MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Hemisphere for Ornament and Crime
MUFF WIGGLER Forum Index -> Eurorack Modules Goto page Previous  1, 2, 3 ... 34, 35, 36, 37, 38  Next [all]
Author Hemisphere for Ornament and Crime
Inju
First, thank you, thank you, thank you for making Hemisphere. It is amazing. I had a request for an option to Volts. I currently have a Minibrute 2s where you can sequence specific volts out. The Minibrute 2s has four patch points you can sequence out from and thus each patch point is valuable to control my Eurorack.

I wanted to use only one patch point of the sequencer to trigger two devices by using voltage changes. My Kick at 4 volts and the Snare at 2 volts. Thus allowing me to sequence both the Kick and Snare from one patch point. However currently this will only work if I have the sequence of Kick on, then Snare off, because the Volt acts as a gate vs. a trigger. In other words once the gate is in the on position, I have to turn it off again in order for it to act as a trigger again, while a trigger quickly turns the gate on and off as a pulse.

Even with this solved there is the problem of triggering both at the same time. I believe I figured this out by using a third voltage, 7 Volts that would trigger both. I can do this if I use mult out to the second Volt and have that in a mult to the triggers of both the snare and drum. I believe there may also be a way to use the Maths for this but have not figgured it out yet.
chysn
Inju wrote:
First, thank you, thank you, thank you for making Hemisphere. It is amazing. I had a request for an option to Volts. I currently have a Minibrute 2s where you can sequence specific volts out. The Minibrute 2s has four patch points you can sequence out from and thus each patch point is valuable to control my Eurorack.

I wanted to use only one patch point of the sequencer to trigger two devices by using voltage changes. My Kick at 4 volts and the Snare at 2 volts. Thus allowing me to sequence both the Kick and Snare from one patch point. However currently this will only work if I have the sequence of Kick on, then Snare off, because the Volt acts as a gate vs. a trigger. In other words once the gate is in the on position, I have to turn it off again in order for it to act as a trigger again, while a trigger quickly turns the gate on and off as a pulse.

Even with this solved there is the problem of triggering both at the same time. I believe I figured this out by using a third voltage, 7 Volts that would trigger both. I can do this if I use mult out to the second Volt and have that in a mult to the triggers of both the snare and drum. I believe there may also be a way to use the Maths for this but have not figgured it out yet.


I read this a few times, and tried to wrap my head around what you're doing. If I understand correctly, you're using a single channel of a sequencer to send CV output, and you want to interpret a 4V voltage as a kick and a 2V voltage as a snare and 7V as both kick and snare.

I'm a little skeptical of this approach. You'd probably find it way more reliable to use separate tracks for snare and kick.

But I don't think you ever got around to saying what your request for Voltage is, exactly.
chysn
Maximars wrote:
- The Tuner should emit even voltages, like 2.000 and 3.000 volts.


I'd say there are plenty of ways to generate output voltages. I don't think there's a good reason to do it from a tuner.

Quote:
- Schmittrigger should have a tuneable gate output like the Voltage app
(for instance, if a pitch-CV goes beyond a certain level, it emits -1V. In short, a bassline-cv-compressor hmmm..... )


Like other logic modules, you can put its output through attenuation and/or inversion to get specific voltages.
Maximars
At the moment I’m struggling with the Schmitt Trigger. Can’t set the threshold of the second unit. No way to reach the settings of the second one. Both work always with the same setting, which makes no sense. Is it possibly a bug?

In my opinion, it would be a great help if the Tune applet emits Voltage and the outputs are not in use at the moment. The References App of o_C did emit even voltages for a good reason.

