Forum

Please consider registering
guest

Log In

Lost password?
Advanced Search

— Forum Scope —




— Match —





— Forum Options —





Minimum search word length is 4 characters - maximum search word length is 84 characters

The forums are currently locked and only available for read only access
JMT: Could LEON be an opera singer?
Jun 3 2004
06:08
alexeiustinov
Member
Members
Forum Posts: 34
Member Since:
Apr 8 2004
Offline

Hello Vocaloid community!

Please check out our new example – Andrew Lloyd Webber’s “The Phantom Of The Opera”, in which we continued our experiments in expanding the range of singing sounds and made an attempt to give LEON academic sounding – singing with a stable Vibrato and quite strong pressure. As always, no manual editing at all! See details at [url:160wkj9s]http://www.ntonyx.com/vocaloid/06-02-2004.htm[/url:160wkj9s]

YV Enhancer 1.0 questionnaire is out. Please take a minute to fill it in now!
Thanks a lot in advance.

Alexei Ustinov,
CTO, JMT

Jun 3 2004
06:12
alexeiustinov
Member
Members
Forum Posts: 34
Member Since:
Apr 8 2004
Offline

[quote:1fic6q0h]andromeda said (in a song comment)

Now come on Alexei!! After doing what I thought was a good attempt with a demo a few days ago, this is two steps backward. The vibrato is ridiculous and the sound of the voice in the lower range at the end is just comical. Surely you cannot be serious?[/quote:1fic6q0h]

I think all that the Vocaloid users (including us) are doing, cannot be recognized as something serious if we talk about art, aesthetics or human music. However, we as researchers and developers must be serious in our efforts to achieve more realistic result in obeying “physical laws” and “intonation norms” of singing.

BTW, some time ago I tried to check out a number of real singers to sing my own song – academic, pop, amateur. They applied a lot of singing techniques, but nobody could bring the sense that I needed. So I had to sing this song myself (I’m not a singer!), and it was the best variant!

Well, I think we can’t talk about understanding “sense of lyrics” in case of Vocaloid at all. (This is the first thing that we appreciate in a singer. Maybe, it’s even more important than beauty and strength of a voice.) But we can try to achieve such control, when Vocaloid follows “physical laws” of singing, and doesn’t break them every time. So, as I have written on our pages – now we only show that Vocaloid can be controlled more effectively, than but hard mouse drawing. And I hope, with YV Enhancer any user will be able to achieve better result in shorter time.

So, I completely agree with you. Thanks!!! :D

Jun 3 2004
06:26
andromeda
Member
Members
Forum Posts: 188
Member Since:
Feb 27 2004
Offline

Dear Alexei – you are so kind to me. I constantly criticise yet you are always so generous in your response. You are doing excellent work on Vocaloid – I just don’t think Leon will be the next Pavarotti. Maybe when Mr Pavarotti retires, he can offer his voice to Zero-G for sampling. <img class=” />
Best wishes
Chris

Jun 3 2004
07:51
bobyordan
Member
Members
Forum Posts: 88
Member Since:
Feb 28 2004
Offline

This was really scary! *joke*

Some long notes sung by Christine, had a good opera feel on the a-words. :)

Both voices sounded like they where processed with heavy effects. But I like it, hard to get this effect with any other software. I guess George Lucas perhaps could be interested in having R2D2/C3PO could make an opera apperance in the next Star Wars movie? ;)

The voice could use more of a morph filter characteristics to sound more realistic? I hear some achievements goin in that direction.

The lyrics was a bit harder to hear on this demo, though.

[b:1kj4eqow]I wonder if you could add a feature in the YV enhancer, that can tweak songs, by analysing the text also? PM algoritms that produces different results according to the lyrics content. Kind of phonetic sensitivity. [/b:1kj4eqow]

I also noticed when I sing a melody self, it gets more of the feel that I sing notes and attach words to them. When the vocaloid sings it sings words with different notes. The difference is asstounding. The flow of the melody can not be achived with the way the vocaloid kind of spits phrases. Is it possible to enhance the switching of the notes so that it creates a continues flow in the melody line?

Cheers
Bob
:roll:

Jun 3 2004
10:08
robotarchie
Member
Members
Forum Posts: 223
Member Since:
Feb 27 2004
Offline

Alexei – I’m no academic (my student days are long gone), so I struggle a bit with the terms of reference/language of research – forgive me, but I wonder about this “sense of lyric” reference you’ve made here. Could you indulge me and explain a bit more what you mean by it? I’m also a little confused by the remarks about art, aesthetics and human music. Are you saying that they are irrelevant or redundant references to “computer music”?

Jun 3 2004
10:55
alexeiustinov
Member
Members
Forum Posts: 34
Member Since:
Apr 8 2004
Offline

[b:39e9swc3]Dear Chris, Bob and Archie![/b:39e9swc3]

I am thankful to you for your criticism and suggestions and I consider you the most sensible members of this jam.

