KVR Audio

jancivil · Post by **jancivil** » Sun Mar 02, 2014 9:08 am

woggle wrote: the idea that say a Warlpiri speaker will use minor thirds as some sort of signifier is really silly - or at least requires some very solid evidence

I don't know, it could be happening. I would want to see it repeatedly. But in terms of most society in the world, certainly where I live, I doubt you could say very much that is meaningful.

You might find things in eg., southern culture, a kind of twang, and 'accent' and say something about it but quantifying it is going to be more complex than musical convention typically uses.

jancivil · Post by **jancivil** » Sun Mar 02, 2014 9:19 am

Musicologo wrote:Cantometrics

well, there is psychology mixed with sociology and not my area...

to the original point, 'harmonic' function of speech seems a stretch and I think I'm prejudiced against the terms. Maybe it does smack of what Lomax was after, but 'harmonic' has a certain connotation

Musicologo · Post by **Musicologo** » Sun Mar 02, 2014 6:18 pm

Well, going further now: you claim it is not very useful to define a portamento and couldn't see the advantage. Well, it's exactly one kind of thing I'm looking for right now. If you think of samples and VST's it is very useful to teach a computer to DO IT, and a computer only understands numbers!...

In case of human voice and instrumental lines that mimic human voice that becomes complex but I still believe doable.

We're just lacking an efficient way to describe it properly. But I'm assuming there must be a way to convey a convincing portamento, a slide, an inflexion, a weep, a wail, etc... without having to actually PLAY it on a midi keyboard. If we can DO it, we should also be able to express it. I mean, when I play it on a keybard I get numbers. I look at a midi and those are numbers. They are even discrete, they are not a continuous!... They have a resolution. They are just a mix of some controllers involving something more than pitch and duration, but I'm certain they can be mapped, described and notated somehow!

The clarinet in rapsody in blue can be transmitted to a computer probably by an exponential/logarithmic line(?)... a portamento/glissando that is both increasing in velocity and tempo* as it approaches the top note.

*(or keeping the tempo, but with progressively short note durations in each pitch.)

woggle · Post by **woggle** » Sun Mar 02, 2014 8:12 pm

there are lots of academic studies on this topic - here is one that looks at Chinese http://www.icphs2011.hk/resources/Onlin ... 0Aijun.pdf

Looking at this study and a few others it seems there might be universals, but these are not of the form of rising thirds etc but rather of the form of certain pitch contours. That pitch space gets partitioned differently in different languages - and not according to some sort of 12Tet notion. And there is lots of variation even within a universal

eg "The speakers may use different strategies to
express the same emotion. For the male speaker, the
„Sad‟ has a kind of level offset tone, whereas it has a
slightly rising tone for female speaker. "

And there is a real paucity of information about most languages, so any conclusions have to be very preliminary

Nystul · Post by **Nystul** » Sun Mar 02, 2014 9:07 pm

This is in Japanese and I don't have a clue what is said, but staring at the piano roll is somewhat educational:

Those who study linguistics aren't typically going to write about it from the point of view of a musician. But with the synthesis tools we have available these days you could experiment and start to get a handle on how things work. I do think even if you have some success in synthesizing realistic speech patterns that your conclusions may not be so cut and dry. Most people can't tell when I'm joking and I'm an actual person.

Need to start carrying some emoticon signs around with me.

woggle · Post by **woggle** » Sun Mar 02, 2014 9:12 pm

slowed down speech, sung, then the singing sped up to the speech rate. Rob has done a lot of work with speech as music. This is a pretty standard tool http://www.fon.hum.uva.nl/praat/

jancivil · Post by **jancivil** » Sun Mar 02, 2014 10:11 pm

Musicologo wrote:Well, going further now: you claim it is not very useful to define a portamento and couldn't see the advantage. Well, it's exactly one kind of thing I'm looking for right now. If you think of samples and VST's it is very useful to teach a computer to DO IT, and a computer only understands numbers!...

