>>11170566That's what I've been doing recently. I posted my thoughts in the thread this anon mentioned:
>>11168598Basically you need to decide first what kind of synth you want to design: additive, subtractive, FM, etc. I've been messing with the additive approach lately. The idea is very simple: given a frequency, you generate a bunch of harmonics with different amplitudes and add them together. There is a lot more to that but this is a good start. The first thing to do is to create some kind of a prototype on PC. Once you get that working you can move on to microcontrollers.
The simplest first step is to generate a single frequency sine signal and generate a wav file and play it on PC. So you have three ways to debug: by just listening to it, then checking the spectrum in Audacity, and also generating points to a file and plotting them on desmos to see if the wave form look right.
Initially you can take the brute force approach and just generate y=sin(x) in a loop. The next step is to optimize it and create a look up table: you only need to generate 2*PI points with a certain step so a typical table size would be something like 1024 or 2048 etc. Then you need to learn about sampling rates, and how to interpolate missing points in your descrete look-up table by doing, for example, linear approximation. After you have a single freq wave working, you can start mixing them by simply adding them together. Very simple. Then at some point you can start learning about microcontrollers. Your micro has to be reasonably fast, at least 100Mhz and must have a FPU. And obviously it should have a DAC so you can output your analog signal and listen to it, or check the waveform on a scope.
Do you program? I can give you some specific points and basic examples. You can use any language on PC (I use C for quick prototypes and C# so I can also plot waveforms). Micros are programmed in C/C++ or sometimes assembly (if you have to).
