I wrote an alternative version of the built in subtractive synthesis demo that provides what I think is a better and more familiar subtractive synthesizer architecture
The subtractive synthesizer is based on the same general outline of the existing subtractive synthesizer demo, but fixes the issues outlined above. The new subtractive synthesizer has controls that are more like those of actual subtractive synthesizer instruments, with separate envelopes and resonant filters.
A general outline:
The code for the demo can be found here.
This is an expansion of pconrad’s frere jacque two voices code that allows sequencing notes and playing them back from separate instruments. One important feature that code lacks is the ability to provide different sequences of notes for each instrument. This is important because in real pieces, it is rare that all the instruments are always playing the exact same notes.
my code is an expansion on that demo, providing this feature, and demonstrating it with two separate instruments.
I made a basic implementation of Karplus-Strong synthesis to create a string-like sound using a delay. this demonstration is implemented simply as an extension of my MiniSubWaves code, creating a variant that has a comb filter applied after the main resonant LP filter. the result is some very rough sounds that are nevertheless quite reminiscent of physical string instruments.
karplus strong synthesis works by utilizing a noise burst that can be modulated by an envelope, that is then passed through a low pass filter. finally, the signal is passed through a very short delay. Very short delays can be represented/simulated by comb filters. comb filters create a pattern that strongly resembles the interference pattern caused by playing a signal along with itswlf a very short delay after. The delay time controls the pitch. Longer delays corresponds to lower pitches, while shorter delays correspond to higher pitches. The equation for the fundamental frequency of the string sound is given by F0=Fs/Fs, where Fs is the sampling frequency, Fn is the note frequency, and F0 is the fundamental frequency. Since allolib currently uses a sampling frequency of 44100 Hz, we can use this equation to calculate the delay required to create a note with the desired fundamental frequency. The expression for the delay (in seconds) is given by D = (44100/Fn)/44100. The demo uses this formula to automatically calculate the delay so that notes can be produced at the proper frequency by dynamically adjusting the delay time.
Finally, we can put all these components together into a simple composition that utilize each of these elements. it will include our improved subtractive synthesis, our multi instrument sequencing code, and finally, our code to synthesize string like sounds using the Karplus Strong algorithm.