Anyone produce highly sampled hi-hats?

[brainwreck]

Originally Posted by Glennbo

Funny you should ask, I just rebooted out of Linux and back into Windows to find that out for myself, and the answer is *eight*. Oh, and I mis-typed the total number of files by 400. The complete total number of files for eight hihats is 7,423 or roughly 927.875 samples per each hihat. There is a separate folder for Ambience that has around 8000 files in it.


09/29/2003 09:38 PM 309,010 TO05_04_DS_FH_R_L04.wav
09/29/2003 09:38 PM 313,430 TO05_04_DS_FH_R_S01.wav
09/29/2003 09:38 PM 322,868 TO05_04_DS_FH_R_S02.wav
09/29/2003 09:38 PM 316,084 TO05_04_DS_FH_R_S03.wav
09/29/2003 09:38 PM 321,886 TO05_04_DS_FH_R_S04.wav
7423 File(s) 2,807,448,704 bytes

D:\Superior\Drummer\Hats>

-

I'm assuming here that the ambience files are for every kit piece, not just hi-hats.

[Glennbo]

Originally Posted by brainwreck

I'm assuming here that the ambiance files are for every kit piece, not just hi-hats.

Correct. Just looking at the file names, I can't really tell how many files in the Ambiance folder are part of the hihats.

[Tod]

Hi Glennbo, have you ever measured the way your hi-hat pedal is putting out CC4? Does it put out the
full 0 to 127 CC4 values, and if so, how smoothly?

SMM and I have been discussing this some and we think we can come up with a fairly natural hi-hat, but
I'm wondering if the midi hat pedals themselves might not be a determining factor in how it works for
e-drummers.

[brainwreck]

Originally Posted by Tod

Hi Glennbo, have you ever measured the way your hi-hat pedal is putting out CC4? Does it put out the
full 0 to 127 CC4 values, and if so, how smoothly?

SMM and I have been discussing this some and we think we can come up with a fairly natural hi-hat, but
I'm wondering if the midi hat pedals themselves might not be a determining factor in how it works for
e-drummers.

With current drum samplers I think this doesn't matter, because we're only looking at something like 4-8 levels of openness. So then, 4-8 CC values will access all of those sample sets.

I don't know what the definition of smoothness would be here, since we're dealing with 128 discrete MIDI values mapped to 4-8 discrete sample sets. If you mean, a consistent line or curve of CC values in relation to pedal movement, the only way to know that would be to set up a ruler on the pedal (there is only around an inch or so of travel) and graph CC values to pedal increments. So the measurement increments would need to be something like 1/128", which is damn small for just eyeballing. A dial caliper would probably make for a better time than a ruler here.

Edit: On my pedal (FD-8) the travel is ~3/4". So in my case, that should be ~0.006" increment of travel per CC value out of 0-127. My drum module (TD-11) puts out 0-90 CC values, which I have altered via a jsfx to 0-127 (resulting in larger CC increments per increment of pedal movement, of course).

[Glennbo]

Originally Posted by Tod

Hi Glennbo, have you ever measured the way your hi-hat pedal is putting out CC4? Does it put out the
full 0 to 127 CC4 values, and if so, how smoothly?

Here's a screen shot of pressing my ancient Roland FD-6 hihat controller from full up to full down, and then starting to go back up. This is as much as I could grab off the screen.

[Tod]

Originally Posted by brainwreck

With current drum samplers I think this doesn't matter, because we're only looking at something like 4-8 levels of openness. So then, 4-8 CC values will access all of those sample sets.

Hi brainwreck, so you are actually only using one CC value per sample level, is that what you mean?

I don't know what the definition of smoothness would be here, since we're dealing with 128 discrete MIDI values mapped to 4-8 discrete sample sets. If you mean, a consistent line or curve of CC values in relation to pedal movement, the only way to know that would be to set up a ruler on the pedal (there is only around an inch or so of travel) and graph CC values to pedal increments. So the measurement increments would need to be something like 1/128", which is damn small for just eyeballing. A dial caliper would probably make for a better time than a ruler here.

Edit: On my pedal (FD-8) the travel is ~3/4". So in my case, that should be ~0.006" increment of travel per CC value out of 0-127. My drum module (TD-11) puts out 0-90 CC values, which I have altered via a jsfx to 0-127 (resulting in larger CC increments per increment of pedal movement, of course).

