Suggestions for ping-pong delay? (FV-1 content)

[Digital Larry]

One of these days I am going to get my Eclipse development environment working again so I can make some more progress on SpinCAD Designer for the FV-1 chip.  My laptop stopped working reliably and now I've checked my code out to a Linux system, but I'm dealing with lots of mystifying issues with the build environment. 

Regarding the program itself, I'm starting to turn my attention from the internal workings (though I do still really want to finish fixing the Ramp LFO simulation so that pitch shift algorithms can be adequately simulated) to tweaking and optimizing the existing blocks as well as adding some new ones along the way.

Right now I'm thinking about the Ping-Pong Delay.  Pretty sure I have this set up so that the long tap is at the end of the defined buffer and the shorter tap is at the halfway point.  Since there are two taps, they can be mixed to mono, or stereo with the input signal using a 3-1 mixer block or two 2-1 mixer blocks, respectively.

Feedback would naturally be taken from the longer tap, although with the feedback loop "block" I've added into the experimental build, feedback can be taken from anywhere to anywhere, so I might just delete the built-in feedback parameter. 

Question here is whether people typically want the shorter tap time adjustable.  Is it true?  And do you want it just freely adjustable, or do you want some help in constructing rhythmic relationships, and just what might those be?  And would you want that adjustment to be on a knob or fixed for the particular algorithm?

[SISKO]

Question here is whether people typically want the shorter tap time adjustable.  Is it true?  And do you want it just freely adjustable, or do you want some help in constructing rhythmic relationships, and just what might those be?  And would you want that adjustment to be on a knob or fixed for the particular algorithm?

It is true. Id make a -continuous- knob that when maxed it would be exactly at the halfway point

[Digital Larry]

Question here is whether people typically want the shorter tap time adjustable.  Is it true?  And do you want it just freely adjustable, or do you want some help in constructing rhythmic relationships, and just what might those be?  And would you want that adjustment to be on a knob or fixed for the particular algorithm?

It is true. Id make a -continuous- knob that when maxed it would be exactly at the halfway point

OK, when max'ed it's at 50% - how about when turned all the way to the left?  Going all the way to zero, especially for audible delays, doesn't seem very useful.

I could have two variations on this - a continuous knob and another one where the pot would be sliced into 8 segments, each with a specific setting.  The settings could be individually adjustable or follow some pattern of 1/16, 1/16:3, 1/8, 1/8:3, 1/4, 1/4:3, 1/2, 1/2:3.  Actually the triplet divisions are smaller, but I just realized that and didn't want to retype it.   

[idy]

Maybe your program set would have one Ping Pong with an infinitely variable tap and a second identical Ping Pong with a pot set up to give you set ratios. The ratios seem a practical way of doing things, several well designed pedals have them.

[Digital Larry]

Maybe your program set would have one Ping Pong with an infinitely variable tap and a second identical Ping Pong with a pot set up to give you set ratios. The ratios seem a practical way of doing things, several well designed pedals have them.

That's not a problem at all.  I'm just fishing for what types of delay ratios people like to use.  I was a little surprised by SISKO's request to have the max delay be at 50%, rather than allowing for a delay to appear closer to the end.  I don't typically use multi-tap delays in a rhythmic sense personally - I like shorter delays more for cheap ambience - less clutter than reverb.  

Any U2 fans?

[Digital Larry]

Here's the approach I've decide to take.  I'm designing a block called the Triple-Tap or perhaps Three-Tap delay.  Its maximum length will be settable from perhaps 50 msec to the full length of the delay RAM (1000 msec at 32678 Hz sampling rate).

Within that block will be three movable taps.  Each tap can have its maximum adjusted from 10% to 100% of the full delay length using a control panel slider.  Each tap will also have an independent delay time input.  When connected directly to a pot, the delay time when the pot is fully down will be about 10% of the max (control panel setting).  When the pot is fully up, the delay time for that tap will be 100% of the control panel value.

