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
Phonetic input via speach.
Apr 9 2004
06:51
bobyordan
Member
Members
Forum Posts: 88
Member Since:
Feb 28 2004
Offline

Hi guys

Just woke up with a wild idea, there is probably a loth of speach recorgnition software out there (in the world). If Vocaloid allowed the user to, via a mic, input words via speach instead of we having to write them, then the workflow could be speed up and the right phonetics entered at once. Eg select a note, speak in the mic, the word (phonetic) turns up on the row connected to the note. The phonetic database is scanned by the software not on spelling but according to how the word (phonetically) is spoken. :)

Apr 9 2004
07:56
andromeda
Member
Members
Forum Posts: 188
Member Since:
Feb 27 2004
Offline

s aU n d z – l0 aI k – @ – gh U d – aI dh i: @

Cheers
Chris

Apr 11 2004
20:22
robotarchie
Member
Members
Forum Posts: 223
Member Since:
Feb 27 2004
Offline

Ak ak ak ak ak (Popeye laughing)

Apr 12 2004
01:34
quetzalcoatl
Member
Members
Forum Posts: 173
Member Since:
Feb 26 2004
Offline

I like ideas like this puddy-cat. I wonder how you would go about contacting Yamaha direct with it.

Edit…

Vocaloid is developed at Pompeu Fabra University in Spain (Barcelona) and financed by the Yamaha Corporation:

http://msl1.mit.edu/furdlog/in…..1&c=1

and .. if you have the time, please try and track down the people involved with Vocaloid at this university, here’s the web site:

http://www4.upf.es/s97is.vts

..and more contacts here…… the language is Catalan from Northern Spain. I am fluent in Spanish but Catalan is not even similar.

http://www.iua.upf.es/

….and in amongst these people must be somebody who is working on Vocaloid :D :

http://www.iua.upf.es/mtg/eng/index.html

I’ll keep updating this as I find out more.

Apr 12 2004
02:02
rickpaul
Member
Members
Forum Posts: 18
Member Since:
Apr 7 2004
Offline

I also really like this idea, especially if it could get timing information from the input voice as tweaking timing is one of the most tedious bits of just getting started out.

Of course, that may be less useful for people who aren’t also singers, but then different of us will likely want to use this stuff for different purposes. My main interest in it is being able to use voices other than my own, be it to do a female demo of a song or just to add different background vocals to complement my own. But, especially for the former use, it has to be able to sound like a really good, English speaking singer. The latter use isn’t quite as fussy in that it will be blended in less overtly, but the better it can be the less covert it has to be in the mix.

To get really out there in terms of “technical demands”, the idea for me might be if it could just record what I sing, figure out phonemes, pitch, bends, timbre changes, etc., but then just substitute someone else’s voice for mine, then let me tweak from there. Of course, in that case, I suppose one of those TC units that changes people to Cher or Barry White might be sufficient, and I do like the idea of MIDI input for those times when I don’t feel like setting up a microphone or want to play with an arrangement at the notation level to suss it out before putting vocals to it.

Rick

Apr 12 2004
02:07
quetzalcoatl
Member
Members
Forum Posts: 173
Member Since:
Feb 26 2004
Offline

Ed Stratton (Zero-G) has already started to think about real vocal input, I remember reading that he reckons it will be another 5 years before it’s possible.

Apr 12 2004
03:46
gray
Member
Members
Forum Posts: 304
Member Since:
Feb 27 2004
Offline

I would be tickled pink if they would just fix what they have now. Samples in more range I believe will fix the awful sound quality. Falsetto shouldn’t be used a full octave above where it starts, maybe more than that. And the gender change is a joke. On bass notes you must actually increase the male gender to keep from getting the chipmonk effect and vice versa on higher notes. Samples should cover 90 percent of modern songs instead of 5 percent. Come on ZeroG. Put some samples out there for us.

Apr 12 2004
07:43
andromeda
Member
Members
Forum Posts: 188
Member Since:
Feb 27 2004
Offline

To Rick
Did you now that there is a cheap and simple notation (score and pianoroll) sequencer that can take voice input and convert to midi events? It also allows you to attach audio files to the score. I’ve used it quite a bit. Though limited in some ways compared with other score writers or sequencers, it is perfect for the way I tend to work. I must point out that I don’t use the voice input. You have to sing in tune for that…..
The web site is at http://www.musicmasterworks.com They also have a section where you can post your demos (MP3 or MIDI) though it seems to have gone rather quiet recently.
Cheers
Chris

Apr 12 2004
09:17
robotarchie
Member
Members
Forum Posts: 223
Member Since:
Feb 27 2004
Offline

Lateral Solutions are a small British software house that has a nice polyphonic capable pitch-to-MIDI program called Guitar Synth. It’s a purely software solution. It’s obviously aimed at people who want to convert the pitch of their guitar strings into MIDI note streams – but a less well known fact is that it works rather well with a microphone and voice input as well. On the “OCR” thing. I don’t think that I’ve heard of a program that can convincingly convert audio into vocaloid usable phonemes as such, but I do think that perhaps a User Group posting area specifically for “discovered” pronounciations might be a good resource? Maybe we could compile a “super user dictionary” file which could shared with lots of alternative “spellings”? Just a thought…

Apr 12 2004
09:24
robotarchie
Member
Members
Forum Posts: 223
Member Since:
Feb 27 2004
Offline

[quote="andromeda":2zeg1qbq]s aU n d z – l0 aI k – @ – gh U d – aI dh i: @

Cheers
Chris[/quote:2zeg1qbq]

The quote above from Andromeda sparked that wee notion by the way!

Apr 12 2004
16:50
bobyordan
Member
Members
Forum Posts: 88
Member Since:
Feb 28 2004
Offline

Awesome, so I have actually bought a VSTi that was some Spanish students first introduction to programing? 8O Hope the person in charge has not graduated, four years ago? <img class=” />

My therapist tells me that things can only get better, from now! :?

Apr 12 2004
19:19
quetzalcoatl
Member
Members
Forum Posts: 173
Member Since:
Feb 26 2004
Offline

Yep it looks like a few people did their doctorates (PH D) partly on the reasearch for Vocaloid. I’m close to tracking down the actual developers .. emails have been sent and received and replied to, APBs have gone out to all units and we’re waiting for an eyeball. I’ll report as soon as we have bring them in for questioning, we’re running low on doughnuts and coffee .. this could be a long stake-out. Stay close.

Apr 12 2004
21:53
robotarchie
Member
Members
Forum Posts: 223
Member Since:
Feb 27 2004
Offline

Bob – did you say VSTi? You’re too soft, man. The Trades Description Act would question that. I wish it was. Cantor is a VSTi. Vocaloid is all but stand alone, ReWire not withstanding (at least till the end of the month i.e the promise of update, we’ll see).

Apr 13 2004
09:59
bobyordan
Member
Members
Forum Posts: 88
Member Since:
Feb 28 2004
Offline

Sorry Archie I really mess up my defintions there :D, I meant that the only thing that does not work so far is the VSTi part of the Vocaloid. ;)

Apr 13 2004
11:28
robotarchie
Member
Members
Forum Posts: 223
Member Since:
Feb 27 2004
Offline

Next time you’re playing a Leon track to someone and they can’t make out the words just adopt a Basil Fawlty tone and say “He’s from Barcelona…”

Forum Timezone: UTC 0

Most Users Ever Online: 108

Currently Online:
5 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