Want to design a new type of synthesizer
I was studying a spectrogram of a violin with meldaproduction's MAnalyzer, when I noticed that the overtones weren't sine waves. I also noticed that these overtones repeated themselves at octave intervals starting at semitones of 0, 19, and 28. What I was thinking was that I could design my own overtones using a graph, like you'd see in algebra, and use them as filters that repeated every octave on white noise to design new timbres.
How do I go about this? How do I use Max/MSP to create my own overtone shape on a graph that can be used to filter white noise? I was hoping to use Max/MSP.
Also, I don't want to hear any bullshit that I need to do things the classical way and use waveforms and bandpass filters. I'm not asking for alternatives; I need to know the answer to my question.
1. Great attitude.
2. What's wrong with using the thread you already started on this?
3. I'm not sure you even know enough to know whether someone has answered your question. I say this because you're rejecting the use of band pass filters when it is precisely this type of filter you're going to need to use in order to isolate particular frequencies from a bit of white noise, which makes me think you don't even know what a band pass filter is.
I think you had some pretty good answers to get you started in the right direction in your original thread. Did you try any of it?
Band pass filters can only ever make a sine wave overtone, which is exactly what I don't want. I want to be able to create a different type of overtones to make more realistic sounding timbres instead of the buzzy or dull kind of sounds you get from using standard waveforms. No one has been able to show how that's possible.
And I'm sorry about the attitude, but I've been trying to make sounds I like for six years and synthesizers just aren't working for me. I'm just starting to lose hope that I'll ever be able to make the music I like, and I'm very frustrated by that.
getting angry with us sure isn't going to solve it.
Have you looked into additive synthesis? The minute you say "overtone intervals", all us geeks are raising our hands saying "additive", you see. In fact, you haven't said anything to make me think additive wouldn't work for you.
... it wouldn't be "unique" or "new", since looking at spectroscopes and breaking down violin sounds has been done by sound designers for decades, though....
Because you're still just using sine wave overtones. I need to design my own overtones. No sine waves.
Ok, so I was right with number 3. You don't understand the properties of a band pass filter else you wouldn't say that they can "only make sine tones." It's not true.
Second I think you need to investigate what overtones actually are. You will find that overtones by their nature approximate sine tones. It's the entire basis of additive synthesis and Fourier's discovery that all complex sounds can be reduced to a number of component sine waves.
You might reinvestigate that violin sound you spoke of, because to say the overtones of the sound weren't sine waves (as in they were saw or square waves) doesn't make sense from a scientific standpoint.
We're you looking at a sonogram analysis of this sound?
I was using MAnalyzer from Melda productions. The left side of the fundamental frequency's overtone was not a standard sinewave overtone; the left side of it had a dip in it that appeared on the left side as it reached towards its peak. I'm trying to get a picture of it now, but I can't seem to find MAnalyzer in my VST effects directory in my DAW so I'm having to reinstall it.
The overtone I saw could not be made with a bandpass filter. Sine waves are evenly distributed on both left and right sides, but naturally occurring overtones don't. I'm not sure who Fourier was or what he discovered, but the theories branching off from it seem too limited to explain acoustic instrument's timbre. If I'm right, we could program a new synth that could make these overtones and create timbres that all previous synths could not since they could only ever use sinewave overtones.
MAnalyzer uses standard FFT analysis to produce its spectrographs and sonograms. The second 'F' in FFT stands for Fourier. Just FYI.
It sounds like you're using the spectrograph view in this analyzer (amplitude + frequency). For a more detailed analysis you're going to need to use the sonogram output.
When you say one of the overtones has a "dip on the side" this makes me think that what you are actually seeing are two overtones in close proximity in this sound you're analyzing.
Not when I look at it in a different analyser. If two overtone were that close, it would sound dissonant. Regardless of whether it's true or not, is there any way to design filters on a graph in Max/MSP?
It should look something like this: http://www.coolmath.com/algebra/11-graphing-quadratics-parabolas/images/04-graphing-parabolas-02.gif
Although it wouldn't be a parabola as that would still just create a sinewave overtone.
No, close overtones don't necessarily sound "dissonant", mostly because the ear doesn't always hear overtones discreetly. The ear hears the combination of overtones as a single sound with a particular timbre. Again, this is additive synthesis 101.
As to your graph question, look into zplane~.
Will do. Thanks for your help, and sorry about the aggression. College has me stressed out and I'm panicking at the idea that I might have to give up six years of work by being burnt out.
Maybe you want to look at the example called "forbidden-planet"?
an overTONE is a cosine wave by definition i think, i would probably just call that additive synthesis what you are after. :)
-110
some aggro fourier rap (white trash m&m style) fo' yo
fourier was a dude
who sat down an rapped down an' proved
dat if you found a sound
dat go 'round
back and thru infinty
like divinity
inside da sound
you have a sine-a
all da time-a
no noise is any finer
than a a bunch a sine-ahs
wid crazy phases, magnitudes ablaze-a
yeh listen to the dude