this is more like it - check it out -
http://www.zero-g.co.uk/index.cfm?articleid=802

then click on the one called 'lola is here '.

Funny song! Although it's clearly not authentic sounding, I like it.

Still sounds like singing through a vocoder.

Too digital, too robot like. Yes, if the aim is to create robot-like singing, the product is a success, but if the aim is to create something remotely resembling a real singing human being they have a lot of work left.

It sounds like HAL is sounding singing in 2001 when he is about to be turned off.

Invitia

It was developped in Spain Barcelona and i had contact with a guy named Xavier Amatriain who is a very good musician to. Besides that he has a phd in Object Oriented Models for Digital Signal Processing with a Focus on Audio and Music. You can find him on this page of MTG

http://www.iua.upf.es/mtg/f_publi.php3?authorid=12

he told me the people at iua were busy on this tool quite a while ago

I think they did a fabulous job with this tool but ofcourse a very convincing replication that would be inaudibly undistinguishable from a real voice is still not existing

then again convincing simulations of whole lot of instruments are still not existing.
hence the success all these sampled audio bank generators have when they offer good controllable sample based vsti`s

remco

I actually really liked the robotic nature of the voice. It has a kind of "otherwordly" quality about it that attracts me.
I second what ET says - we aren't even at the stage where a simple instrument like the guitar can be said to be 100% authentically reproduced by synths. The human voice is somewhat more complex than that, so I think it's not a bad job at all.
Doesn't sound anything like Battlestar Galactica

Actually, I thought it's not bad at all..... I can certainly see uses for it - especially for some types of electronica where vocals are not really that important, but you need some kind of hook to please the masses.

I'm not in any way interested in vocals - in fact alot of singers banging on about the usual stuff, i.e "I love you so so so so so much, let me pump your booty babe" or going on about what a tough world it is, nobody loves them, full of angst shite shite shite shiiiiiiiiiiiiiiiiiiiiiiiiiiiiite I hate it - that's one of the many reasons that electro dance and dub and ambient noodlings got me hooked many years ago - no vocals - it was bliss to my sanity.

Anyway - I can see where Vocaloid could be extremely useful - just simple vocal hooks that really don't need to be convincing. Some french accented chick singing "take me higher, gimme E, gimme E, gimme an E Bob, oooh take me higher Bob, ooooooooooh Bob" would be perfect for this plugin.

I don't think anyone is really taken in by advertising blurb, so if you just take Vocaloid at face value - a vocal tool for people that either a) can't sing, or b) don't want to sing....but realise that the masses unfortunately don't consider music to be music without vocal nonsense over the top of a perfectly good composition......(deep sigh) then it could be a god send.

And I enjoyed those demo MP3s - they were quite good. It's easy to jump on a plugin - but considering how difficult it's got to be to get anywhere near a human singing voice, then Vocaloid is a pretty neat toy.

No offense, but as a vocalist, Lola sucks. It sounds like the speech synthesizer on my 1993 Macintosh 660AV. Honestly, I'm not sure, judging from the samples, that Lola can do anything you can't do with the Fruity Speech Synthesizer and Fruity Vocoder.

But perhaps I'll withold judgment until some really talented programmers get a hold of Lola's pipes.

Quote:

"I love you so so so so so much, let me pump your booty babe" or going on about what a tough world it is, nobody loves them, full of angst shite shite shite shiiiiiiiiiiiiiiiiiiiiiiiiiiiiite

Quote:

"take me higher, gimme E, gimme E, gimme an E Bob, oooh take me higher Bob, ooooooooooh Bob"

Dunno what you're on, but I want some...

God, why are the demos that are a year old more impressive? Ro-Bot is hee eeree! That is terrible. NOW, this would be perfect if I was doing a movie soundtrack that took place in space at some space station, and some weird freakified robot was singing. THEN it be believable. But that's the only place I could envision me using it.

But let's be honest people, the public is picky, ESPECIALLY when it comes to vocals, because lyrics is everything to them, even if it makes you ill Kritikon. The words, they relate to them, so it gives it meaning to them. That's what matters. And if it sounds like some damn robot singing to them, how are they getting that 'connection' of another human 'speaking' to them. I'd rather have a partially bad real vocalist than a robot anyday though. But then again, Lola wouldn't have an attitude, would show up on time, and wouldn't demand front stage either.

Oh, and I LIKE a good vocalist for electronic music. This wouldn't cut it.

Devon

Quote:

I'd rather have a partially bad real vocalist than a robot anyday though. But then again, Lola wouldn't have an attitude, would show up on time, and wouldn't demand front stage either.

There are plenty of those around mate..... and they seem to make a good living out of it too - beats me.

This makes me wonder if this is what it sounds like
if you would sample all the letters and then piece
them together to make words and then run it through
a vocoder.

//Daniel

In a way it's like computer graphics. When will we see characters that look just like humans? Some animators get very close but it's still lifeless I feel.

Same for vocal synthesis. Think of how many mouth movents there are and you'll get the picture.Then there's the tonge and throat and lungs topped off with emotion .

It will get better but to get the actual performance
you want might be tough.

I would love to have somthing like a vocoder where you can pick a voice. Sing into the mic and have someone elses voice come out. This would be hard in realtime but even if you could record your crappy vocal and hit ......replace with lola ect.

IT would analyze every naunce of your voice and replace with a new singer.

Maybe one day.

Mush

Have you guys had a chance to look at

http://jasminemusic.net/vocaloid/index.htm

Here you will find dense midi controller streams that can drive voice generators like the vocaloid. The mp3 examples though have been created using the Yamaha VL physical modeling generator. In fact, I think this page is a good demo of the VL capabilities. You should listen to the mp3s and also look at the midi files to see what's in there!

Surya

ps: I think Jasmine is working with Yamaha on the vocaloid, and might have something to show for it soon!

this sounds as bad as the other one. the only difference is there's less syllables to confuse the fuck out of us