Softpot Ribbon Controller

Scott Stites · Posted: Tue Aug 12, 2008 1:00 pm Post subject:

Where do you get them there math hats? Very Happy

I've got 2M2 hysterisis resistor, bipolar powered comparator through a diode to CMOS.

widdly · Posted: Tue Aug 12, 2008 7:39 pm Post subject:

Scott Stites · Posted: Tue Aug 12, 2008 7:49 pm Post subject:

Scott Stites · Posted: Wed Aug 13, 2008 7:09 am Post subject:

The differential voltage works so well, I think it would be possible to incorporate a quantizer, yet maintain analog glide. My initial objection to quantization was that I didn't want any steppiness in glides, etc. However, using the differential voltage, the initial note could be quantized and the differential voltage could be summed with the quanitzed voltage to produce analog slides, etc.

At the appendage up position, the DAC would grab that value and hold it. It would have to be a fairly high resolution DAC to snag the appendage up voltage without some apparent shift in voltage (pitch if its controlling a VCO).

All that's down the road. I want to finish the basic stuff first. Re-triggering is what I'll look at next, but probably not until the weekend.

I'd also like to have a basic design posted if people want to use that without all the jewjaws before I get into all that. In fact, quantization with scales and all that good stuff would be best under FPGA or PIC control, which is not my field of knowledge at all. Maybe a board that takes the basic analog voltage and timing info from this one and does all the cool scaling, etc. would be the ticket. Some sort of State Machine. Maybe the boards could be called "The Finger" and "The Digit". Laughing

frijitz · Posted: Wed Aug 13, 2008 7:47 am Post subject:

Scott Stites · Posted: Wed Aug 13, 2008 10:24 am Post subject:

Thanks Ian, that's a pretty good idea! Now if I can get my head around these differentiators.

I guess deriving the re-triggers is going to be a balance between how much voltage movement will be required vs how fast it's moving. The voltage to excite the re-trigger will have to be moving faster than the fastest appendage can slide up or down the ribbon (a point above a fast slide so that hammer-ons can be detected accurately). Yet, a hammer-on voltage interval of (at minimum) a half step would desirably be enough to do it.

Scott Stites · Posted: Wed Aug 13, 2008 11:43 am Post subject:

I was just thinking that a cool thing to do would be to provide an output/function that held the initial note only, which could be bent around with the differential voltage.

Then it occurred to me - I already did! Laughing

The "reference" sample and hold, which is differentially summed with the position sample and hold (S&H3) is already holding that information. All I would have to do is mix the differential voltage with that output and...bam!

One app that applies right away (along with the obvious pitch bend) is to be able to hammer a note, then slide the filter around that. Or, one could connect the differential output to a vca, hit a note, then control the volume of the note by sliding the appendage up the ribbon. A special version of the differential voltage where negative excursions get inverted to positive would allow one to even fade in the highest notes by sliding down.

frijitz · Posted: Wed Aug 13, 2008 11:45 am Post subject:

Scott Stites · Posted: Wed Aug 13, 2008 11:24 pm Post subject:

Scott Stites · Posted: Fri Aug 15, 2008 7:19 am Post subject:

There are a number of opportunities to make this thing a really versatile instrument. Hopefully it can provide a lot of versatility, yet remain fairly easy to set up and use.

The greatest power, of course, is that it provides a simple control surface with which to generate and control output voltages and synchronization. Right now, the number of unique unique simultaneous control voltages it produces is nine (not counting some inversions or signal input mode). Differential voltage really let the cat out of the bag - six of the control voltage responses were derived from that function.

Right now I'll try to describe the three main generated voltages from which all the others are derived:

1. Continuous Positional Output Voltage. While moving up or down the ribbon, the voltage tracks the position. When pressure is released, the voltage retains the last positional voltage that was generated before pressure is released.

