I did some analysis on Sylenth's detune a while ago because I wanted to see if I can implement it in my own little synth. Here are my results for anyone interested in the technicalities. All others: better skip over this post and play with your synths in the meantime.
Sylenth's detune is linear with regards to half tones. If you for example set it to five voices and you detune the "outermost" voice by 2 half tones then the five voices will be detuned as follows (in half tones):
-2 | -1 | 0 | 1 | 2
This is just an example because such a detune will obviously sound like crap.
Please note that this is not a linear detune with regards to the frequencies! If the center voice for example plays at 440 Hz then the detune of the voices in the example above is the following in frequencies:
349.3 | 415.4 | 440 | 466 | 493.7
As you can see neighboring frequencies do not have the same constant distance between each other as was the case when we expressed the detune in half tones.
The detune only looks linear in a spectrum analyzer because most analyzer use a logarithmic scale.
fluffy_little_something wrote: ↑
Thu Jan 10, 2019 9:29 am
Whatever the spread mode is, beyond 12 o'clock max it sounds unpleasant in my view.
That's because the maximum detune for Sylenth is around 3.3219 half tones (for some people this number might immediately ring a bell as it is log_2(10)). This is in each direction from the center! So it is possible to detune the oscillator by a tritone
from the lowest to the highest voice. That's why Sylenth should be your go-to synth if you want to make the devil's music.
Joking aside, I wonder why they have chosen 3.321 half tones as the maximum. Are there any useful patches that use such a heavy detune? Perhaps some modulations with a quick ADSR?
There's a reason why Sylenth is "a couple of hours" better than other synths. Its knob does not go from no detune to 3.321 half tones linearly but instead it seems to map by f(x) = x^2. So if the knob at min is represented by x=0 and at max by x=1 then the detune in half steps is calculated as f(x) = 3.3219 * x^2 (see the red graph in the attached image).
I think it can best be explained with the attached image. The x axis shows how far the detune knob is turned to the right. If you multiply this number by 100 you get how much the knob is turned in percent. The y axis shows the resulting detune in half steps. The red line shows Sylenth's detune mapping and the black line is a simple linear mapping which might be used by synths with "less hours" in them.
With a linear mapping you already hit 0.5 half tones when the knob is turned 15% of the whole range. With Sylenth's mapping you hit it at around 40%. This means that the knob gives you much more room for the interesting detune range where the detune is not yet too heavy. Or put differently: the detune starts rather slow and ends rather quick.
You can also see from the graph that at twelve o'clock (0.5) it is already detuned by 0.8 half steps which is the reason why it starts to become rather unusable from that point.
I hope this answers dav0001's question.
You do not have the required permissions to view the files attached to this post.