Whenever I tune my oscillators, I want to have a even voltage at their v/octave input, something like 2.000 volts. Most of the time patches are crowded and it is not so simple to allocate the oscillators straight voltages. So i would pull the 1v/oct cable, patch it into the CV output of the Tuner, tune it to C for instance, patch it back and you’re ready to go. Left jack 2V, right jack 3V, fixed, little thing, big improvement! One applet for all tuning purposes thumbs up
Inju
Hi Chysn,

I apologize for the confusion. I'm currently trying to figure out a way to easily sequence my drums using the Minibrute 2S. The Minibrute 2S has an amazing sequencer where you can simply depress 16 pads to send out one of several variables assignable to two outputs. These variables include;

• 1V, 2V, 5V, and 8V. These are referred to collectively as the xV type. (Each one of these can easily be changed up the voltage parameter you choose by turning the corresponding encoders which are above each of the 16 corresponding pads. In other words I can choose the pads to go up to 8 volts, but simply turning a nob above the pads, I can easily change their voltages from 1 to 8V. The increments is the voltage you select divided by 100)
• Env (for Envelope)
• Sin(e), Tri(angle), Saw Up, Saw Down, Sqr (Square), and SnH (Sample & Hold). These are referred to collectively as the LFO type.
• Pitch, Gate, Velocity and Pressure is also assignable to each pad and encoder.

The amazing thing about the Minibrute 2S is that for each of the 16 steps, I can assign a constant voltage to come out. Thus I can assign one step to 1.5 volts and another step to 7.5 volts, or any voltage that I want.

I am trying to figure out a way to trigger the Kick and Snare modules using only ONE of the TWO outputs. Currently I would need to assign each output channel, one to a Kick and the other to the Snare.

By allowing a trigger vs. a gate on the Volt in hemisphere, every time the step sends out a certain value, the Volt sends a gate signal to the trigger of my kick drum.

By allowing a trigger versus a gate, this has some amazing applications besides the use I wish to use it for. For instance, by using a mult to a KB out that sends out Voltz/Octave signals to your oscillator and the other to the Volts, I can program that every time a C2 note is pressed a kick drum would be triggered and if a C4 is pressed a snare drum would be triggered. However, since the volts being sent out by the keyboard or in my case sequencer/pads is a constant signal, the gate never closes unless it receives a different signal that is of a different voltage. In other words, every time a voltage is registered on the Volt, instead of it being a gate, allow it to simply be a trigger. Add a third variable where if the signal is above a certain amount, both triggers fire.

Hope that makes sense. Also, though no-one really talks about it. I love the scope widget you have in Hemisphere. I use it all the time and compare it to my scope on my laptop. The spinning of the encoder to get the resolution just right takes time for each wave, but still it is amazing! Once again thank you. If you have a Patreon or Paypal account, please pm it to me and I'll send you some funds. smile
chysn
Maximars wrote:
At the moment I’m struggling with the Schmitt Trigger. Can’t set the threshold of the second unit. No way to reach the settings of the second one. Both work always with the same setting, which makes no sense. Is it possibly a bug?


It's as designed. Both channels use the same threshold. See the documentation here: https://github.com/Chysn/O_C-HemisphereSuite/wiki/Schmitt-Trigger

Quote:
Whenever I tune my oscillators, I want to have a even voltage at their v/octave input, something like 2.000 volts. Most of the time patches are crowded and it is not so simple to allocate the oscillators straight voltages. So i would pull the 1v/oct cable, patch it into the CV output of the Tuner, tune it to C for instance, patch it back and you’re ready to go. Left jack 2V, right jack 3V, fixed, little thing, big improvement! One applet for all tuning purposes thumbs up


Yeah, I can understand that. It would be easy enough to do in a future release.
Mdashdotdashn
@chrsn: I'm curious if you found anyway to measure the performance/intensity of calculation of any app. I've currently have a terrain synthesis oscillator that's really nice but it is also close to brining the teensy to its knees.

One approach would be to bang a digital output with 0 at the start of the audio interrupt and 1 at the end and pass it through a slew limiter. But before messing up with hardware, I was wondering if there was any built-in solution you coul see.
chysn
Mdashdotdashn wrote:
@chrsn: I'm curious if you found anyway to measure the performance/intensity of calculation of any app.


While your app is running, return to the main menu by holding down the right encoder. While in the main menu, your app will continue running. From the main menu, press the left encoder button. This will bring you to a sort of diagnostic screen. The CORE line will tell you the processor's usage percentage.
Mdashdotdashn
chysn wrote:

While your app is running, return to the main menu by holding down the right encoder. While in the main menu, your app will continue running. From the main menu, press the left encoder button. This will bring you to a sort of diagnostic screen. The CORE line will tell you the processor's usage percentage.