Aah, 3/4", that could explain why there are only 91 values (0 to 90). I imagined the pedal travel would be much more then that. Still, if the
pedal were able to put out 128 continuous CCs, I think a natural transition from open to tight could be made. Some adjustment would have
to be made for pedals that don't put out 128, like your pedal. As you say, the jzfx don't give you more values, it just spreads them out a little,
which I can see, would make a difference for you.

[brainwreck]

Originally Posted by Tod

Hi brainwreck, so you are actually only using one CC value per sample level, is that what you mean?



Aah, 3/4", that could explain why there are only 91 values (0 to 90). I imagined the pedal travel would be much more then that. Still, if the
pedal were able to put out 128 continuous CCs, I think a natural transition from open to tight could be made. Some adjustment would have
to be made for pedals that don't put out 128, like your pedal. As you say, the jzfx don't give you more values, it just spreads them out a little,
which I can see, would make a difference for you.

No, but given the limited number of openness sample sets, I could do that with current drum samplers and the result would be the same. In other words, if a sampled hi-hat is limited to 8 openness sample sets, I could very well use just 8 CC values, because the switching among openness samples is very course (limited to 8 max sample sets) any way.

Acoustic hi-hats usually have somewhere around that size gap/pedal travel, depending on the player of course and whether a fancy cam mechanism is being used for hi-hat. I have seen some players use a gap such that you can't look and tell that the hi-hat isn't closed, and I have seen others where there is more than an inch gap. I think it is style dependent. But for an e-hat, an adjustable amount of pedal travel for the same CC output values would be nice (which I could do using a hi-hat stand and a goedrum hi-hat controller), along with sampled hi-hats having more sample sets. But a 3/4" pedal travel shouldn't be a hinderance in way of what we are discussing. It's a pedal travel amount, unrelated to CC output by the module. I probably use something around that amount of pedal travel on a real hi-hat (I think less). The problem with the TD-11 is that Roland chose to limit the CC output to 0-90 rather than 0-127, hence having to put together a jsfx plugin to get 0-127. But again (this probably isn't driving home with you because you aren't using and examining an e-hat in person), given the courseness of sample switching, this shouldn't matter.

[Glennbo]

Originally Posted by Tod

Hi brainwreck, so you are actually only using one CC value per sample level, is that what you mean?



Aah, 3/4", that could explain why there are only 91 values (0 to 90). I imagined the pedal travel would be much more then that. Still, if the
pedal were able to put out 128 continuous CCs, I think a natural transition from open to tight could be made. Some adjustment would have
to be made for pedals that don't put out 128, like your pedal. As you say, the jzfx don't give you more values, it just spreads them out a little,
which I can see, would make a difference for you.

Sounds to me like his pedal is either screwed up, or Roland has changed the way they work, coz mine outputs the full entire spectrum of 0-127. My Roland FD-6 has a lock nut that you use a drum key on to unlock so you can mess with the travel of the pedal, but I've never changed mine from the stock factory position. I'm using this with an equally ancient Roland TD-6 controller brain.

BTW, I posted this on my Studio Tests site the other day. It's a test of my hihat controller driving Superior Drummer 2.

https://soundclick.com/share.cfm?id=13756019

[brainwreck]

Originally Posted by Glennbo

Sounds to me like his pedal is either screwed up, or Roland has changed the way they work, coz mine outputs the full entire spectrum of 0-127. My Roland FD-6 has a lock nut that you use a drum key on to unlock so you can mess with the travel of the pedal, but I've never changed mine from the stock factory position. I'm using this with an equally ancient Roland TD-6 controller brain.

BTW, I posted this on my Studio Tests site the other day. It's a test of my hihat controller driving Superior Drummer 2.

https://soundclick.com/share.cfm?id=13756019

Output of CC4 values is to do with the drum module, not the pedal. I think you two guys might be confusing pedal travel (a physical thing) vs. CC output value range. I imagine that the FD-8 and FD-7 are very similar pedals. The FD-8 has an adjustment screw for pedal travel, and the max travel is around 3/4".

The actual difference here is that the TD-11 outputs CC4 values of 0-90 rather than 0-127.