2. Fixed Positional Voltage (CV2). Each time pressure is applied, the initial positional voltage is sampled and held. While retaining pressure on the ribbon and moving the appendage up or down on the ribbon, this voltage does not change. Its action on release depends on if "hold" or "normal" is selcted. If "normal", the fixed positional voltage will go to the last positional voltage sampled by CV1 before pressure is released. If "hold" is selected, the voltage will remain unchanged until the next pressure is applied, whereupon it will sample that initial value, and the process repeats. Mixing differential voltage with this voltage will provide a means of sliding or vibrato.

3. Continuous Differential Voltage (Diff CV). When the ribbon is first pressed, this output is at zero volts. When the appendage slides "above" the original point of contact, while pressure is applied, the differential voltage will rise (positive) relative to the original position. When the appendage slides "below" the original point of contact, the voltage will fall negative, relative to the original position. For pressure release, like CV2, its behavior depends on if the "Hold" or "Normal" mode is in effect. In normal mode, the voltage will return instantly to zero (if it's at some non-zero value) when pressure is released. In hold mode, it will lock to the current value when pressure is released. In that case, when the next pressure event starts, it will return to zero volts, and the process will repeat.

I've uploaded a block-like schematic of what I have working so far (it all works as of this writing).

All three voltages are derived from manipulation of two different Sample and Hold's - S&H3 and S&H4 (S&H1 and S&H2 are part of the analog shift register responsible for grabbing stable voltages for S&H3).

Starting with note off - the comparator is low. Flipflop 1 Set input is high, so Flipflop 1 Q output is high.

If the CV2/Diff Hold switch is closed, this high signal is transferred straight to the S&H4 clock input, which allows it to track its signal input, which is fed from the output of S&H3. The output of S&H3 and S&H4 are therefore at the same potential, so there is 0V differential - the Differential CV is 0V. CV1 and CV2 are at the same potential as well - which would be the last voltage sampled by S&H3.

If the CV2/Diff Hold switch is open, the clock input of S&H4 is low, so it is holding the last sample it took - which was the voltage initially loaded into S&H3 at the initial press of the ribbon the last time it was pressed. The differential voltage will be the value of S&H3 voltage plus or minus S&H4 voltage.

In this state, the reset pin of Flipflops 2 and 3 are held high, which holds their Q outputs low, and /Q outputs high. Because they are held in reset mode, they ignore the clock pulses coming from the third output of the CD4017, which is used to sequentially clock the first three S&H's.

When the ribbon is pressed (note on) Set goes low on Flipflop 1. The Q output does not go low yet - it's waiting to be reset by Flipflop 3.

At the same time, Reset on flipflops 2 and 3 also goes low. This allows Flipflop 2 to be clocked by the third sequential clock pulse from the CD4017. These two flipflops form a divide by four counter. This is intended to give S&H3 plenty of time to acquire a stable voltage before its output is sampled by S&H4 (by plenty of time, I don't intend to imply observeable latency - there is no apparent latency).

At the rising edge of the fourth clock pulse, Q of Flipflop 3 goes high. This resets Flipflop 1, which brings its Q output low and its /Q output high. When /Q goes high, this initiates a pulse that is blocked by the switch's action on the AND gate feeding the summing OR gate on the input of the S&H4 pulse input.

If the CV2/Diff Hold switch is closed, S&H4 been constantly tracking S&H3 already. When Flipflop 1 Q goes low, this locks the current voltage on the ouput of S&H3 into S&H4. The voltage on CV2 is now fixed at this same voltage (because CV2 always reflects the contents of S&H4). As the appendage is dragged up and down the ribbon, S&H3 voltage varies and so does the differential CV output, because its output is added and subtracted with the fixed output of S&H4.

If the CV2/Diff Hold switch is open, S&H4 been holding the initial voltage value recorded when the ribbon was first pressed on the previous note - it hasn't been tracking S&H3. When Flipflop 1 /Q goes low, this creates a sample pulse that "locks" the current voltage on the ouput of S&H3 into S&H4, creating a new initial note voltage. CV2 puts out this same voltage now. As the appendage is dragged up and down the ribbon, S&H3 voltage varies and so does the differential CV output, because its added and subtracted with the fixed output of S&H4. The voltage on CV2, which is S&H4s output voltage, remains fixed at the initial voltage of S&H3 when the pressure was first sensed on the ribbon.