Fantastic ! Thanks. Urg. 80%
coolshirtdotjpg
chysn wrote:
coolshirtdotjpg wrote:
One feature that would be great, and I wonder how hard it would be to implement is: saving tempo in the metronome app, or even just another parameter with common tempos.


It's a good idea. For Hemisphere 2, I plan to make the metronome more central to the system, by assigning it to a long-hold of the Down button. That will replace the category filtering system, which I don't think is all that useful. So some form of preset system would fit in pretty well.


That would be amazing. Thank you, thank you, thank you!
coolshirtdotjpg
chysn wrote:
Inju wrote:
First, thank you, thank you, thank you for making Hemisphere. It is amazing. I had a request for an option to Volts. I currently have a Minibrute 2s where you can sequence specific volts out. The Minibrute 2s has four patch points you can sequence out from and thus each patch point is valuable to control my Eurorack.

I wanted to use only one patch point of the sequencer to trigger two devices by using voltage changes. My Kick at 4 volts and the Snare at 2 volts. Thus allowing me to sequence both the Kick and Snare from one patch point. However currently this will only work if I have the sequence of Kick on, then Snare off, because the Volt acts as a gate vs. a trigger. In other words once the gate is in the on position, I have to turn it off again in order for it to act as a trigger again, while a trigger quickly turns the gate on and off as a pulse.

Even with this solved there is the problem of triggering both at the same time. I believe I figured this out by using a third voltage, 7 Volts that would trigger both. I can do this if I use mult out to the second Volt and have that in a mult to the triggers of both the snare and drum. I believe there may also be a way to use the Maths for this but have not figgured it out yet.


I read this a few times, and tried to wrap my head around what you're doing. If I understand correctly, you're using a single channel of a sequencer to send CV output, and you want to interpret a 4V voltage as a kick and a 2V voltage as a snare and 7V as both kick and snare.

I'm a little skeptical of this approach. You'd probably find it way more reliable to use separate tracks for snare and kick.

But I don't think you ever got around to saying what your request for Voltage is, exactly.


This is a feature of the stillson hammer mk1, and it's pretty cool, basic just mapping a threshold to trigger a gate output. Unfortunately it doesn't seem particularly well suited to hemispheres since it really only gets interesting when you have at least four outputs. This would an interesting idea for the original ornament and crime where X volts cause a gate at X output, or alternatively Y volts opens up inputs X through Y. This doesn't seem particularly hard to implement, but just doesn't seem that useful with only the two gate outputs for each applet.

[edit, I say it doesn't seem that hard to implement with precisely zero knowledge of programming for O+C so of course, forgive me if I'm being ignorant.]
mxmxmx
Mdashdotdashn wrote:
@chrsn: I'm curious if you found anyway to measure the performance/intensity of calculation of any app. I've currently have a terrain synthesis oscillator that's really nice but it is also close to brining the teensy to its knees.


fwiw, it's probably obvious but you generally don't want to use floats on this thing; MK20 doesn't have an FPU. extern/dspinst.h includes a bunch of things you might find useful ... (not sure which oscillator we're talking about, of course, but superficial sleuthing turns up this).
coolshirtdotjpg
Another thought, the O+C hardware seems perfectly suited to some sort of quadrophonic panning control, but again hemispheres might not be the right way to implement it.

I'm imagining this in combination with a quad VCA setup so that each of the 4 outputs of O+C are controlling each channel of the VCA to allow for quadrophonic panning/mixing.

Chysn, is there a way that the input of one app could be "normalled" to another, if so, you could have each side of a quadrophonic app sending voltage to the appropriate outputs with the right side receiving X and Y information from the left side?

If so that would be super interesting.

Also, I hope that my floating ideas is not seen as "gimme gimme gimme" feature requests. Since you seem to be really open to ideas from the community, I am just throwing out ideas and seeing what sticks, it is interesting to see what is feasible, what features seem useful to people other than myself etc. Like most folks my ideas far outstrip my ability to implement any of them in code.
chysn
coolshirtdotjpg wrote:
Chysn, is there a way that the input of one app could be "normalled" to another, if so, you could have each side of a quadrophonic app sending voltage to the appropriate outputs with the right side receiving X and Y information from the left side?


