How to use an fm synth

Fd80f81596aa1cf809ceb1c2077e190b
0
rouncer 104 Jun 28, 2009 at 22:57

fm synthesis is about this strange happening when vibrato is twisted well past normal vibrato speed, it will start to
generate aliasing tones or phasing tones im not sure, but at the beginning, if youve only got a sine wave oscillator with
a sine wave vibrato you cant do much with it yet, but you could have a listen to what it sounds like, but thats just
the beginning.

to improve your chances after just using sine waves, you can use any complex operator you wish, and you can even apply fm
to the fm and it actually works and will give you more sounds.

what ive tended to notice is you should position your fm frequency still, and you should modulate your fm amplitude, amplitude
tends to sound better when its moving, and youll get a “moving” sound when you put an envelope on your amplitude which will make
it sound *alot* better, because real sounds are always constantly changing and youll gain that characteristic.

You can actually put an envelope on the frequency also, but its not quite as thrilling as the amplitude.

an “fm” pluck sounds pretty good, and you do it with decaying the amplitude of the fm oscillator and it sounds quite convincing.

so when i jump on an fm synth, im constantly testing frequencies, enveloping the amplitude and modifying the operator waves until
i get the sound i like, which sometimes doesnt happen, im not quite good at it yet.

But its quite fun, i got more than i expected out of fm, because i didnt expect much i in fact thought it didnt work at first.

and thats about all i can tell you at this stage, but i cant wait for whats next!

Whats similar to fm is rm or am.

FM - take one sine wave and use its amplitude to modulate the frequency of another sine wave

AM - take one sine wave and use its amplitude to modulate the amplitude of another sine wave

RM - take one sine wave and add it to the second sine wave

They all sound pretty similar, but I like fm because its sound is more obvious than am and rm, but am is still
nice it tends to make cleaner sounds if thats what your after, but you use all 3 the same way.

Im a bit confused about am and rm, some people think rm is just am, and rm is no different, but I think differently.

3 Replies

Please log in or register to post a reply.

A8433b04cb41dd57113740b779f61acb
0
Reedbeta 168 Jun 28, 2009 at 23:48

I don’t think RM is adding two sine waves…I hadn’t heard the term before (it apparently stands for Ring Modulation), but according to a Google search it is multiplying two sine waves. This is just AM if you allow the envelope to be negative, though usually in AM the envelope is a much more slowly changing shape than the signal being modulated.

Fd80f81596aa1cf809ceb1c2077e190b
0
rouncer 104 Jun 29, 2009 at 00:58

Yeh, im not quite sure on that… your probably right.

6837d514b487de395be51432d9cdd078
0
TheNut 179 Jun 29, 2009 at 02:10

They are all modulations, which essentially adjust their respected parts. FM only modulates frequency, not phase or amplitude. AM only modulates amplitude, not frequency and phase, and so on. As an example, an FM applied to a constant frequency would make it sound like its pulsating (1950’s style alien saucer sound). An AM applied to a constant amplitude would make it sound like it’s fading in and out. PM is a bit of an odd one, but it generally sounds like static or like FM, but with a higher pitch / noise.

Ring Modulation is just a DSP technique that applies a carrier signal to the original source, and in this case it multiplies the two. There are many DSP solutions that do this, but the most popular one is called a Vocoder, which is an RM with an equalizer and band filter. It’s quite a common practice when people want to simulate robotic voices.

RM = input(t) * carrier(t)
where
input = input signal (range: -1.0, 1.0, such as a song or voice)
carrier = carrier signal (range: -1.0, 1.0, typically a modulated sine wave)
and t is some time value.

The real fun in DSP is when you get involved with Fourier Transforms. Converting data into frequency domain, manipulating frequencies, and then convert the signal back into time domain. It’s a little weird at first, but amazing once you see the possibilities with it.

– Edit

http://www.dspguide.com/pdfbook.htm

I forgot to mention, but check this site out. He offers a free PDF of his book and it’s amazing. The best information you’ll get anywhere, including how to use and interpret Fourier Transforms and perform DSP.