Reply to thread

Message: [QUOTE="Bombastus, post: 91019, member: 2178"] A multi-tap delay can produce several separate echoes from one input even when its feedback path is set to zero. That distinction matters because taps decide when the first echoes appear, while feedback decides whether delayed material is sent around the circuit again. Treating those controls as interchangeable makes complex delays harder to program. You can build a complete rhythmic phrase from taps alone, then add regeneration only where the pattern actually needs to continue. That separation is not merely theoretical. Some multi-tap processors expose exact tap counts and level decay without requiring feedback at all, while others let one selected tap become the source that feeds the whole pattern back into itself. [HEADING=2]Delay taps build the first echo pattern[/HEADING] A tap is a playback point on a delay line. Put taps at different times and one dry sound can produce a finite sequence of echoes at those positions, with each event potentially carrying its own level, pan, filtering, or other processing. The crucial word is finite. If you program four taps and leave feedback off, those four delayed events can happen once and stop, because nothing is being routed back into the delay input to create another pass. This is why [B]independently timed echo taps[/B] are useful for rhythms that must stay exact. You can place one repeat on a dotted subdivision, another slightly late, and another at a longer interval without relying on a feedback percentage to generate their timing. Some hardware and software designs make the distinction unusually obvious. A ten-tap delay can determine the exact number and spacing of echoes with tap controls and a decay shape instead of a feedback loop, proving that multiple audible repeats do not automatically mean regeneration is happening. Feedback-style delays behave differently. A single delayed event can return to the input repeatedly, so one tap may create a long train of evenly spaced echoes even though the delay line has only one main playback point. [HEADING=2]Feedback decides what returns to the input[/HEADING] Feedback takes some delayed signal and routes it back into the delay path. Each trip creates another opportunity for that material to emerge again, which is why raising feedback generally extends the audible tail instead of adding a new independently placed tap. That routing distinction explains a common mixing mistake. Lowering a tap's direct output does not necessarily remove its influence from a feedback network, because some designs separate the level you hear immediately from the amount that tap contributes to the returning signal. One documented quad-tap implementation works exactly that way. Its tap outputs can be reduced while a tap with feedback still contributes signal to the summed return path, meaning another tap can reveal that regenerated material later in the pattern. That is a useful mental model for any advanced delay. Direct tap level answers how loudly an event reaches the output now, while feedback level answers how strongly material from that point is allowed to participate in future passes. The two controls can therefore produce opposite results. A loud tap with no feedback can hit once and disappear, while a quieter tap with substantial feedback can keep generating later echoes long after its first appearance. [HEADING=2]Selective feedback reshapes the pattern over time[/HEADING] The interesting part begins when a multi-tap delay lets you choose which taps regenerate. Instead of feeding the entire audible pattern back at one shared amount, selective routing can make only certain timing positions survive into later cycles. That is the deeper idea behind [B][URL='https://goldmidi.com/community/threads/frctl-audio-delivered-dlay-1-2-0-update.76246/']DLAY's per-tap feedback lane[/URL][/B]. Its current implementation lets individual taps contribute different amounts to the echo loop, so a dense first pass does not have to repeat as the same dense block. Imagine six programmed taps forming the opening rhythm, but only the third is allowed to feed the loop strongly. You hear all six on the first pass, then later regeneration can be dominated by that third timing position rather than replaying the original six-event shape wholesale. That gives feedback a compositional role beyond making a delay longer. You can let an accented event persist, keep secondary taps finite, or make the tail simplify as it develops instead of preserving the density of the initial pattern. It also changes how you troubleshoot a delay that feels cluttered. Cutting the number of taps may damage the first rhythmic phrase, while reducing feedback on selected taps can preserve that phrase and clean up only the repetitions that follow. The same logic works in reverse when a pattern feels too static. Keep the initial taps restrained, then increase regeneration on one or two strategically placed events so the tail develops a different emphasis from the opening sequence rather than becoming a quieter copy of it. Once taps and feedback are treated as separate jobs, the controls stop fighting each other. Tap placement establishes the first-pass structure, and feedback determines which parts of that structure earn another trip through the delay path. [/QUOTE]

Name