Possible, sure. That's vaguely how master clock forwarding works. The first digital input gets the real data, and the third digital input becomes a sort of "virtual input."

Realistically, there are interface limitations that make it hard to do that kind of thing elegantly and frustration-free, so I've kept fancy things to a minimum.

Quote:
Also, I hope that my floating ideas is not seen as "gimme gimme gimme" feature requests.


Not at all, I understand. At this point, memory being what it is, I sort of have the luxury of having little latitude to do anything else with the firmware. smile
coolshirtdotjpg
chysn wrote:
coolshirtdotjpg wrote:
Chysn, is there a way that the input of one app could be "normalled" to another, if so, you could have each side of a quadrophonic app sending voltage to the appropriate outputs with the right side receiving X and Y information from the left side?


Possible, sure. That's vaguely how master clock forwarding works. The first digital input gets the real data, and the third digital input becomes a sort of "virtual input."

Realistically, there are interface limitations that make it hard to do that kind of thing elegantly and frustration-free, so I've kept fancy things to a minimum.

Quote:
Also, I hope that my floating ideas is not seen as "gimme gimme gimme" feature requests.


Not at all, I understand. At this point, memory being what it is, I sort of have the luxury of having little latitude to do anything else with the firmware. smile


Haha, that sounds about right! Just out of curiousity, is it possible for the user to compile the firmware with only certain applets, so that if I ever do what I keep threatening to do and come up with an applet for it, could I swap out something I wouldn't use out to save on memory?
chysn
coolshirtdotjpg wrote:
Just out of curiousity, is it possible for the user to compile the firmware with only certain applets, so that if I ever do what I keep threatening to do and come up with an applet for it, could I swap out something I wouldn't use out to save on memory?


Yeah, of course. See the hemisphere_config.h file. Remove, or comment out, the unwanted applets from this file, and then remove the associated HEM_applet.ino file from the project.
Mdashdotdashn
chysn wrote:
coolshirtdotjpg wrote:
Just out of curiousity, is it possible for the user to compile the firmware with only certain applets, so that if I ever do what I keep threatening to do and come up with an applet for it, could I swap out something I wouldn't use out to save on memory?


Yeah, of course. See the hemisphere_config.h file. Remove, or comment out, the unwanted applets from this file, and then remove the associated HEM_applet.ino file from the project.


Also you can remove all applications but hemisphere and the configuration apps.
Mdashdotdashn
mxmxmx wrote:
Mdashdotdashn wrote:
@chrsn: I'm curious if you found anyway to measure the performance/intensity of calculation of any app. I've currently have a terrain synthesis oscillator that's really nice but it is also close to brining the teensy to its knees.


fwiw, it's probably obvious but you generally don't want to use floats on this thing; MK20 doesn't have an FPU. extern/dspinst.h includes a bunch of things you might find useful ... (not sure which oscillator we're talking about, of course, but superficial sleuthing turns up this).


Your search lead to the right area although it wasn't really in par with my current state. The oscillator is here and involves currently multiple evaluations of perlin noise. I'm totally aware of the lack of FPU, I've got used to fixed point back in the day the raspberry pi was in the same situation and it's used heavily in all my processing (hence the templated code). There's a fairly comprehensive class for it here

Edit: It seems a "blank" app selected in both side uses still 25% of the core. Is that the amount of time it takes to peek/poke data to the outside ?
mxmxmx
Mdashdotdashn wrote:


Your search lead to the right area although it wasn't really in par with my current state.


alright then.

