Hardware voice / speech / song synth?

The Lounge General Discussion
#1

I’ve just come back from a week long intensive singing summer school so this is contrary to the immense irreplacable human beauty of all of that but… I keep wanting to play with speech / voice synths and am sad there aren’t any obvious hardware ones.

I’ve played with Synthesizer V (Synthesizer V | Dreamtonics株式会社) and it’s ridiculous how good this is getting (though still obvs not immense irreplacable human beauty and obviously fiddly considering how powerful it is). Also love the Plogue Chipspeech (chipspeech) stuff - massive fan of anything they do really as Bidule was my gateway into a lot of electronic music stuff and they’ve done no wrong since). Still, I would love something in hardware that I could enter lyrics into.

Everyone laughed at the CT-S1000V (CT-S1000V, singing synth from Casio) but if it came in a desktop module and maybe with an actual hardware keyboard to enter text rather than an app I’d buy it.

I know most of you don’t do words and music while I’m moslty a lyricist / poet rather than musician, but even for the musical part of it it’s a shame it’s largely unexplored in hardware.

I had no idea about the Monomachine doing voice stuff and found this video which I’ve fallen in love with https://www.youtube.com/watch?v=KAsGcsv023M which makes me want to get one but I’m now dreaming of a box that does this and only this and is a little more coherant.

Something that is a little less synth and more speech / singing and allows little multi-part harmonies to be written across an Elektron synthesiser would be absolutely amazing. Especially if you could directly type in lyrics into it (standard Elektron text input is fine but another hardware method would be amazing too). DigiSong / DigiSpeak type box would be so dreamy.

Shall I track down a Monomachine (are there more speech synth demos out there?) Or just use software as fiddly and clicky as it is?

Am I the only one dreaming of this as the next Elektron box? Ha, they could even add that chorus effect all you keep asking about.

Maybe I’ll just sample loads of syllables into my Digitakt.

#2

you could do it oldschool like senor coconut/ATOM and use a sample. sampler all the sounds and create speech by combining them. he goes into it a bit in this podcast:

it’s an inspiring listen and worth checking out the whole thing but he goes into his methods fo speech synthesis a bit… enough to get the idea of what he was doing.

i think that’s the way to go to get something truly unique

3 Likes
#3

there’s also chipspeech (doh! which you mentioned in your post)

and vocaloid

there’s one other but i can’t recall the name of it. it’s sort of along the lines of vocaloid

#4

vocaloid does english but to make it coherent you have to enter phonemes and results can supposedly be unpredictable.

really good if you want to write lyrics in japanese though.

1 Like
#5

After doing a brief investigation of the official site, the start up cost for a single english language voice vocaloid along with vocaloid studio6 is around $350USD. I wonder if the casio keyboard isn’t the better option afterall :grimacing:

1 Like
#6

yeah. another reason to DIY and do it the hard way by chopping up things and looping vowels and stuff.

2 Likes
#7

speak ‘n’ spell plus pro edition.

2 Likes
#8

Po35 is good fun

2 Likes
#9

Gave away my PO-33 but this one does look fun!

Thanks everyone. From a little test run, the problem with sampling is that much more than even acoustic instruments you pretty much have to sample every vowel at every pitch within a range otherwise it sounds funny and not in the good way.

Still think it’d be beautiful if someone put the magic of the Synthesizer V engine (the magic is probably because it’s not real time and doing all sorts of working out when you input notes and words) or even just something with a massive library of samples that works out legato / consonant pairings etc like some of the simpler vsts do, into hardware.

But also think I should stop being silly and just sing and work with singers and stop trying to replace the most human of all things.

Going to keep playing about to see what I can come up with. Also, as it will probably end up sounding like a bad imitation of it, it’s worth plugging one of my all time favourite records (exactly 10 minutes long so I put it on often for short walks when I have to get somewhere on time), Hockets for Two Voices by Meara O’Reilly. The fact that she sung this so purely and perfectly it sounds superhuman is a marvel.

Also found this video where she tries to do it “live” with samples (and slightly fails but it’s one of those things that’s beautiful to see fail slightly as it exposes the humanity of art and singing and how that’s the best part).

Oh, and there’s also my hero Petra Haden who does the opposite (making her voice into every possible instrument).

1 Like
#10

FilipNest:

As far as software, which was not your question :

and rather than reposting, here are links to three posts i’ve made on this. ( I may have made other posts too. )

The Evolution of Yamaha's Vocaloid - #7 by Jukka
Music Synthesizer Technologies made using AI Methods - #21 by Jukka
CT-S1000V, singing synth from Casio - #96 by Jukka

Also look for newer AI stuff too, i’m pretty sure it’s either new or on the way.
More hardware for it is probable but i don’t know of anything specific.
I would look for a follow on from Casio as for instance.

With the price on the Casio CT-S1000V seems like a place to start, despite you keyboard aversions.
Watch the Benn Jordan video if you are interested but on the fence.

Oh and you are not forced to use an app with the CT-S1000V.

1 Like
#11

I love Holly Herndon’s voice resynthesis work, but I think she uses custom software she made

https://youtube.com/playlist?list=OLAK5uy_lcNh3aAR4cANXULZPxOOOW4oxtQQ9t9lI

#12

Yamaha FS1r, but now they are expensive and by all accounts you need a spare decade to work out how to program it.

3 Likes