So you’ll know doubt recall, if you’ve been following me for any length of time.. and closely enough to notice this sorta thing or remember it.. lol.. that I’ve been kind of excited about the possibilities of vocal synthesis. Vocal synthesis is a technology via which we use special kinds of synthesizers to… well sing. This is,  I think, an interesting subject.. if only from an academic perspective.. but for me it takes on special importance because.. I can’t really sing.. and it, at least potentially, could be the best way to bring lyrics to my music… or at least one key to it.

Generally.. for music to be popular, it needs vocals..  I’m not really aiming to be a pop sound artist.. but.. as kind of avant guard and experimental as many of my musical inclinations might be.. the music I make is not like.. an unapproachable high brow thing… it is a product of a certain type of aesthetic thinking.. it is a kind of searching, a kind of adventure.. and exploration. I love the feeling like.. I’m doing something original.. developing a unique voice… I’m.. even if it’s only on a compositional kind of level.. or on a kind of.. purely aesthetic level.. talking about certain things.. its a kind of sonic philosophical discourse.. it’s often described as “conceptual music.” All of this doesn’t make it unapproachable unless the soul metric of approachability is that it sounds like everything  you’ve already heard in your life..

On the subject of lyrics 

The possibilities of lyrics is.. both exciting and  terrifying for me. Without lyrics the music seems like it’s open to infinite interpretation.. and lyrics seem to.. suggest “and this is what this piece of music is about,” the implication being that it potentially limits interpretations..  Further.. can I even be a great lyricist? I mean.. as I strike out trying to do this, will it somehow work against what is good in my work?

I have, as it turns out, been writing lyrics since I was at least in the 8th grade… They have never really been anything I’ve been overly proud of.. never sure if they suck.. but occasionally I show them to people.. and some people really like them.. so what the hell, right?

My lyrics start out in a spirit that is very much coming from the same place as my music.. a kind of anarchical relationship to structure.. which is where the searching starts.. lets throw out notions of right and wrong, and build a new world with new values.. new ideas of right and wrong.. on the aesthetic level here of course.

My lyrics tend to be about a kind of cosmic searching.. a search for enlightenment.. It’s as if the writer has a certain persona.. and that persona can be both super smart, and super monstrous.. at best there can be a kind of mind bending quality to it that challenges, at least for the moment, you’re perception of the.. well just how you think about things… and it does this in a way that feels dangerous.

sometime the next day:

I’ve been writing all this crap and not posting it, so hopefully I can just get through this one fast.. already I’m not quite sure of the flow.. but what the hell, lets finish this puppy…

Cantor first impressions

I downloaded the Cantor demo sometime ago.. and wasn’t supper happy with it.. but it is pretty close to the only game in town.. at least on the Mac.. But the demo, as it turns out, is a version behind the real version.. so it was time to check out the current version.

The first challenge is learning how to use it. VirSyn’s ideas about interface design are a little.. well both dated and idiosyncratic..  and it doesn’t help that vocal synthesis is kind of a big jump away from how traditional synthesizers work…

You have this interface, and you draw in notes.. and on the notes you tell it what word you want that note to be… from here we can adjust various parameters of the note to create a vocal performance…

The first thing you start dealing with are phenomes.. which are component sounds of our language.. you type the word, it has a big old dictionary from which it finds the word.. and the component phenomes.. and you can adjust the phenome spacing in the note. There’s more to it then this, but this is the basics of it.

The first mega huge problem you run into is the interface.. which proves to be a little bit of a pain in your butt… as far as how to get working.. and even once you have it worked out.. it’s rather clumsy.. as a for instance.. hitting command Z for undo.. very frequently doesn’t undo what you just did in Cantor, but undoes the “why yes, I’d like to add Cantor to my project in my DAW” command.. resulting in the loss of hours of work.. and then some of the features don’t even seem to quite work.. but then.. I’m just a learner, so who knows.

After some time mucking around with it I did find a workflow that made sense.. It’s one of these programs where you have to learn to work around it’s limitations.. which.. sucks.. but.. the results were exciting.

How Cantor differs from it’s competition.

On the PC side of things there’s.. I wanna say its made by Yamaha… a similar instrument.. but the difference there is that the phenomes are not synthesized, they’re sampled… The result is that you can actually get fairly realistic realizations of singers.. where as with Cantor.. there’s no hiding the fact that its synthesized.

The choice, at least if you’re on a PC, between going with a sample based or synthesis based.. singing instrument.. is probably an aesthetic / what sort of applications do you want to explore with it sorta thing: There are lots of places where robot type voices sound pretty cool, after all… which might be particularly true for someone like myself.. who does mostly electronic music anyway. 

Some other options I’ve considered:

One option is to attempt to sing… and use a program like melodyne to fix my wrong notes. Here the problems are a couple fold:

1. It “may be” that hitting the wrong notes is not all that’s wrong with my singing. There’s software out there, of a not unsubstantial cost, that one can use to fix these problems as well… but of course thats more money and more work.
2. The kind of gear needed to record really good vocals isn’t cheap, and I haven’t really invested in it.
3. The work flow would involve.. write the song, practice the song, record the song, edit the badness.. which ends up taking quite a lot of time and energy.. so that… even given the problems we find with Cantor.. it actually has a much better workflow for.. kinda how I’m doing things now.

Another option is Vocoding 

I don’t think I’ve really played with Vocoding too much.. the VirSyn effects bundle I recently invested in includes a Vocoder.. One that got a really good review in a recent issues of Sound on Sound just recently.. so I’m ready to try it out… 

Basically.. you have two signals.. signal one we look at and analyze… what is the amplitude of what frequency? You have a kind of grid of so many frequencies.. on the second signal.. we apply EQ… that imprints on the second signal.. what happened on the first.. if this frequency is this loud on the first, turn the EQ that far up on the second.. for that part in our grid. 

So this is how vocoding works… and the results tend to be.. “and now we have a funky talking robot.” There’s more applications for vocoding then this.. but this would seem to be the main one. 

latter: 

There’s probably more to say, but I must post… 

This entry was posted on Friday, January 9th, 2009 at 11:59 pm and is filed under Music Production, My Studio, musical composition, sound synthesis. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.