Two delay lines, not one delay with a spread control

Open almost any stereo delay and you will find one delay time and a control that pushes the two sides apart from it — spread, offset, stereo, whatever it is called. The two channels are locked together and the stereo image comes from the difference you dial in around a single value.

That is a good design for what it does. It widens. It cannot produce a rhythm, because both sides are still counting the same subdivision.

Tempo Delay decouples them. The left and the right each have their own note division — sixteen per channel, straight, dotted and triplet — or their own free time anywhere from 1 to 5000 milliseconds, and their own feedback amount.

A quarter on one side against a dotted eighth on the other is a three-against-four figure that drifts out of phase and back into alignment on a cycle you can hear. That is not a wider version of a delay. It is a different thing arriving on the same track.

You could put a mono delay on each side of a split and get two independent times. What you would not get is the two of them talking to each other, or a way to treat the result as one image.

The ping-pong toggle swaps the feedback matrix from parallel repeats to cross-channel ones, so each line feeds the other and a single tap walks across the field. And a mid/side width matrix from 0 to 200 per cent sits after the whole thing, so the result can be collapsed to mono or opened out as one decision rather than two.

Two uncorrelated repeat trains at high feedback stop being a rhythm quite quickly. Each side is producing its own series, the two series do not line up, and past a certain density the ear gives up on counting and hears a wash instead.

The settings that keep the effect legible usually have one side carrying more feedback than the other, so there is a dominant pulse and a second one commenting on it. That is a musical judgement rather than a rule, but it is the one worth starting from.