Once again pressure is released. If the switch is closed, S&H4 locks the last value of S&H3 onto its output. This causes the differential to snap back to 0V (if it was at some other value) and causes CV4 to snap to the last positional voltage recorded by S&H3.

When pressure is released and the switch is open, S&H4 does not take another sample of S&H3 and remains fixed at its current voltage, which was the initial note voltage recorded when the ribbon was pressed. The differential output stays at whatever value it was at when the ribbon was released - it will only return to 0 at the start of the next note.

Cheerios,
Scott

appendage_rough1_nov.PNG

Description:

CV1, CV2 and Differential Control

Filesize:

82.27 KB

Viewed:

376 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Scott Stites · Posted: Fri Aug 15, 2008 8:42 am Post subject:

Here's a diagram of the six control voltage permutations I mentioned in the previous post.

The bottom of each graph illustrates that the initial point of contact is always 0V. This can be anywhere on the ribbon controller, not just the center. The horizontal line at the bottom shows the direction of movement on the ribbon cable while applying pressure after the initial contact. The graph above shows the direction of travel and polarity of voltage on each CV output as the appendage slides up and down on the controller from the initial contact point.

This is not just kitchen sinking for kitchen sinking's sake - I've tried these out, and they can have some very profound uses. Take number 5 for example: say you are using CV2 (fixed initial note, no slide) and you're using a slide control out to control filter cutoff. If you hit a note at the upper end of the ribbon, you don't have any more room to slide up. However, you've got plenty of room to slide down, which will also open the filter just as sliding up from the initial point of contact will do (and sliding back up to intial will close it back down). In fact, playing this really opened my eyes to how musical opening a filter on oscillator slides going up and going down is. I used the continuous output to slide a VCO, and the number 5 output to control the filter cutoff. The initial note is darker - as the note slides up, it gets brighter (as is normal). Sliding back down also opened up the filter for some gnarly bass.

This arrangement is also very, very good for smaller ribbon cables that would be used as controllers (like a pitch or modulation wheel). In that case, you could view it as sort of a one axis JAG. Very Happy

Cheerios,
Scott

rib_control_diff_perm.PNG

Description:

Differential CV Permutations

Filesize:

12.25 KB

Viewed:

248 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

fonik · Posted: Fri Aug 15, 2008 10:31 am Post subject:

this is very exciting, scott. great, great, great. i jst printed out your last two posts and will read them carefully.
first the klee and now this controller! thank you, scott for sharing.

Scott Stites · Posted: Fri Aug 15, 2008 12:05 pm Post subject:

I'm pretty excited about it, Fonik! Actually, if everything works out to how I'm envisioning it, it'll be a nice mate for the Klee. The Klee will be able to feed it control voltages and triggers, and the ribbon cable will be able to feed the Klee the same. I think it would make for some pretty cool interaction with the Klee.

Hey Ian: The idea for retriggering with a differentiator is looking like it's going to work just fine. I plan on feeding the differentiator with permuation 6 on my differential permutation chart - that way if a step is of a particular interval up or down, the waveform seen by the differentiator will be the same.

I'm shooting for an interval of, at minimum, a half step (83 mV) to excite the retrigger.

I think I might get away without using a lag. The target voltage here won't move nearly as quickly as the breath controller would at the point the re-trigger would be used. I set up a differentiator with a gain of 3.3.
Sending a negative going 83 mV pulse to it generates a positive spike of around 140 mV. If I feed it a negative going 7V triangle wave at ~4 Hz, the differentiator output stays below 100 mV. I figure a triangle wave at 4 Hz would be like sweeping up and down the entire ribbon four times a second at a very high range setting, so hammer-ons of a halfstep or so in either direction ought to retrigger while slides are ignored.

Once I confirm that all the above does actually work, the next trick will be defeating the re-trigger on (non-trilled) note-on and note-off events. I think I can solve that with some timing signals and CMOS logic.

