cudnylon

I've been going through countless threads but nothing seemed to help. I designed a delay with a tapin~ going to a tapout~ with 16 taps (8 left, 8 right) and a feedback loop. i'm getting major volume issues comparing the dry and wet sound. also the dry/wet slider crackles when moved back and forth. finally i have a feedback loop with a filter set up to send all 16 taps back to the input, but that isn't working either.

- is using normalize~ for 8 outputs each the best method or should i use *~/something else?

- how can i eliminate the slider noise. i tried to use the ($1 50) > line~ method but it still clicked, sometimes even worse.

- why is the feedback loop completely nonfunctional?

i gather all these issues are coming from the same source (volume issues).

Here is a quick example patch of what i'm dealing with (messages send delay time to taps, which also make a click when selected).

problems-with-delay-using-multiple-taps-volume-issues

havent looked at your patch, but i guess you overlooked that 

two outputs of a [tapout~ 100. 200.] are a regular case of 

having 2 parallel signals, which means that if you mix them 

together, you must ether lower the volume of the sum, or at

least make sure that the individual signals do not exceed -1. 1.

in other words, any kind of autogain or limiter inside the

feedback loop must be implemented per tap. (and not only one.)

what might also work in practice, (in a situation where feedback is

not going to be over ~0.95, your effect could be built not to 

allow higher values) is to lower the input into the delay

section by doing [clip~ -1. 1.] [*~ 0.05] right on the main 

input. that may make any kind of gain control inside the feedback

havent looked at your patch, but i guess you overlooked that 
two outputs of a [tapout~ 100. 200.] are a regular case of 
having 2 parallel signals, which means that if you mix them 
together, you must ether lower the volume of the sum, or at
least make sure that the individual signals do not exceed -1. 1.

in other words, any kind of autogain or limiter inside the
feedback loop must be implemented per tap. (and not only one.)

what might also work in practice, (in a situation where feedback is
not going to be over ~0.95, your effect could be built not to 
allow higher values) is to lower the input into the delay
section by doing [clip~ -1. 1.] [*~ 0.05] right on the main 
input. that may make any kind of gain control inside the feedback
unnecessary.

problems with delay using multiple taps, volume issues