DSP Robotics Support

Not nearly as Penny as I - haven't a clue what "2^24" is.

k brown wrote:Not nearly as Penny as I - haven't a clue what "2^24" is.

2^24 is "two to the twenty-fourth power".

(The forum software doesn't let us input superscripts...if we didn't have that '^' shorthand we'd have to type in "2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2 x 2". )

The math I learned in art school didn't cover that.

Wow, that discussion really moved on while I wasn’t looking!

I give much thanks to tulamide and deraudrl for enhancing my education about number representations in the float system. And the link tulamide showed really surprised me. 0.01 isn’t in fact accurately represented.

Those who know me realise I’m rather pragmatic, so if something works as I want I’m normally happy. I was looking for a number whose reciprocal didn’t go on forever, so 1/200 did the trick, and I got the same green float number out of the preset as I put in, so job done.

But it’s interesting to learn about what goes on under the hood. The representation of 0.01 I would say is “close enough” and was therefore fit for purpose on this occasion, but maybe not in other situations.
But it was insufficiently elegant and the principle explained here can be used elsewhere.
I personally like the idea of sticking to powers of 2 for such matters; it’s as easy to remember as using 1/200 etc.

Cheers

Spogg

Thanks for that affirmation - we'll see if Manfred can get it to act up!

I do recommend that you polish up the math skills, you kind of need it for synth programming

adamszabo wrote:I do recommend that you polish up the math skills, you kind of need it for synth programming

In my case there’s an appropriate English joke: You can’t polish a turd.

Cheers

Spogg

Hehe, yeah I understand I remember I HATED math in university, we learned shit I knew I would never use in life, until I started messing with synthesizers and programming, those skills proved useful and were a great help. Not much the math, but the method of thinking one needs to use.

I'm a science guy. I think, that knowing is key to progressing. Just as in this thread. If I know what a float datatype really is, how it works, it will lead to a better usage of it. In general, I had similar experiences to Adam. My majors in school (is that the correct word? it sounds strange. i mean the specialised intensive courses you need to select in higher schools) were math and german, but only because I couldn't select the latin/german combo. I was really bad in maths, because I saw no real world application for all the strange graphs and formulas.

Then I started programming. And suddenly, everything that seemed so difficult in school, became simpler just because I could see or hear the result of the math. It made sense.

And once you understand the concept of binary math, a whole new world of possibilities opens up. Back when I programmed my tracker in the early 80s, it was binary math that allowed me to develop the sequencer, the sound mxing and all that stuff. I had no idea, how an analog oscillator worked, but I knew how to put loads of information in just the low nibble of a byte.

Spogg wrote:I give much thanks to tulamide and deraudrl for enhancing my education about number representations in the float system. And the link tulamide showed really surprised me. 0.01 isn’t in fact accurately represented.

A completely off-topic aside: some years ago (on Usenet, to give you an idea how long), I was involved in a rather spirited discussion concerning the metric system as it applies to woodworking. It was the usual bit about how much easier it was to deal with metric divisions by 10 versus the US custom of quarters, eights, sixteenths, whatever. (Not to mention such oddities as why sheets of Baltic birch plywood are 60" square and not the 1.5m you'd expect given the countries of origin. Still not sure about that one.)

The discussion degraded rapidly when the, um, gentleman suggesting we all throw away many $$$$ worth of router bits, drills, etc made the rather churlish comment that "You Murricans should just get with the program because metric is more accurate anyway: that's how computers work." He was seriously dumbfounded to learn (or rather forcibly taught) about the errors inherent in doing decimal arithmetic in binary computers.

More on-topic, I for one am glad that musical scales weren't invented by that level of metric-centric fanatic, although that might well explain why "death-metal" is a thing.