Clipping, Distortion and Aliasing
Analog & digital clipping
Clipping in analog hardware modulars happens because no voltage can go above the PSU’s voltage. Clipping in software modulars happens because the developer has to put a limit to the virtual voltage, otherwise it can go to very high values until they overflow* or get clipped by the hardware audio interface and in the way cause very high levels at your speakers/monitors.
Clipping in analog hardware causes distortion. Clipping in digital also causes distortion but in worst cases can also cause aliasing. Off-course both distortion and aliasing can be good or bad depending on what you are doing.
Because of this, It is essential to allow the musician to have some leeway before clipping starts to happen. For example in Eurorack, PSUs produce a maximum range of -12V to +12V. Yet you would see VCOs and audio signals having voltage between -5V to 5V, which is less than half the range of what the PSU can offer. This is good to give the musician some room for loudness fluctuation (i.e dynamic range) without hitting the clipping level (i.e 12V). In digital modulars, it’s the same story.
Clipping In SoloRack
In SoloRack, VCOs and audio sources give -1V to +1V. Which is considered FULL SCALE in most DAWs (i.e 0db FS). Internally, SoloRack modules by default clip at -4V to +4V (i.e +12db FS). Which is a considerably high level. This is intended to give more of a dynamic range. However, sometimes it’s useful to set a lower clipping level per module. You can do this by right clicking on a module and selecting a level from the Clip Level menu.
Where high clipping doesn’t work well
In amplifying feedback loops. Clipping is eminent. This means that no matter how high you set the clipping level at, it WILL clip. Preventing distortion and/or aliasing here is impossible. In this situation, setting the clip level to a high value will only amplify the signal path. So it is best here to actually do the opposite, set a low clip level like +3db FS or lower and work with that. And consider the clipping as an effect instead of an artifact. Which can be quite interesting in many ways.
Notice that some modules that have a feedback loop exposed through a send/receive, like the S620 and the S224 are by design, highly vulnerable to feedback amplification. Therefore the default clipping level for such modules is +6db FS (instead of +12db FS) just as a precaution. But you may want to lower that level manually even further if you like to experiment with feedback loops.
It is also recommended that you use the new SA13 Clipped 2 DAW as an output module if you want to experiment with feedback loops to prevent hurting your speakers and ears in case things go out of hand quickly. Or at least place an external limiter in your DAW after SoloRack
What oversampling does
Aliasing happens whenever any harmonics are supposed to have frequencies that are above half of the sampling rate (i.e. the nyquist limit). Those harmonics eventually get mirrored back below nyquist which is what gives the aliasing artifacts**. Oversampling increases the internal sampling rate at which SoloRack works. This essentially increases the bandwidth (i.e frequency range) and hence, makes the nyquist limit higher. Thus it lessens the probability that the harmonics will go above nyquist. Every time the oversampling rate is doubled. Audible aliasing (if it exists) is more or less halfed.
The reason that clipping may cause aliasing is because clipping causes distortion. And distortion causes higher harmonics to occur, which increases the probability that those harmonics will go above nyquist.
Obviously the higher the pitch of the sound the more probable aliasing becomes. Though if the pitch is too high, killing aliasing would usually make no musical sense because what would be left is only a few highly pitched sin waves. Even if done in pure analog. At which point the sound of aliasing it self would probably be more interesting.
* overflows can cause very high CPU usage.
**This is a very layman’s way of saying it.