[b:39e9swc3]Chris a.k.a. Andromeda.[/b:39e9swc3]
If I was Pavarotti, I wouldn’t give my samples for Vocaloid, the same way as if I was Schwarzenegger I wouldn’t give my 3D visual image to Hollywood, not mentioning the mass market, because there’s no guaranty that these samples (or images) would be used correctly. If in some future time PM (or something similar) becomes OBLIGATORY part of the synthesis, limiting “character of movement”, as well as “used vocabulary” and “reproduced sense”, such world-famous samples could be used by anybody. And nobody will be able to manage it in improper way.

[b:39e9swc3]Bob[/b:39e9swc3]
You are talking about very important things! But let us make our first step!
We are not Yamaha, we are not even Zero-G, to make high-tech products in a short time :-( but of course we realize the relevance of these things.

[b:39e9swc3]Archie,[/b:39e9swc3]
Accept my apologies for the language. BTW we try to go around things which are not simple to discuss even for native speakers :-(
“sense of lyric” – I want to say, that the same words can be sung in a lot ways. And “sense of words” is not always transferred correctly (convincingly) by a live singer. So could we hope to get some sense form a robot?
“Computer music” – it is not important what way do you create music. Only a human being is a “recipient”. And this “recipient” possesses certain peculiarities of reception, SO if we want our music to be listened to we have to follow the laws of reception. This way I think electronic music inevitably subdues to art and aesthetic experience of previous generations.

I have to leave our forums for a week, because I am going on a business trip today, however I’ll return to you soon with some news.

Thanks a lot again!

Jun 3 2004
17:25
robotarchie
Member
Members
Forum Posts: 223
Member Since:
Feb 27 2004
Offline

Crikey…. run off and leave me none the wiser for a week, why don’t you? Come to think of it, I’ll probably need the week to get my head round this latest reply. Is it just me or is the air around here getting a bit rare….? Maybe I need a holiday (oops – I mean business trip). Speak to you when you get back – enjoy your trip, Alexei!

Jun 3 2004
21:09
bobyordan
Member
Members
Forum Posts: 88
Member Since:
Feb 28 2004
Offline

[quote="AlexeiUstinov":211gmmya][[b:211gmmya]Bob[/b:211gmmya]
You are talking about very important things! But let us make our first step!
We are not Yamaha, we are not even Zero-G, to make high-tech products in a short time :-( but of course we realize the relevance of these things.
[/quote:211gmmya]

Yes I agree, let’s take the enhancements, one at a time. :) I will hold my horses a bit. :D

Some additional thoughts on the phonetic sensitivity is that an a opens the mouth more/in another way than eg an e. This results in more or less head resonance etc. I think this is probably relevant controlling the filters etc.

Happy businessing. :)

Jun 4 2004
06:09
renardfox
New Member
Members
Forum Posts: 1
Member Since:
May 11 2004
Offline

I think what Alexei meant by “Sense of Lyrics” was that human singer inevitably understands the [b:2osrtq9u]meaning of lyrics[/b:2osrtq9u], the words find emotional resonance in his/her soul and hence the nuances in singing appear, and sometimes these nuances make the final result (performed song) differ from what was originally intended by the author (take the numerous cover versions of famous songs, e.g. I Will Survive- some of them sound somber, some funny etc). So even not every human singer succeeds to transcend into the song what was meant to be there.
Vocaloid, on the other hand, is a machine and no lyrics can find any emotional reaction from him (I even should say “from it” but it’s kinda hard for me to call Leon or Lola “It”.), so you have to somehow DIY. You can make it sing any obscenities about itself and it will have no problems with that whatsoever, singing-wise. Not feeling (sensing) what you sing, not believing it makes singing fake.To avoid that is the developer’s task
That’s all I have to say about that.

Sep 13 2004
09:41
alexeiustinov
Member
Members
Forum Posts: 34
Member Since:
Apr 8 2004
Offline

Hello everyone!
Now that we have all three Vocaloid voices we have remade our original example with MIRIAM doing Christine and LEON impersonating The Phantom. This example seems to sound much better now!
Please check out the new improved mp3 – either here at the downloads section or on our web site [u:2efv7fz3]JasmineMusic.com[/u:2efv7fz3]

Forum Timezone: UTC 0

Most Users Ever Online: 108

Currently Online:
1 Guest(s)

Currently Browsing this Page:
1 Guest(s)

Top Posters:

gray: 304

robotarchie: 223

andromeda: 188

quetzalcoatl: 173

Giuseppe: 167

Luka Mitutoyo: 147

Member Stats:

Guest Posters: 1

Members: 7988

Moderators: 0

Admins: 1

Forum Stats:

Groups: 5

Forums: 27

Topics: 764

Posts: 3555

Newest Members: kprem, yang7764, lisajim, gameace, Sunnydoot, Anildoot

Administrators: administrator: 268