Then there is the possible confusion of CC4 values to sample set openness. It might not be clear here that there is not an an actual continuous change going on for hi-hat openness in drum samplers. There is a discreate switching among a very limited number of sample sets. So if a drum sampler only has say 8 sample sets for openness, having more CC4 values than that (and finer increments) has no benefit.

So just to make sure that we are all on the same page: Do you guys understand that pedal travel amount, CC output range, and the number of sample sets for openness are discrete things?

[Glennbo]

Originally Posted by brainwreck

Output of CC4 values is to do with the drum module, not the pedal. I think you two guys might be confusing pedal travel (a physical thing) vs. CC output value range. I imagine that the FD-8 and FD-7 are very similar pedals. The FD-8 has an adjustment screw for pedal travel, and the max travel is around 3/4".

The actual difference here is that the TD-11 outputs CC4 values of 0-90 rather than 0-127.

Then there is the possible confusion of CC4 values to sample set openness. It might not be clear here that there is not an an actual continuous change going on for hi-hat openness in drum samplers. There is a discreate switching among a very limited number of sample sets. So if a drum sampler only has say 8 sample sets for openness, having more CC4 values than that (and finer increments) has no benefit.

So just to make sure that we are all on the same page: Do you guys understand that pedal travel amount, CC output range, and the number of sample sets for openness are discrete things?

I knew my TD-6 was responsible for the range of CC messages, but the foot controller has a potentiometer or possibly photo-optical sensor (I've never opened mine up to see what's inside it, but it is mechanical), so I see possibility of the pedal not outputting the full range, which could result in less than the full range coming out in the form of CC messages from the brain.

BTW, I do have the "Thin Data" setting in my TD-6 enabled, which keeps it from sending even more CC data than the screen capture I posted, so it has the capability to chunk even more linear data out than my example shows.

[brainwreck]

Originally Posted by Glennbo

I knew my TD-6 was responsible for the range of CC messages, but the foot controller has a potentiometer or possibly photo-optical sensor (I've never opened mine up to see what's inside it, but it is mechanical), so I see possibility of the pedal not outputting the full range, which could result in less than the full range coming out in the form of CC messages from the brain.

BTW, I do have the "Thin Data" setting in my TD-6 enabled, which keeps it from sending even more CC data than the screen capture I posted, so it has the capability to chunk even more linear data out than my example shows.

What I have read says that both the FD-7 and FD-8 use an FSR (force sensing resistor) for controlling input to the module. I don't know if that is correct.

I was a bit pissed at first that the TD-11 outputs so few CC4 values, but then I figured out that not very many CC4 values are required to access the full set of samples both in the module and in drum samplers. The only caveat being here that some drum samplers that don't have adjustments for CC4 values to openness ranges might require values greater than 90 to access some of the samples, which is why I put together a jsfx for the TD-11 for expanding output values 0-90 to 0-127 (for Ezdrummer 2).

By the way, you can test the hard switching of sample sets in your drum sampler by using the jsfx that I linked to above when discussing this issue with DeathbyGuitar. There is also a project in that post ready to fire up, only requiring you to insert your drum sampler after the jsfx on the track, hit play, and move the slider on the jsfx to hear the switching between openness ranges. You should notice that there is no continuous changing going on. At a given CC4 value, the sample set switches, otherwise staying exactly the same between CC4 values. But what might fool you here a bit is that there are hit variations at any given velocity range. In other words, within a given CC4 range, the sampler will rotate among a number of variations, but the openness of the samples will remain exactly the same throughout that CC4 range.

[ashcat_lt]

The CC value coming from the brain is (at best) 7 bits deep, but that basically gets downsampled by the sampler to the equivalent of 2-3 bits. The data is "stair stepped" or rounded to the nearest meaningful value. Resolution is completely dependent on the number of samples available. "Fixing" the pedal can't help. Massaging the CC data can't help. I'm not sure where the confusion is here.

I'm pretty sure this won't actually have acceptable results, but maybe if one was to map this CC to modulate the sustain/release envelope either instead of or on top of the sample switching...??? I'm not sure if EZD or AD can do that, but if so, or for folks with samplers that can, it might be worth a try.

@Glennbo - I notice in your pic that your brain puts out low values when wide open. That's kind of the opposite of what I expected, but I guess that's the way it works? This would mean that the person expanding from a max value of 90 is actually getting more closed samples out of it. Wide open should still be wide open either way.