Cheerios,
Scott

frijitz · Posted: Fri Aug 15, 2008 9:45 pm Post subject:

Scott Stites · Posted: Fri Aug 15, 2008 10:34 pm Post subject:

Hey Ian,

State Machine · Posted: Sat Aug 16, 2008 7:48 am Post subject:

Oh dear, I have to catch up here .......... Scott, absolutely fantastic stuff here. Ian, great dialog as well and contributions .......... OK, I can say anything intelligent until I read more here ............. I can say this, the amount of control will be astounding Laughing

Reading more here .............. and, wow, I can't believe it Scott, your allowed to stay up that late .......... geez ....... Wink

Oh, hey Fonik !!!!!!!!!

Bill

State Machine · Posted: Sat Aug 16, 2008 7:51 am Post subject:

Coriolis · Joined: Apr 11, 2005 Posts: 616 Location: Stilling, Denmark

Will there be a lot of talking here in the future of "Giving Klee the Finger"? Rolling Eyes

C

Scott Stites · Posted: Sat Aug 16, 2008 11:48 pm Post subject:

I figured I'd leave the name "The Finger" for Eric Barbour to use. Laughing

I recorded a couple of quick samples tonight. Bear in mind (again) that this ribbon is just flopping loose, so playing it is like trying to hit a moving target. I basically have to try to hold it by the edge while molesting it. About 1/4 of the left side (low notes) is hanging up in the air attached to the breadboard, so it's unplayable. Basically I'm just pecking around on it.

Anyway, the samples demonstrate a few different things - retriggering on hammer-ons, modulation by sliding, and the difference between CV1 and CV2.

Each sample uses the same patch (I recorded the two sequentially). The patch is two VCOs through the 2040 filter through a VCA, with a little reverb. The trigger out and gate out are connected to an ADSR envelope generator. The envelope generator is providing the envelope for the VCA and the filter. The differential output, in Mode 5, is also controlling the filter cutoff. This CV works by starting at 0 wherever the ribbon is pressed. Then, while pressing the ribbon controller, moving right or moving left from that point will increase the CV in a positive direction. The Appendage is in CV hold mode, so when the pressure is released, the differential output stays at the voltage it was at until the next pressure is applied, whereupon it returns to zero and uses the new initial position on the ribbon as its zero point.

I have the EG set with a rather pointed attack/decay, with a low sustain to demonstrate the retriggering.

appendage_cv1_retrigger.mp3: CV1 will always glide in pitch as the ribbon is pressed and the pressure slides up and down the ribbon. Because the Appendage is in hold mode, when pressure is released, the pitch stays where it's at until the next pressure point is applied, whereupon the pitch goes to the new contact point's position. As pressure is kept on the ribbon, moving up and down causes the filter to open up - it closes when it passes the original pressure point. Releasing the pressure then re-applying it starts the filter at the original level. That's why some notes are more subdued and others are brighter - it's very easy to control at will.

Re-triggering applies to hammer-ons. While pressure is applied, hammering on a different point (up or down) will cause a trigger to fire (past a certain interval) on the hammered note - this is re-trigger pulse 1. When the hammer-on note is released, a second trigger fires - this is re-trigger pulse 2. Both of these triggers can be made available as separate outputs to control external events. In this patch, they're mixed with the initial note-on trigger.

appendage_cv2_retriggin.mp3: CV2 does not naturally glide. It grabs the initial voltage value wherever the ribbon is pressed and holds that note. Sliding while pressure applied still applies the modulation. All of the filter sweeps heard in this sample are simply me sliding up and/or down on the ribbon while keeping pressure applied. Releasing and applying pressure again sets the voltage to zero wherever the ribbon is pressed, and that is the new "reference point" for the CV. It starts off with the same voice as the CV1 sample. After a little while, I increase the resonance of the filter so the sweeps can be more clearly heard.