Each tap (and the end of the delay as well) will have individual audio output connections, so mixing, feedback etc. are entirely up to you.

By using the FV-1's incredibly useful scale/offset function, any number of different things can be made to happen.  Here's a few things that come to mind.

Set the full delay length to 1000 msec.
Set delay time 1 to 600 msec
Set delay time 2 to 700 msec
Set delay time 3 to 800 msec

If you connect the control inputs of all 3 delays to the same pot, or pot plus scale/offset, then the relative delay relationships between the three taps will be maintained even as you twist the pot.  I don't mean that the delay time difference between them will stay the same.  Rather, the ratios will stay the same.

Or, you can connect each delay time input to a different pot for individual control.
Or, you could have different scale/offset parameters between a single pot and the three delay time inputs so that the relative delay relationship CHANGES as you twist the pot.  This one might require some pencil and paper time to figure out how to get useful results.

Another thing that pops to mind is using this as a multi-tap chorus with the possibility of complex relationships between the various taps.  For example, you could add a Sin/Cos LFO and take the Sin output, through scale/offset to delay 1.  You could take Cos output, through a different scale/offset, to delay 2.  Finally you could add or multiply the sin and cos, and run through a Power function, plus of course scale/offset, to get each tap moving at the same speed but different depths/directions/control curves.

I'd also be interested in putting feedback around this, but from the tap mix point rather than just a single tap.  One paper I read recently (and my own experience) indicate that uncontrollable feedback occurs quickly with such a structure, unless you modulate the tap positions.  In that case you can get much more gain in feedback before howling occurs, although the end result is "ghostly voices".  Wow sounds great huh!    Anyway I'll be experimenting with all of this.

I'm having to overhaul some of the things I use to more easily make all this happen, so it may be a week or two before it's ready.

[Digital Larry]

I created the 3-movable tap block, with fixed center and end taps.  I'm still not exactly clear what creative use can be made of the fixed taps when the delays are kinda long.  I tried adding feedback from the end along with the 3 movable taps set for a fairly short time, resulting in a sort of

brrapp.....  brrapp... brrapp... sound.

With the delays much shorter and the input connected to an LFO, one can sweep the movable delay past the fixed center delay tap for a through-zero flanger type effect.  Thought I got a bit of this, need to experiment more.

Armed with my newly created tools, I quickly created a ten-tap delay with individually adjustable gain and time for each tap.  The first 5 taps are mixed to one output and the second 5 are mixed to a separate output.  All of these design choices are somewhat arbitrary and I'm looking forward to experimenting with them a bit to see what works best.  It would be easy to create a block with more taps and/or differing mix/output/fixed tap options, although the control panel to adjust everything is starting to take up significant vertical space on the screen!

I have just a couple more things to add before I put out a new beta release of SpinCAD Designer at my forum.  It may take a week or so!  Some things such as supporting "Undo" turned out to be more difficult than anticipated.

[ElectricDruid]

Armed with my newly created tools, I quickly created a ten-tap delay with individually adjustable gain and time for each tap.  The first 5 taps are mixed to one output and the second 5 are mixed to a separate output.  All of these design choices are somewhat arbitrary and I'm looking forward to experimenting with them a bit to see what works best.  It would be easy to create a block with more taps and/or differing mix/output/fixed tap options, although the control panel to adjust everything is starting to take up significant vertical space on the screen!

If you've got ten taps, why not try alternate ones to left and right for a bit of stereo space? (You said you'd got two outputs, or did I misunderstand?)

T.

[Digital Larry]

You can place any of the ten taps anywhere along the length of the delay.  The "first 5" are simply the ones that go to output #1.  There is also an 8-tap block with equally spaced taps preconfigured for alternate left/right output.  These things all have different applications and alignment with people's tastes.  Also given the FV-1's pedal orientation, lots of people use it mono even though I think that's missing a lot.