[brainwreck]

Originally Posted by ashcat_lt

The CC value coming from the brain is (at best) 7 bits deep, but that basically gets downsampled by the sampler to the equivalent of 2-3 bits. The data is "stair stepped" or rounded to the nearest meaningful value. Resolution is completely dependent on the number of samples available. "Fixing" the pedal can't help. Massaging the CC data can't help. I'm not sure where the confusion is here.

I'm pretty sure this won't actually have acceptable results, but maybe if one was to map this CC to modulate the sustain/release envelope either instead of or on top of the sample switching...??? I'm not sure if EZD or AD can do that, but if so, or for folks with samplers that can, it might be worth a try.

@Glennbo - I notice in your pic that your brain puts out low values when wide open. That's kind of the opposite of what I expected, but I guess that's the way it works? This would mean that the person expanding from a max value of 90 is actually getting more closed samples out of it. Wide open should still be wide open either way.

^ Yep. The only thing that matters is that a given pedal position results in the module outputting a given CC4 value for triggering the desired sample set. As long as those relationships are set to preference, everything else (many CC4 values, etc.) is meaningless.

I think that modulating an amp envelope isn't going to do it. The greatest change of a hi-hat that is more closed or more open is the timbre. A hard closed hi-hat has a tap/tick sound; an ever so slightly more open hi-hat has that sound to a lesser degree with a touch of sizzle; those two relationships changing as the hi-hat opens up more, along with other things in the timbre and sustain. But something which might pass the sniff test well enough here for a limited number of sample sets is, modulating the mix between two neighboring sample sets for the inbetween CC4 values. But the hits would need to be phase aligned. And the end result might turn out to sound like a mess. The only way to know would be to try it. And an easy route without getting into all sorts of mapping and other issues would be to render out a number of neighboring sample sets, add them to a pair of tracks, align them, and setup parameter modulation such that one track's output descreases as the other track's output increases. Can you foresee any problems with that approach? If the result is ok, there may even be some further benefit in setting up envelope relationships between neighboring sample sets rather than overall mix of volumes.

[brainwreck]

I think I will give the above a go over the weekend, unless someone can provide reasonable arguments for why it wouldn't work.

[Glennbo]

Originally Posted by ashcat_lt

@Glennbo - I notice in your pic that your brain puts out low values when wide open. That's kind of the opposite of what I expected, but I guess that's the way it works? This would mean that the person expanding from a max value of 90 is actually getting more closed samples out of it. Wide open should still be wide open either way.

I also spent a lot of time in Superior Drummer 2 getting my hihat to have the majority of the samples assigned to the first quarter or so of pedal travel from fully depressed. Initially, I was hearing more sample switching than I liked hearing, and it was because I do more of the fine movement on my pedal (both on V-Drums and real drums) from fully closed to just barely letting my foot up off the pedal. When it was switching to the wide open samples too quickly, it made for real ragged sounding hihat work, but honestly, since I spent the time tweaking it up for the way I play, I am not bothered at all in the way it works now, no matter how many sets of samples are coming into play.

Only if I isolate the hihat where nothing else it playing do I notice any sample switching. If I'm playing along with other music, like when recording a track to a song, or playing one back, I really don't notice it at all.

[brainwreck]

Originally Posted by Glennbo

I also spent a lot of time in Superior Drummer 2 getting my hihat to have the majority of the samples assigned to the first quarter or so of pedal travel from fully depressed. Initially, I was hearing more sample switching than I liked hearing, and it was because I do more of the fine movement on my pedal (both on V-Drums and real drums) from fully closed to just barely letting my foot up off the pedal. When it was switching to the wide open samples too quickly, it made for real ragged sounding hihat work, but honestly, since I spent the time tweaking it up for the way I play, I am not bothered at all in the way it works now, no matter how many sets of samples are coming into play.

Only if I isolate the hihat where nothing else it playing do I notice any sample switching. If I'm playing along with other music, like when recording a track to a song, or playing one back, I really don't notice it at all.

It could be really useful to know at what specific CC value each openness range is set to switch for your setup.

[Glennbo]

Originally Posted by brainwreck

It could be really useful to know at what specific CC value each openness range is set to switch for your setup.