Before the two minute mark, the Appendage is re-triggering like it was in the CV1 sample. Now hammering on triggers the same note, but with different filter cutoffs. At the two minute mark or so there is a pause as I disconnect from the combined trigger output and go to just the intial note-on trigger only. At the end of the sample, you can hear the difference in how the filter hammer-ons work.

BTW, CV2 can be made to slide simply by mixing in either CV1 or one of the differential voltages. This same voltage used to modulate the filter could be used to bend the note up as one rolls the finger to the left or the right - it's actually a very sweet sounding vibrato. I didn't have it in this sample though...

Cheerios,
Scott

appendage_cv2_triggin.mp3

Description:

Appendage CV2 Output - Retriggered the single trigger

Download (listen)

Filename:

appendage_cv2_triggin.mp3

Filesize:

4.98 MB

Downloaded:

1192 Time(s)

appendage_cv1_retrigger.mp3

Description:

CV1 Output, Retriggering

Download (listen)

Filename:

appendage_cv1_retrigger.mp3

Filesize:

3.48 MB

Downloaded:

1132 Time(s)

Coriolis · Joined: Apr 11, 2005 Posts: 616 Location: Stilling, Denmark

Man, you are selling this thing very well...

C

Scott Stites · Posted: Sun Aug 17, 2008 8:05 am Post subject:

If one were to set up a selector switch to a mix input from any of the differential permutations, that would allow one to set the response of CV2 to any slide response one wanted for that CV.

1. Normal: Fixed initial, no slide

2. Slide 1: Left slides down, right slides up

3. Slide 2: Left stays fixed, right slides up

4. Slide 3: Left slides down, right stays fixed

5. Slide 4: Left slides up, right stays fixed

6. Slide 5: Left stays fixed, right slides down

7. Slide 6: Left Slides up, right slides up

8. Slide 7: Left slides down, right slides down

Then, of course, one could invert the selection. This would bring in the "left slides up, right slides down". Very Happy

Taking it a step further, one could provide a selector switch for CV1 that did sort of the same thing, though CV1 slides naturally. By sort of, I mean the responses would be different yet again (because of summation and subtraction).

Going in the same order, the permutations would have this effect:

1. Left slides down, right slides up (normal).

2. Left slides down at twice the rate and right slides up at twice the rate.

3. Left slides down at normal rate, right slides up at twice the rate.

4. Left slides down at twice the rate, right slides up at normal rate.

5. Left stays fixed and right slides up at normal rate.

6. Left slides down at normal rate, right stays fixed.

7. Left stays fixed and right slides up at twice the rate.

8. Left slides down at twice the rate, right stays fixed.

This could be inverted as well. In that case, inverting the response of the permutation as it's mixed in would cause number 2 to create no slide at all.

Reason I mention this is I've got a cool patch running that has a high oscillator controlled by CV1 and a low oscillator controlled by CV2, only CV2 now slides up left or right. The interplay is cool. Cool

This stuff could go on forever, you realize......

State Machine · Posted: Sun Aug 17, 2008 8:25 am Post subject:

Scott, the results are amazing. Great expression for a ribbon just hanging in the breeze !!! Just imagine when it's actually mounted to a flat surface. Cool

Great research and development .... Wink

Scott Stites · Posted: Sun Aug 17, 2008 12:46 pm Post subject:

Well, certainly I'm going to try inserting a DAC. Main thing is the DAC will need to be high resolution but inexpensive and not unobtainium (I'm considering stacking two DAC08s and choosing the resolution from the 16 bits supplied).

This would eliminate the analog shift register, though I'd still need to maintain two S&Hs for the CV outputs. They would need only be simple S&Hs - as long as they don't have any significant lag in the ms or less range. Very Happy

The DAC would also tighten up timing; the analog shift register is a hedged bet that the ribbon voltage has been acquired.

However, the DAC is primarily for infinite hold, but I'm willing to put in defeatable 0.083V/step (half step musical interval) quantization on the outputs as well. Slide would still be possible via analog means, so it wouldn't be steppy. Just the initial note value would be quantized.

