Well… how do I put this down…
What am I doing “technically” (to the sound) when tweaking the “width” parameter in the default track-DSP? 
I did a quick test using a mono rimshot sound, and at maximum (FF) width, Renoise seems to have simply inverted the left channel. So from 00 to FF, it’s probably morphing/fading the left channel somehow between the normal and inverted state?
Weird. I always just assumed it added a slight ms delay to one of the channels, hehe.
Edit: I quickly made an animation to show the effect. This is width of the rimshot sound changing from 00, 20, 40, 60, 80, A0, C0, E0, FF.
(Excuse my cursor, hehe)
Just tried it in Soundforge (Process > Pan/Expand), setting width to full, and it appears to do the same thing. The only difference was that it inverted the right channel instead of the left.
You mean the left channel? (left is top half of the image, right is bottom).
I’m not quite sure myself why it amplified like that. I forgot to ask if anyone knew why. It almost gets 2x as loud, which could definitely cause some problems if a person wasn’t careful, heh.
I tried again with Soundforge, doing some partial (25%, 50%, 75%, etc) widening, and it doesn’t make the altered channel louder. In fact it makes it quieter. 
Ah well!
Maybe Renoise tries to compensate the volume which would get lost until you multiply the values with -1?
Say the left channel has the values 2, 4, 6, 8.
Inverting it 90 degrees would mean multiplying every value with -0.5 .
the result would be -1, -2, -3, -4 -> sound is quieter.
Only if you invert the sound 180 degrees, you get the same volume
(because multiplying with -1 would result in -2, -4, -6, -8) .
So my guess is that Renoise amplifies the volume until you have inverted the sound 180 degrees. But I’m just guessing here… Don’t trust me too much 