I don't know of any way to find out what set of samples are currently playing in Superior I just know that until I put the majority of the samples more to the closed or down position of the pedal, I was hearing real open samples jumping out when I had the pedal mostly down, and it made it sound real ragged. It was because it was jumping between fairly closed to pretty open samples when playing something like straight eighths, so you'd hear something like "tick, tick, shtick, PSHSSST, tick, tick".

After spending literally hours and days screwing with it, I finally arrived at a curve that feels and sounds pretty smooth when just playing the hihat like I would with a real hihat.

[brainwreck]

After having a quick play around with mixing two neighboring openness samples rendered from Ezdrummer 2, I think it is a promising approach toward filling in the space between sample sets. My intitial thoughts are that it will take some doing to find the right mix of volumes, and I think that the more open sample of neighboring samples might need a shortened sustain. I won't be able to begin better exploring this until the weekend, but if anyone else gives it a go, I would be curious to know what you think about it. I'm kicking myself for not thinking of doing that before sending off for a set of hi-hats and stand, but I still want to explore that route too. Any way, thanks for the discussion on the topic. It seems that that an idea from Todd has graduated to an idea from ashcat_lt, to idea from myself, which may evolve further by someone else.

[Tod]

Actually I do understand most everything we're talking about. The one thing I was not aware of is the travel of the pedal. But I do totally
understand what's all involved here, my main interest being that of a developer and primarily using Kontakt.

Originally Posted by ashcat_lt

I'm pretty sure this won't actually have acceptable results, but maybe if one was to map this CC to modulate the sustain/release envelope either instead of or on top of the sample switching...??? I'm not sure if EZD or AD can do that, but if so, or for folks with samplers that can, it might be worth a try.

That's exactly what I've been talking about ashcat, something like 6 to 8 openness samples with modulated ahdr envelopes in between the samples.
I actually did a test that I posted in the other thread using just one sample that was a sizzle sample. My only objective with this test was to
see how well the modulated ahdr might work, and I think it shows how feasible this might be. I think my intention with this test was a little
misunderstood, I was only demonstrating how the ahdr worked, but I was very much thinking ahead.

https://drive.google.com/file/d/1hbJ...ew?usp=sharing

The real snag I'm seeing here is the output of the pedal itself. Thanks Glennbo for the mp3 and the picture showing the values of your pedal.
I see their values are 2 to 3 increments apart, did the values go all the way to 127?

I was hoping to be able to approach this with the thought of using the whole 128 values, making very slight changes with each single value, which
would be the ideal. But I can see it's not going to be that simple. When SMM and I get started putting this together, I'll be able to tell a lot
more.

[Glennbo]

Originally Posted by Tod

Thanks Glennbo for the mp3 and the picture showing the values of your pedal.
I see their values are 2 to 3 increments apart, did the values go all the way to 127?

I was hoping to be able to approach this with the thought of using the whole 128 values, making very slight changes with each single value, which
would be the ideal. But I can see it's not going to be that simple. When SMM and I get started putting this together, I'll be able to tell a lot
more.

I'm using a function in my V-Drums brain called "Thin Data", which keeps the hihat controller from creating millions of events, which I'm pretty sure is why there are jumps between the values. I turned that function on when I first started recording the midi output of the V's because it was so dense and computers were slower then. I could probably disable it, but it's working great for me as it is, so I'm not going to fix it. The full range is spec'd at 0-127.

[Tod]

Originally Posted by Glennbo

I'm using a function in my V-Drums brain called "Thin Data", which keeps the hihat controller from creating millions of events, which I'm pretty sure is why there are jumps between the values. I turned that function on when I first started recording the midi output of the V's because it was so dense and computers were slower then. I could probably disable it, but it's working great for me as it is, so I'm not going to fix it. The full range is spec'd at 0-127.

Aah yeah, thanks Glennbo, I saw where you mentioned that before and I should have figured that.

[brainwreck]

Glennbo and DeathbyGuitar, I think I am liking e-hats more after spending more time with them. It's not an issue of understanding how anything works. It is more about coming to terms with e-hats not being acoustic hi-hats and getting used to playing them as such. In other words, getting used to pedaling between the handful of stepped ranges for sampled openness. I still think a sampled hi-hat with many more ranges would be a big improvement, but I can play it as is now.