The nature of the design will preserve the availability of the control, gate, and trigger outputs (all of the features) discussed here.

The main challenge for me was a way to acquire voltages and hold them in "memory", with the parts available to me, while at the same time generating gate and trigger signals (my original intent did not include re-triggering, but Ian showed me the light on why that would be a good thing). Now that I have an idea of the timing, I think a DAC will eventually take over a bit of the work load (I'm thinking I might be able to use it to discern re-triggering as well). A good basic core design will be nice for those who want a simple frikkin' modwheel-scale controller, thank you very much, which is something I'd also like to have. The core will be expandable to the big Kahuna, which I also have every intention of building. The differential voltage perms are a very powerful thing, even for a small modwheel style controller.

But first, I need to firm up/simplify/document this design. DAC probably won't happen 'til after Malaysia.

Cheers,
Scott

Scott Stites · Posted: Tue Aug 19, 2008 11:39 am Post subject:

I took a break from the Appendage BB on Sunday, for the most part. Yesterday I moved the main gate/trigger stimulus to the output of the divide by four counter. I noticed that synchronization with differential and its permutations would cause a 'knock' on the first new note if there was a wide gap (when using differential CV to modulate something). That's now fixed.

I recorded most of the breadboard circuitry yesterday, with the exception of the trigger generators and differentiator values. I need to re-organize the breadboard - it's crowded with a lot of unused stuff. I killed a CD40106 yesterday because of that, so I'm documenting it as it is now.

Having said that, the schematic is really not all that complicated. There are a few things I want to try to enhance it a bit more.

One of the things that makes it rather difficult to use yet for some aspects of pitch control is that the pressure (area of finger) laid on it will cause the pitch to shift. For pressure-related vibrato, that's pretty nice. For trying to end a held note at a specific pitch, it can be a bitch - it very easily can drop as the finger is removed. Part of it is due to technique - how one removes the finger, for example, but I think it could be made easier. Tapping the finger for short staccato notes doesn't incur this particular wrath, BTW. CV2 is immune as well, as it grabs only the initial pitch value (unless, of course, it's mixed with differential voltage to provide a slide).

The softpot is very sensitive, and there's a fine line between enough pressure and no pressure. It's fairly difficult to unflatten the finger enough to get the desired ending pitch without inadvertently triggering another note. What I plan to do is create a release time on the order of milliseconds - if when releasing the pressure, there are brief openings and closures, it will not produce a gate/trigger event, but instead just update the sample and hold. After a certain amount of time expires (unnoticeable during normal play), then it will accept a new note. This (I think) will make it easier to gradually withdraw pressure (unflatten the finger) to end with the desired pitch during slide play.

It would be very easy to add a "hold" switch - a momentary switch that ensures the pitch you're at will not deviate, no matter how you remove the finger. That breaks my philosphy of having this a one appendage device, but, at the same time, would be a very simple thing to operate.

I have another line of attack planned as well - this will serve to make pitch deviation on release an easier thing for fast, heavy play. As mentioned before, mixing differential voltage with CV2 causes a very nice, controllable slide. CV2 itself is fixed at the initial value when pressure is applied - the differential voltage slides it up and down past this point. The idea is to be able to modify the differential voltage to have a certain space between upward and downward movement where the voltage doesn't move at all. This would have the effect of placing the finger on the ribbon, then having a "null" spot of variable width. This null spot would be at the initial note value. Sliding up past the right side of the null would raise the pitch, sliding down past the left hand side of the null would lower the pitch. This would keep pressure deviation from happening at the initial note value (though that would go away once one slid past the null). Better still, though, it could provide a nice, more easily nailable point to slide to. Say you enter a note, slide up, then want to slide back to that original note. This would make that very easy to do.

On the kitchen sink front, I think it would be entirely possible to turn the single axis ribbon output to a dual axis output. Left finger would provide an output and right finger would provide a separate output. Sort of like a two axis joystick split out on a single axis. But, that comes later (it will require a bit of re-tooling of this idea, which I want to thoroughly wring out first).

Cheerios,
Scott