A mass spring system and a filter are the same thing, actually.
1 Like
So please tell me more! How does the mass spring go about filtering the audio?
The same way that a spring reverb would.
You use a transducer to turn an electrical signal into a vibration in one end of the spring. And then another (or possibly the same) transducer turns the vibrations of the spring back into signal
For a spring/needle, you would apply impetus to the needle and measure it’s position.
There are direct analogues between inductors/capacitors/resistors and inertial mass/spring/friction and also acoustic spaces (tubes/boxes/turbulence)
2 Likes
Indeed ! And yet none of the filters I tried reacted to a CV changing direction in a way that would come anything close to the elasticity I’d expect from a mass / spring system, I got some interesting results but nothing close to were I was trying to go to… So I am guessing but, maybe fine tuning the parameters of the system may be crucial ?
On a side note I think modules oriented to patch programming, like lower level building blocks, could be fun too. If I want to patch something that is just giving me the sin(x) I can use a phasor based oscillator, but it will surely have a lot more parameters and a GUI completely polluting the patch… I for one would love lower level building blocks !
Or maybe they exist and I haven’t been able to find them, it is really possible. 
1 Like
Good call, I will try that !! Thank you.
The fun part of Inertia is how it’s simultaneously covering both function generator and filter (and oscillator) territory. Most function generators don’t have resonance, and most filters don’t give you separate cutoff frequencies for rising and falling. Also there’s a clever switch that lets the first two knobs swap between rise/fall time, and rate/skew, so it makes some amount of intuitive sense regardless of how you’re using it.
Also you can get some interesting results (including highpass filtering) by mixing the 1st and 2nd order outputs.
1 Like
Exactly !! (hysteresis on a cutoff, whaaaat ? )
I think this modules offers something unique, and that is no small feat in the middle of the 11209 modules that are right now present on modulardgrid (not that this is the start and end of everything you know, just to take a rather absurd step back).
I keep looking, I have found funky looking ways of shaping CV with mixes of filters and mixers, but still nothing elastic, alive. I still have to try @StochasticTelegraph 's proposition with code though, and it looks like a really interesting possibility for sure.
Inertia is related to acceleration
unit for acceleration is [metre per second squared]
Meter per second,per second
How is the progression of speed…in the time
In the algorythm,you have to put a threshold to stop the movement when the speed is near zero…Floats are not perfect on computer
Are you a real person ?
3 Likes
Probably not what you’re looking for, but I was reminded of the bouncing balls idea.
I have done an algorythm in Blitz3d in the past…something simple (not more than 75 lines I think) that could add inertia effect via analog stick on a shoot them up…
I have lost the code
But I have some remembers
Thanks for the suggestion !
I have spent quite some time with the CF100 in that kind of configuration and you are right that it kind of works for certain kinds of CV signal. It does not work with less abrupt slope changes though, I could not get that kind of reaction to a classic function generator à la Rampage with slow rise and fall :
I wonder if @AlliewayAudio Cartoon Running might get you in that ballpark - on some settings. It’s very chaotic and unruly, but with a CV input some of its behaviors might do what you want.
1 Like
That is a great suggestion, I thought about it and went to try other things… Off to try that !
1 Like
While this doesn’t do it, if you roll your own solution my old “LFN” could be part of the solution. As the manual says, it’s a source of randomness, actually noise through a graphic equalizer with the bands super low (like 1/10 of a hz, etc…) VCV Library - Squinky Labs LFN.
btw, the various slew limiters are also filters, one pole LP filters with saturation on the inputs and feedback. back in the day that kind of non-linearity was called “TIM” when it occurred an an audio amplifier.
BlitzPhys.txt (1.1 KB)
Blitz3D code…Ten years ago
If @StochasticTelegraph can translate it…Basic 2D inertia
I’d guess that if you turn up the time scale on your scope there, you’ll see that the filtered signal has some small fluctuations as well as a lag. The frequency of the ramp (~0.75Hz by the look of it) is much lower than the cutoff frequency of the filter (8Hz, if it’s turned all the way down), which means that the filtered signal can bounce back pretty quickly with minimal overshoot. To get more overshoot, you could either lower the cutoff frequency (by sending a negative CV to the FM input) or turn up the resonance.
It is actually possible to get an asymmetric rise and fall using the filter technique too. You just need a slope detector feeding from the output of the filter back to its FM input. Here’s an example of that with both Zzzorb (blue trace) and CF100 (yellow trace).
asymmetric_cv_inertia_using_filters.vcv (8.0 KB)
I couldn’t say how closely any of this mimics the behavior of the Inertia module, but it’s a fun experiment.
In the patch, I used a sample delay and comparator as a slope detector, since the existing slope detectors that I tried would interpret very slow changes as neither rising nor falling, but rather “steady.” This meant that I also needed to set the CF100’s Noise Level (right click menu) to Off, because tiny variations from noise would be interpreted as slope changes. As an added bonus, this brought the CPU usage of the CF100 down to about the same level as the Zzzorb, so that’s nice.
2 Likes
Ok,I see
Amplitude control the rise,fall,filter freq,res,pitch,FM…etc
Yes I did see that, I had not thought of the negative CV to lower the cutoff, and indeed I felt that it was needed, thanks for this idea !! This and the “Cartoon Running” module both are good options, and both work in some situations, and are more or less interesting depending on the CV they chew on (some settings work only on stepped CV, others only on slowly rising and falling function generators).
I mean, it is always interesting, but not often looking or sounding like a natural response to a simple movement (for instance I’d like to see the trace follow more closely the input when the signal is “steadily” falling or rising, and showing signs of oscillation when the input signal is changing direction, then more or less quickly stabilizing to a “closely following” state again…).
It really is interesting to try all those less usual waveshaping tools on CV…
1 Like