Yeah, no. By portamento I mean 'is continuous over a range', there are not points on a graph with it.

What is it about ones and zeroes that makes a continuous change of pitch over a range a mystery to describe so that we're programming a lot of single points? How are you going to describe this in terms of points? It's continuous; how many do you imagine there are? I think the concept you're missing is 'infinitudes'. Think about it, do some investigation into this.

'Because samples': well, no what I do is choose a portamento sample. The resolution of pitch bend is boilable down to the discrete, 14 bits, versus a bend of a string on a guitar, or a 'bottleneck' slide: What do you think happens in these cases? I believe it is not a description of all points in between, as I believe that is not possible because of the nature of it. How is portamento built into a voltage-control synth? i'm trying to find out.

Now, we deal with how long it will take, and we may stretch the sample accordingly, or something is scripted. I think that isn't a description of points on a graph either, the computer knows what 110% means somehow. I think the technical solutions lie elsewhere.

jancivil · Post by **jancivil** » Sun Mar 02, 2014 10:30 pm

Musicologo wrote:The clarinet in rapsody in blue can be transmitted to a computer probably by an exponential/logarithmic line(?)... a portamento/glissando that is both increasing in velocity and tempo* as it approaches the top note.

This particular phenomenon as an exercise (in which technically you'd want to model the clarinet mathematically) is problematic as the thing differs in fact from say a string bend the way it does out of limitations of the clarinet; it's a trick. That portamento isn't a given, which is the one reason I brought it in. The goal however is to "do" a totally smooth portamento. The clarinetist has to do some work, as opposed to slide trombone. Some people totally pull the illusion off as a continuous slide. Examination of what has to happen reveals why this is not always true.

I didn't say 'it's not useful to define', it's been defined, I said that notating it by points is doing too much for something that isn't necessary, essentially.

Musicologo · Post by **Musicologo** » Tue Mar 04, 2014 7:34 am

What's the difference between frank sinatra singing new york, new york, and liza minelli singing new york, new york, and a clarinet playing the melody of new york, new york?...

If you exclude differences in rhythm, pitch, dynamics, timbre what's left?
"gestures"? "articulations"? a "combination of several of those things?".

If you tell me, try to mimic frank sinatra voice i would DO something different than "try to mimic liza voice". and I would know instinctively what that would be, even if I don't know exactly how to verbalize it.

That's where we are. What is the difference between you saying "these pretzels make me thirsty" angry or ironically or sad. and how to translate that into numbers and teach a computer.

"emotional prosody" was an interesting key term. I've found a myriad of articles and some interesting articles regarding studies in my mother tongue.

Apparently, regarding speech, the variables "mean of the fundamental frequency"; "variation of the fundamental frequency" and "duration of the phrase" are good predictors and parameters to mess around. This is already something...

Musicologo · Post by **Musicologo** » Mon Mar 17, 2014 1:21 am

Some sounds behave so strangely, so strangely, so strangely... wait...

http://aeon.co/magazine/altered-states/ ... -in-music/

pensaku · Post by **pensaku** » Mon Mar 17, 2014 5:32 pm

Adding some thoughts after Musiclogo's nailing it...

We repeat some sound a lot so it's so familiar it sounds like it means something but not always.

If you just shift little pitch of it it sounds like something else like in chinese or voice actor's "Per Se" talk.

I mean... it's educational thing, it's not set thing.

nasenmann · Post by **nasenmann** » Mon Mar 17, 2014 7:11 pm

Musicologo wrote:Some sounds behave so strangely, so strangely, so strangely... wait...

http://aeon.co/magazine/altered-states/ ... -in-music/

That sound bite is really illustrative and inspiring, thanks for the link!

memyselfandus · Post by **memyselfandus** » Sun Mar 06, 2016 3:12 pm

Very cool

jancivil · Post by **jancivil** » Mon Mar 07, 2016 12:59 am

phuckit

zethus909 · Post by **zethus909** » Wed Mar 09, 2016 2:53 pm

everything around me is an instrument..

Harmonic function of speech