Also, I contacted Roland support on the TD-11 CC4 range of 0-90. I didn't really need anyone to tell me this, but the support person confirmed that this is the normal output range for this drum module, telling me (of course) that their more expensive drum modules have a range of 0-127, which is lame of them if you ask me. I mean, intentionally limiting lower cost modules to CC4 range 0-90 with no good reason for doing so and no mention of it in their manual. Although what I have works just as well as any other controller when using my little jsfx CC4 expander plugin, I did have an idea on a hi-hat controller if someone runs into needing an inexpensive one that would likely work every bit as well as anything commercially made, with a full CC4 value range. You could connect up a low voltage dc signal (from any DC power supply) through a fixed resistor in series with whatever flavor of variable resistor that you like (forming a voltage divider) and grab the signal at the nodes of the variable resistor to send to a spare audio input on your interface. The resistor values would depend on the power supply output, but shooting for a variable voltage of around 1 volt or so should work with any audio input. Then a simple jsfx could convert the incoming sample values to MIDI CC4 values 0-127. This should be a simple ordeal for anyone familiar with jsfx MIDI scripting and it should be very low cost, depending mostly on the materials from which you make your mechanical pedal. It seems that this is essentially what these drum modules are doing any way (dividing a low voltage and converting to MIDI CC4 values). Optionally, if you don't want to completely diy, you could use any e-hat pedal as the variable resistor as outlined above. So you would only need a fixed resistor, a power source, and cabling.

On the acoustic hi-hat front, I had ordered a set, but then Guitar Center canceled my order 5 days later, which was a bummer. Apparently they sold the set to someone else, then canceled my order. I didn't realize it happened until I went to check on the shipping status of my order. So I will have to keep an eye out for another set (not from Guitar Center). And now I have a stand on the way with nothing to put on it.

[DeathByGuitar]

Originally Posted by brainwreck

It is more about coming to terms with e-hats not being acoustic hi-hats and getting used to playing them as such.

Pretty much. I'm all for obsessively researching the outer-limits of a technology for a given purpse (I do it all the time) but at the end of the day you gotta take the good with the bad. In my case, using sampled drums sucks because it's in-fact not the real thing and not me playing the parts, but on the flip side it would cost me like $4k or more to get the equivalent setup I'd need (drums,cymabals, mics, interface with enough inputs) not to mention a good sounding room that I could have access to, and at the end of the day nobody really cares besides guys like you and me.

Also, I wonder if the new TD-17 module has that same CC restriction. Have you checked out 2Box drum brains as well? I haven't tried them except for seconds at a time but maybe they might have more CC range. Maybe not.

With guys like you and me on the forum likely being the rule not the exception, I sometimes wonder what kind of crazy project we could all create for the community. One of these days...

[Tod]

Originally Posted by DeathByGuitar

With guys like you and me on the forum likely being the rule not the exception, I sometimes wonder what kind of crazy project we could all create for the community. One of these days...

I have no doubt we could create something that's pretty good. I'm more than sure that with
the right samples I could create a hi-hat in Kontakt that will be close. My partner and
friend SMM and I have discussed this and our plans are to give this a good shot as soon as
we can get to it, but that could be a while yet.

Between all us here, if indeed we can put something together, it's still going to be the
individual pedal CC4 output that needs to be addressed. With Kontakt's scripting I can do
a lot of things to try to accommodate, but there's only so far you can go and still retain
the full aspect of it all.

[ashcat_lt]

Originally Posted by brainwreck

You could connect up a low voltage dc signal (from any DC power supply)...

Probably not, unfortunately. Your interface input is going to be AC-coupled, which means DC-blocked (actually HPFed at a cutoff somewhere below 20Hz), which means it'll get weird fast.

You can use an AC signal, though. Something right in the middle of the spectrum like 1KHz so it's not likely to be affected by the frequency limits of the rest of the system. If you happen to have a spare output from your interface, that could be your source. JS ToneGenerator sent to the hardware output, through the voltage divider, and back to an input. Then...whatever needs to be done to get that to CC values. You could still use an existing Hat pedal or really any commercial volume or expression pedal.

