasant listen.

Peter

I'm happy to hear the song worked for you. Some things are really obvious (like using a mirroring structure) in retrospect, but sometimes it feel like I blindly stumble on solutions. Songwriting can really feel like fitting in a puzzle - especially when it's setting lyrics to a preexisting melody.


Thanks!

I think I pitched the female voice a bit out of a comfortable range. and it's a bit thinner than I'd normally use. On the other hand, there's only a small range of parameters where the "male" voice seems to work.

Hopefully I'll learn how to create a better performance with the female voice in the future.

I appreciate you listening, as well as the feedback!

Hi, Peter.

I'm pretty impressed with this iteration of Eleanor Forte. While there are some demos that sound very "Vocaloid", I've heard a couple that show what the voice is capable of.


I'm glad the song worked as more than just a tech demo.

Thanks for listening!

A quite cool song, deliberately corny lyrics aside, it's hangs together really well and the drum treatment works really well - though the slap kick isn't improved by it.
The voices - well you did a cool job tweaking them to get a great result. That mob should be paying you demo fees.

First, the song... Love it, David. Cool bass! A very clever write, and the perfect vehicle to use the AI voices. You're right, I don't understand it, but it sounds good. Really well constructed production.

well heck i wasnt expecting this.
us vocallists are gonna be redundant...lol.
great job with the backing band too.
I MEAN IT SINCERELY MATE. this must have been a lot of work. also what i like bout it is, it has some lovely
song "hooks" in the arrangement.
ive only got two words BLOWN AWAY.



best to you.
om

It's good to see you got the ladies together again, and a gentleman too this time. Your work is brilliant, not just because you know how to program the Synthesizer V so well, but because the lyrics and music fit the AI gang very naturally. The only thing you are missing with this production is a animated video of the girls (and guy) belting the song out.

Bravo!