Does anyone know of any attempts to do harmonic analysis on growls, screeches, screams, or grunts? By "harmonic analysis" I mean actual decomposition and analysis with oscilloscopes, sound software, etc. - stripping out all the noise (I love growling - do it myself - but in the strictest acoustic sense there's a lot of noise in growled vocals) and analyzing the main tone and its harmonics (if there are any). This would/will tell me what note the vocals would be if they weren't surrounded by most-excellent death metal charisma.

That's the project that I'm starting for my Physics of Music course, and I would like to say with some authority if anyone's seriously analyzed any sort of death metal vox before. Any leads?

Along with that, I'd love to get samples from anyone willing to send me a file or two. I'll be emailing a lot of the pros to see who'll contribute, but there's no reason I can't get some unsigned voices as well. I welcome all input!

If you do want to send something (gugglezatch@gmail.com), 3 things:
1. Some sound formats suck. If it's something crazy that only plays on an Apple IIE, I won't be able to analyze it.
2. Pure vocal samples are much better than clips with other instruments included - it's well beyond me to strip out guitar tones or - even worse - drum blasts.
3. Direct, sustained vox are best, with no words. Simply a screech, scream, grunt, or growl, at any level, will be great. Constant tone helps too - just empty your lungs as consistantly as you can.

Yeah, and if you contribute, I'll share the results with you when it's all done (at least your own results, and those of anyone else who is cool with me sharing graphs and/or stripped sound files of their vox, that is)

Is there anywhere else (other forums) anyone can think of for me to post this - maybe get some more leads/takers? Sorry if this is a bit off-topic, it's the best place I could think of.

-Chuck-

i dont know what exactly the point of doing this is

i can do a midtone growl, then make the same exact motion only singing and make the same town. just rid of the gravel sound surrounding the tone

__________________

If you've got the transition down... good for you, dude. Most folks don't, really. And there's bound to be a difference in the range of harmonics produced from a clean note and a growl at the same frequency.

By that I mean that higher octaves - many frequencies - are produced when one tries to just hit, for example, middle C (440Hz). With anything but a signal generator that produces a sine wave at 440Hz - any other instrument - higher octaves are emitted along with the primary note (in this case, a boatload of higher C's). So, I'm interested to see how screaming and growling stifle those higher harmonics (or produce more?).

In the process I might find out if there are any characteristics common to death metal vocals. It would be interesting to know if growling is just clean singing padded with random white noise or if growling could possibly be manufactured from clean singing - put through some sort of reverse filter in CoolEdit, for example.

And anyway, why not? Science yields a lot of unexpected results. Like Teflon.

-Chuck-

ill play around with it tonight, nothing better to do. i have cool edit. what do you recommend? a scientific filter?

__________________

I have CoolEdit 2000, but I only got the first growling sample today. Haven't decided what specific filters I'll try for. I'd suggest somewhere around a 60% noise reduction for starters. The filters on your version are probably different from mine - the only named ones I saw were facile - like "get off the phone!".

-Chuck-

yeah

theres tons of filtering/limiting/compression stuff on that program. hard to use though, they have very simliar tools filed under very different areas

__________________

440Hz would normally be considered an A.

__________________

Too true. I don't quite claim to have mastered any of the frequency scales yet... nor do I see it happening before the end of the semester. I knew there was a good chance of innacuracy, seeing as how I only play music thus far, and haven't yet analyzed it. Yet.

Thanks for the correction.

-Chuck-

heh, too bad you won't get very far. To say that a growl or scream is a note with noise is very inconclusive. Your speaking voice is at a certain frequency (or note), and when a person sings [correctly], their voice's pitch is higher. A growl is lower than that, and a scream is just a growl with your mouth open. The only person I know of that does anything like what you say would be Chuck from Death when he covered Painkiller. To say the least, he is singing it, just with a lot more crunch and air behind it. Hence, clipping.

__________________

You're working off assumptions and so am I. We'll see how the project goes. And it doesn't quite follow that, simply because growling is lower than normal voice, it would be harmonically unanalyzable.

My hope is that a simple principle will apply - that anything produced in a consistant manner by the vocal chords has a base frequency. It might very well be that growls are random vibrations of the vocal chords, but we shall see. Hopefully there is a base frequency, even if it doesn't sound like anything remotely tonal. To wit, it very well could be that screams and growls are simply [insert euphemism for 'crap' here] emitted at a constant frequency of 380Hz, or 397, or whichever.

If I knew already I wouldn't be doing it in a class. Well, that's wrong - I'd still do it, but I'd already have something to back up my hypotheses. I'm not above doing a known experiment for a nice grade in a 4 credit-hour course, but that's not the case here.

I'm really curious to know if you've got something other than an intuitive opinion to back up what you've said. That's all I have at this point, early in the game, and opinions hardly determine a victor in modern scientific matters. If, however, you know of some analysis or study that has been done, kindly post a link and I'll check it out.

-Chuck-

ADDITION
I sing both 'correctly' and 'incorrectly' - the incorrect range being below my normal range, utilizing a higher larynx to produce quieter - albeit clean - tones. These have discernable frequencies, as matched to a piano.

, if you turn on a tuner and set it next to your mouth, you will get a certain pitch. That's not an opition, it's common sense.
Aside form that, when you are growling, you might be doing what Tom from Slayer is doing. He does not sing (well, on Hell awaits and reign in blood he doesn't). He might be just putting a crunch behind his normal speaking voice. He very well could talk like that normally. And the same goes with Matt from Iced Earth. On the album Alive in Athens, when he talked, he almost sounded as if he was singing-that is, on the same range.

__________________

also

some other people sing and scream at the same time


alexi laiho on hatebreeder, something wild and tokyo warhearts

deron miller on foreign objects- univeral culture shock

__________________

"heh, too bad you won't get very far. To say that a growl or scream is a note with noise is very inconclusive."

Saying this, and then going on to say that growling is lower than normal speech (agree), gives the impression that you do not think that growling is analyzable by virtue of the frequency range it inhabits. From your last post, I get the impression that that's not what you meant, however that principle is what I was refuting.

Apparently we both feel that speech, growling, and clean singing can all be determined to have a base frequency... therefore I'm not sure what you meant with "heh, too bad you won't get very far". If it can be analyzed, that's all I'm looking for. How would that be not getting "very far"?

The fact that it should be easily guess (and be right) that it has base frequency, and to do scientific research of it is insignificant. But if you feel that there's more to it, or that you get to so something cool and know the end result before you start I'm definatly not stopping you.

__________________