Original Werner Kaegi’s paper describes sound synthesis technique which is called VOSIM (for VOice SIMulation). It’s an algorithm which is used in speech synthesis to generate vowels (but it could be used for the fricative consonants or whole syllables as stated in paper). This method is fairly simple and well fitted to be implemented as ‘Logue oscillator. Other methods like e.g. FOF or formant synthesis are more complex in implementation or harder computationally.
Only NTS-1 unit was tested on real hardware by me.
If you use this oscillator and you like it — please donate: PayPal
Follow the instructions in README.md to compile oscillators if you want to do it manually. Feel free to modify the code, it’s easy and fun.
|param||logue param||closest term in Kaegi paper|
||param1||reciprocal of pulse
||param2||number of pulses per period|
||param3||LFO Target: 1 means delay, 2 means attenuation|
To understand these parameters better read Kaegi’s paper and consider the graphs below: one oscillator cycle is shown on the graphs.
M = 0.3
b = 0.9
T = 0.1
N = 4
Since clever trick used in ‘Logue SDK always maps phase to
[0.0, 1.0) interval we can calculate the
Freq param with the formula
Freq = (1.0 - M) ÷ T
Freq = (1.0 - 0.3) ÷ 0.1 = 7
Current implementation allows negative values of attenuation
b so waveform will look like on the figure below