Random tap tempo?

[Zijnzijn Zijnzijn!]

Wondering if anyone makes/is able to make a random tap tempo thing. Wanting to run it into my Canyon on s/h mode to randomize the glitches!

[rsfc]

Look a diy synth circuits for random pulses, gates circuits. Want to make something really vile? It would be easy to extract random pulses from a DIY Geiger counter.

[Teej212]

the tap tempo tremolo from eletric druid does a random waveform, s that what youre wanting?

maybe you could tap the voltage coming from the lfo generator, attenuate it to 3.3v (assuming thats what the canyon runs at?) and then use that in the exp in.

alternatively you could use an 8 step on random mode to get some pseudorandom

[oldangelmidnight]

It's a trigger, not expression.

[eatyourguitar]

there are a lot of ways to do it. or you could buy a eurorack module that would do it easy. it already exists in many forms both retail and DIY. the better question would be what is the statistical distribution of the length of time between pulses? is it normal (flat), gaussian (bell curve) or something else? what range of variation are you looking for? and are you looking for the variation to also have variation over some of the parameters over time? you can actually design a random system that meets %100 of your requirements if you understand your own requirements. you should also consider what parameters you would like to be hardwired and what exactly is user controllable. and also consider that the range of adjustment should not exceed your requirements.

[multi_s]

is it normal (flat), gaussian (bell curve) or something else?

the Normal distribution is the Gaussian distribution. i think you meant "uniform" for your flat reference.

[eatyourguitar]

is it normal (flat), gaussian (bell curve) or something else?

the Normal distribution is the Gaussian distribution. i think you meant "uniform" for your flat reference.

you are correct, I would like to say uniform distribution. uniform is also flat in lay terms. the book says all outcomes are equally likely to occur. obviously this implies all outcomes between the min time A and the max time B. such that

delay_time = uniform[A, B]

then when you switch the preset you use

delay_time = normal[A, Q]

where A is the average delay time and Q sets the maximum time variance in absolute value

but wait there's more!

you could program anything you want in software. there really is no limit to your creativity. I think in analog you will get uniform and normal and maybe log and lin. after that you need software. if you just want to make a random tap tempo accessory why not program a PIC for free?

[multi_s]

delay_time = normal[A, Q]

where A is the average delay time and Q sets the maximum time variance in absolute value

i hate to be a nitpicking ass hat (or maybe i don't, who knows) but just say Q is the variance or standard deviation, not "absolute maximum variance", this terminology has no meaning. The normal distribution is well defined by only 2 parameters, mean and variance (standard deviation is the square root of the variance, sometimes people substitute one for the other, they contain the same information).

if you find a function in a math package like Normal(a,q) etc it means draw a random variable from a normal distribution with mean a and variance q (or more likely standard deviation q, but depending on the implementation....). these are well defined quantities. for example in octave you could code x = normrnd(a,q) and it will pull a random draw from a normal distribution with mean a and standard deviation q.

since the normal distribution spans from -inf to +inf (regardless of teh value for standard deviation), there is no "maximum time variance" if you are pulling a variable from it, it could be any value, but it is just very unlikely with the normal distribution that it will be more than a few standard deviations away.

you can get octave for free and try it out and see. just keep drawing random normal variables then plot the data, as variance increases, spread from the mean is more likely, but variance does not put a maximum on how far that will be. you can try it yourself and see. don't take my word for it.

[eatyourguitar]

all very good points. I am not trained. I just read a lot of books. in this case I can see how I was giving maybe some bad information but you fixed it for me so now I know.

as with stocks the same with AC power. we measure everything in Root Mean Square because that is more relevant to everything that matters. I never considered that a bandpass filter would would have an infinite band but with extreme attenuation at the sides. I guess it is theoretically possible but in practice there would be guaranteed hard limits on both the shortest and longest delay times. you would see the bell curve of delay times.

[Orgeldream]

It's a trigger, not expression.

So the controlling input is for a momentary switch?

Maybe an arduino controlling a relay would be good (and cheap)?
You could have a variety of random flavours e.g. random subdivisions of the tap tempo speed. I have some code somewhere you could use as a starting point if you're interested in going this route.

[crochambeau]

Feed a comparator some white noise (I just described the Crustacean crackle output) and convert the pulse to whatever the supported circuitry needs in a tap tempo signal. Variable rheostats on the low and high points of the white noise volume control will allow you to optimally set the range on density (or, if you prefer "speed" of the random pulses).