Influence your own random number generator

[davidhorman]

Years ago I modified a sound capturing program written in PowerBASIC with the intention of giving myself a usable random number generator. It works by taking a sample of sound from whatever recording source you've selected in Windows (it should be something analogue, like Line or Mic In, and if your card is "noisy" enough it doesn't need anything connected to it) and considering the parity of the total of the bytes in that sample. Or something like that - it was some time ago and I've lost the source code.

So, follow the link, select an analogue source for recording in Windows, run the program, and press Cap. The waveform will be displayed. The white line shows the bias toward parity 0 or 1. By chance, the line should tend to the center. Use your Psi powers and nudge it left and right! :D

http://www.geocities.com/david_horman/psi.zip

David

[Darat]

Can't get to the link.

[Matabiri]

Or influence the cosmic rays at this random number generator:
http://www.random.org/

[ktesibios]

One trouble I see with the idea in the OP is that the noise floor of some computer sound cards is anything but random. Sticking the analog circuitry in the same box as a lot of fairly high level pulse trains is a recipe for getting assorted tweets and whistles. There's a good reason that most professional DAW systems put the analog front and back ends and the A/D & D/A conversion in a separate chassis.

Rather than use the noise of the sound card it would be much better to use an analog noise generator as the signal source.

Of far more practical interest would be a method of influencing a mechanical random number generator- specifically the one that uses the little white cubes with the black spots on the faces. ;)

[Abdul Alhazred]

I think I've got it as far as mathematics and computer science are concerned (I have an MS in it), but very likely I need a lesson in physics, though I have some grounding in that field, too.

Yes I know and understand that in a mathematical sense no deterministic process can be truly and unpredictably random.

But for practical purposes, what is wrong with the old GWBASIC of 20 years ago:

10 RANDOMIZE TIMER

or any modern equivalent, to plant the seed?

I really do know what that construct means under the covers, I am not taking it on faith. I am not being argumentative about it. I am really just asking for information.

Yes it's not truly random, but why isn't it random "enough" whatever that means?

Don't ask me what "enough" is. Please tell me your opinion and why.

If you understand what I just said, I want your opinion.

Thank you.

[davidhorman]

Can't get to the link.

Oops. Copy and paste it into a new window.

One trouble I see with the idea in the OP is that the noise floor of some computer sound cards is anything but random.

I'd contend that the overall parity of each sample is sufficiently random. But you could feed in any source of sound rather than rely on just the noise in the circuit - a mic, or in my case I use the input from my TV card (which comes line out from the TV card and goes back in to my soundcard).

But for practical purposes, what is wrong with the old GWBASIC of 20 years ago:

10 RANDOMIZE TIMER

or any modern equivalent, to plant the seed?

Yes it's not truly random, but why isn't it random "enough" whatever that means?

By randomising it with the timer, you're just picking one out of a finite number of pseudo-random number sequences. Okay, you're picking that sequence at random, but the sequence itself is deterministic. That doesn't rule it out for psi research though - you could use your powers to run the program at just the right time to come up with just the right sequence of numbers to show a significant result (if such a sequence exists).

David

[WinterMute]

or any modern equivalent, to plant the seed?

If you haven't seen it this is an interesting open source project for generating random numbers. SGI (Silicon Graphics) had a similar project some years back but they used a lava lamp to generate random numbers. But I think SIG's project relied on a pseudo random number generator.

Lavarnd:
http://www.lavarnd.org/what/index.html