Feedback in Uhbik-D

[Genetic_Junk]

I'm trying to understand how the feedback works with the five taps. Do all taps send their output to all taps that have feedback enabled? So if feedback is only enabled on tap 1, then the output of taps 1 through 5 will only feedback within tap 1?

[xh3rv]

If we ignore feedback, the signal chain looks like this:
( input ) -> tap -> ( wet ) ...

In fact, though, all taps do this:
( input ) + ( feedback ) -> tap -> ...

Enabling feedback changes what the tap writes from:
... -> tap -> ( wet ) ...

to:
... -> tap -> ( wet ) + ( feedback ) ...

Worth noting that taps apply their gain/pan settings for ( wet ), but not for ( feedback ).

This is a pretty good look at the type of tape echo hardware the software metaphor follows, maybe helpful as a visual thing (the last part of the video looks inside the box)


(Seriously bemused that the counter-clockwise rotation of the 'tape'/disc is counter-intuitive for when the output of a tap plays but clear for how feedback gets written onto the tape )

[aaron aardvark]

xh3rv,
Since my Line 6 PodXT has an option where it models the Binson EchoRec, I found that rather interesting. Does the Uhbik-D specifically model the EchoRec? Or perhaps the Uhbik-D doesn't model anything in particular? That EchoRec gives the distinctive Pink Floyd "One of These days" echo.

[Genetic_Junk]

If we ignore feedback, the signal chain looks like this:
( input ) -> tap -> ( wet ) ...

In fact, though, all taps do this:
( input ) + ( feedback ) -> tap -> ...

Enabling feedback changes what the tap writes from:
... -> tap -> ( wet ) ...

to:
... -> tap -> ( wet ) + ( feedback ) ...

Worth noting that taps apply their gain/pan settings for ( wet ), but not for ( feedback ).

This is a pretty good look at the type of tape echo hardware the software metaphor follows, maybe helpful as a visual thing (the last part of the video looks inside the box)

Thanks very much for the detailed explanation. Is feedback a 16/16 buffer shared by all taps? And if a tap has feedback enabled it will receive a portion of that buffer equal to its length?

[aaron aardvark]

and here is a bizarre video to go with the Pink floyd song:

[xh3rv]

xh3rv,
Since my Line 6 PodXT has an option where it models the Binson EchoRec, I found that rather interesting. Does the Uhbik-D specifically model the EchoRec? Or perhaps the Uhbik-D doesn't model anything in particular?

I don't think it models anything specifically, the same concept of permutations of reading/writing per head is in a lot of machines - Roland Space Echo and Korg Stage Echo, for example. Dunno what Urs was looking at, but I'm sure he's familiar with all this.

Thanks very much for the detailed explanation. Is feedback a 16/16 buffer shared by all taps? And if a tap has feedback enabled it will receive a portion of that buffer equal to its length?

In a sense, yes - it's actually the same buffer as input. Maybe it's clearer to say 'input' and 'feedback' are both 'streams' rather than audio committed to tape or memory. In hardware, one tape is enough, and in software just one chunk of memory containing a sequence of samples (like an audio file) should be enough. In hardware the tape moves past different heads/taps at a steady rate; all writing is done at only one place, reading is done at multiple places. In software I'd assume it wouldn't be the 'tape' / memory but rather the location of the 'heads' - the counter/index to look up where to write or read form in the chunk of memory - that increments at a steady rate. The effect is the same. (Well, software is so often more complicated than I assume it should be, but ... )

[Urs]

Uhbik-D works like any common tape echo/delay - you have a single "endless tape" with one monophonic recording head and 5 playback heads. The signal of each playback head can be mixed/panned to the output and/or it can be fed back into the input.

As a special function, if no delay is set to any feedback then a sixth tap set to sync at 1/1 is taken for the feedback, but it isn't mixed into the output.

[Genetic_Junk]

I understand now. Thanks xh3rv and Urs for helping.