Soo, on another Web-Forum the topic of Drum-Machine Groove and Jitter came up.
An there is this thread here:
How solid is the Digitakt clock for audio out? Plus a swing question which mentions the jitter tests at Litmus — Innerclock Systems as well, but does not go much further.
I find the description of the innerclocksystems measurements quite fuzzy, and was extremely surprised that a machine that feels so tight should have any jitter at all for internally sequenced events.
I mean, this is all block-based processing and jitter comes from only reading events at the beginning of a block, without any further timing information (i.e. play immediately).
That rubbed me wrong, so I put in a bar of 16th into the internal sequencer, pulled down swing to the left (i.e. 50%) and recorded a while at 120bpm (44.1kHz audio). Once with internal clock, and once as slave to the E-RM Multiclock. In both cases I found strictly alternating distances between recorded events of 5497 samples and 5526 samples.
Now I am not sure how that comes about. It is not jitter, as it is not nearly random, but instead this is a minimum swing. Either I have missed some override for swing, or this is on purpose, to make it feel more groovy, or this is a bug in the swing calculation or in the data-knob-reading routine. As this is not a potentiometer but an encoder and I rotated further than left, it is very improbable that is hardware variance. The manual states that swing can be adjusted from 51% to 80% with the default being 50% but I think that is just awkardly phrased and this is less than 51%.
So, does anyone know where this comes from? Is it a bug? Is there a quick way to reset parameters to default? To minimum / maximum?