[karbomusic]

I probably missed something but the one I prototyped and tested in the other thread did all this for about $20, would have been half that but the FSR was $15. I can GitHub the code if anyone wants it as it is incredibly simple.

https://www.youtube.com/watch?v=6jVPnQAzE_s

[Tod]

Originally Posted by karbomusic

I probably missed something but the one I prototyped and tested in the other thread did all this for about $20, would have been half that but the FSR was $15. I can GitHub the code if anyone wants it as it is incredibly simple.

https://www.youtube.com/watch?v=6jVPnQAzE_s

Hi Karbo, I don't know anything about what you're doing, at least on the technical side, is this
something that any e-kit drummer can attach to his pedal and make work?

[karbomusic]

Originally Posted by Tod

Hi Karbo, I don't know anything about what you're doing, at least on the technical side, is this
something that any e-kit drummer can attach to his pedal and make work?

Hi Tod, We covered it lightly with minerman in the other thread. That test at the time was just testing the FSR he had purchased for his E-Kit, so yes anyone can use that basic 15.00 part and attach it to a high-hat pedal and plug into their drum brain - probably some per-brain-you-have caveats but can't answer/fix those since I don't have a drum brain. I know people do it though.

Since I don't have a drum brain, I just built a small midi controller using a tiny Arduino Micro. So for that test, I coded up the MIDI stuff, connected the FSR and plugged the Arduino into the computer via USB, it then shows up in Reaper as a MIDI device which is how I was able to control AD Drums in the video.

[Tod]

Originally Posted by karbomusic

Hi Tod, We covered it lightly with minerman in the other thread. That test at the time was just testing the FSR he had purchased for his E-Kit, so yes anyone can use that basic 15.00 part and attach it to a high-hat pedal and plug into their drum brain - probably some per-brain-you-have caveats but can't answer/fix those since I don't have a drum brain. I know people do it though.

Since I don't have a drum brain, I just built a small midi controller using a tiny Arduino Micro. So for that test, I coded up the MIDI stuff, connected the FSR and plugged the Arduino into the computer via USB, it then shows up in Reaper as a MIDI device which is how I was able to control AD Drums in the video.

In the video it shows that you can get full use of the 128 values and it looks very
good. But then they have to plug that into their brain, so they are still at the
mercy of their drum brain, is that right?

I did follow you in the other thread as you progressed with this, but it's been a long
time since I had to deal with any electronics, so I mainly just observed.

[ashcat_lt]

If you've already got the CCs into Reaper, why the hell does it need to go back to the brain?!? Just send it to your VSTi and be done with it.

[brainwreck]

Originally Posted by DeathByGuitar

Also, I wonder if the new TD-17 module has that same CC restriction. Have you checked out 2Box drum brains as well? I haven't tried them except for seconds at a time but maybe they might have more CC range. Maybe not.

With guys like you and me on the forum likely being the rule not the exception, I sometimes wonder what kind of crazy project we could all create for the community. One of these days...

I have decided that I'm not going to get into the module hopping thing. I really only need good triggering for a drum sampler. And what I have does that fine within the limitations of e-drums. My only real complaint was the hi-hat CC4 thing, which I have taken care of. I'm sure I could get an inch better this and that for more money, but at the end of the day, it doesn't add up to much for me personally. Any more money in the direction of drums will go toward acoustic cymbals first, and a good kit later (for playing away from home), if I decide to keep at playing drums, which I think I will.

Yea, one of these days... Something that we could be doing now, is knocking up little pieces in jsfx, which I am doing as I can, toward putting together something bigger. Not having a solid background in math is a big hinderance there, but I'm working at that too, as time allows.

[brainwreck]

Originally Posted by ashcat_lt

Probably not, unfortunately. Your interface input is going to be AC-coupled, which means DC-blocked (actually HPFed at a cutoff somewhere below 20Hz), which means it'll get weird fast.

You can use an AC signal, though. Something right in the middle of the spectrum like 1KHz so it's not likely to be affected by the frequency limits of the rest of the system. If you happen to have a spare output from your interface, that could be your source. JS ToneGenerator sent to the hardware output, through the voltage divider, and back to an input. Then...whatever needs to be done to get that to CC values. You could still use an existing Hat pedal or really any commercial volume or expression pedal.

I didn't think about that. Dohh. I suppose a simple high frequency oscillator circuit would work for that purpose. Or do what Karbo is talking about, using a microcontroller. A free way of doing it would be to send a signal out, send it through the voltage divider, and send it back in. But that would increase latency of pedal control.