Mdashdotdashn wrote:

Edit: It seems a "blank" app selected in both side uses still 25% of the core. Is that the amount of time it takes to peek/poke data to the outside ?


possible, idk. there's the DAC (~4 us, IIRC), ADC (dunno), display (mostly DMA, but not all of it), polling of inputs etc. o_C v1.35 uses DMA for the ADC now, too, i don't think hemispheres does; so that might help bring that number down a little, but probably not much. either way, (sorry i keep rehearsing this...) OC was designed to be *quantizer*, there's only one SPI bus on MK20, and so on; it'll be of limited use for DSP, even within the limits of what teensy 3.2 can/could theoretically do.

i have a teensy 4 for beta testing on my desk now, that might open up some options (if i/we/someone actually port the code) but some of the bottlenecks (1 bus for DAC + OLED) are set in stone, so there's that.
Mdashdotdashn
[quote="mxmxmx"]
Mdashdotdashn wrote:

either way, (sorry i keep rehearsing this...) OC was designed to be *quantizer*, there's only one SPI bus on MK20, and so on; it'll be of limited use for DSP, even within the limits of what teensy 3.2 can/could theoretically do


I'm complaining at all, the O_C is a wonderful platform and there's *tons* of applications that uses low dsp power that can be used. This was more of an experiment to see how far I could take it. But I've got a bag full of other things to try.
pld
Mdashdotdashn wrote:
Edit: It seems a "blank" app selected in both side uses still 25% of the core. Is that the amount of time it takes to peek/poke data to the outside ?

The time measured is the total execution time (in cycles) of the core interrupt handler, so that includes all the DAC, ADC filtering, framebuffer switcheroo, setup next DMA transfer to screen, etc. I only have an ancient 'scope shot that shows ~20us of pre-processing but that's probably gone down a bit.
All the app computations have to happen in the window where the screen DMA transfer takes place so there's little wiggle room since the whole ISR has to complete in < 60us, and it can't use 100% either since the screen redraws and UI happen in lower priority interrupts/asynchronously.
Ebotronix
hi chysn!
Trending dual slope detector app
ChgValue: The assigned output emits a trigger when the signal changes its value by more than 1/4 semitone (or about .02V)

is it possible to reduce the 0,2 V to get a trigger from the microtonal scales?
Timmy
Mdashdotdashn wrote:
I've currently have a terrain synthesis oscillator that's really nice but it is also close to brining the teensy to its knees.


Yes, Perlin noise is an intriguing possibility. Here is some discussion of implementing it on the MI Peaks and on O&C from about 4 years ago (neither implementations ever happened): https://forum.mutable-instruments.net/t/dead-mans-catch-modified-firmw are-for-peaks/7248/203 which was inspired by this blog post: https://dmachinery.wordpress.com/2014/08/15/perlin-pattern-generator/

Of course, that was as a modulation source operating in the time domain. I suspect it might be very useful as a source for quantisers. Not sure that perlin noise offers anything special as an audio noise source in the frequency domain, though, and in any case, with the 16.6kHz DAC sample rate on the O&C, it is hardly worth bothering unless you like Merzbow (although I am told by Merzbow connoisseurs that the Viznutcracker, sweet! app in O&C sounds nothing like Merzbow's best...oh well).

Doesn't Sonic Potions produce a module or modules that use Perlin noise?
chysn
Ebotronix wrote:
hi chysn!
Trending dual slope detector app
ChgValue: The assigned output emits a trigger when the signal changes its value by more than 1/4 semitone (or about .02V)

is it possible to reduce the 0,2 V to get a trigger from the microtonal scales?


Eh... You can change the value and recompile if you want, but that value is based on quite a bit of experimentation. I don't "smooth" incoming CV like the native firmware does (did?), so if you go much lower, the tradeoff is false triggers because of the ADC noise, or whatever.
coolshirtdotjpg
chysn wrote:
coolshirtdotjpg wrote:
Just out of curiousity, is it possible for the user to compile the firmware with only certain applets, so that if I ever do what I keep threatening to do and come up with an applet for it, could I swap out something I wouldn't use out to save on memory?


Yeah, of course. See the hemisphere_config.h file. Remove, or comment out, the unwanted applets from this file, and then remove the associated HEM_applet.ino file from the project.


That's great. Forgive my ignorance on that. Very exciting work that you've done here.
MUFF WIGGLER Forum Index -> Eurorack Modules Goto page Previous  1, 2, 3 ... 34, 35, 36, 37, 38  Next [all]
Page 35 of 38
Powered by phpBB © phpBB Group