Edit: Quick scanning results in dumbassery. I just repeated what you already said.

[brainwreck]

Originally Posted by Tod

But then they have to plug that into their brain, so they are still at the
mercy of their drum brain, is that right?

No. The arduino is sending out MIDI messages over usb, which is connected to the computer for use by a drum sampler. If you wanted to use it with a hardware module (I don't know why you would), as long as the hardware module has a MIDI input, you could send the output of the Arduino to it.

[ashcat_lt]

Originally Posted by brainwreck

A free way of doing it would be to send a signal out, send it through the voltage divider, and send it back in.

What I said.

But that would increase latency of pedal control.

I mean, the output is just a steady signal so it's only the input latency that will impact it, which would have been true with the DC thing anyhow. Except really we won't want to just look at one sample, but rather some average over time, which will...complicate things a bit.

[brainwreck]

ashcat_lt, see my edited post above. :/

[brainwreck]

Originally Posted by ashcat_lt

I mean, the output is just a steady signal so it's only the input latency that will impact it, which would have been true with the DC thing anyhow. Except really we won't want to just look at one sample, but rather some average over time, which will...complicate things a bit.

I didn't think about that. It's a good point.

So then, we just need to read the peaks? We have a time predictable (sample accurate) signal to work with in the first place, so that should help in timing when to read peaks. Assuming that is true, couldn't we just take the difference between output and input? That is, for a linear resistance. If there is some other resistance curve, that would have to be accounted for.

[karbomusic]

Originally Posted by ashcat_lt

If you've already got the CCs into Reaper, why the hell does it need to go back to the brain?!? Just send it to your VSTi and be done with it.

Small distinction... One point is about using the FSR plugged directly into a brain. Many do it already AFAIK. The other point is my building of a MIDI device that uses an FSR to send MIDI CCs, bypassing a brain altogether. The latter is twofold - 1) yay, it's easy to build a simple controller/brain, 2) I didn't have a drum module when I wanted to test the FSR for minerman (other thread) so I had to build something to test it that could actually send the CCs.

In both ^cases the CCs aren't in Reaper yet. If they are already in reaper for BWs idea then that's the part I TLDR'd.

[Tod]

Originally Posted by brainwreck

No. The arduino is sending out MIDI messages over usb, which is connected to the computer for use by a drum sampler. If you wanted to use it with a hardware module (I don't know why you would), as long as the hardware module has a MIDI input, you could send the output of the Arduino to it.

Okay, thanks brainwreck, I looked arduino up so now I have a little better understanding. It looks
to me like Karbo has a good start on this, he's put forth a pretty good effort.

I still say though, that I believe a sampled hi-hat can be made to work nearly seamlessly, if a hi-hat
can put out CC4 in a smooth and seamless way.

[ashcat_lt]

Originally Posted by brainwreck

So then, we just need to read the peaks?
We have a time predictable (sample accurate) signal to work with in the first place, so that should help in timing when to read peaks.

Yes, and that's a good point I probably would have figured out on my own eventually.

Assuming that is true, couldn't we just take the difference between output and input?

I think maybe ratio is better. return divided by send. End up with a number between 0 and 1 and multiply that by 127.

Though there could be some loss (or gain!) in the system, so you'd probably need a constant "fudge factor" in there along the way. Not all interfaces (you might be surprised how few) will be at unity even with straight wire from output to input.

Originally Posted by karbomusic;

In both ^cases the CCs aren't in Reaper yet.

Well, sure, but you can't plug the USB output of your Arduino thing into any drum brain I know of, so you'd end up plugged into the computer along the way, in which case... I get it, though.

@Tod - CC4 is limited to 127 levels and MIDI timing resolution, which is more than most people need for most things, including this.

[karbomusic]

Originally Posted by ashcat_lt

... so you'd end up plugged into the computer along the way, in which case... I get it, though.

Right, because I needed to test the FSR somewhat end-to-end and had no way to get CCs into the computer while testing the FSR somewhat real-world.

If I were to do anything beyond testing with the Arduino, it would be the brain and replace the commercial drum brain (that I don't have) - it wouldn't be 'that' difficult to do since it's not going to do anything but get pads/notes/CCs into reaper. I have no plans to do it, I don't really need it but the occasional POC makes things easier later if I change my mind.