07:28 diagonal1x: Maybe i believe others who seem to single voiced i.e unanimously suggest that in mathematics they name that discrete Fourier transform where there is fast version of that.
07:29 diagonal1x: The last which was presented by Gauss in the early 19th century
07:38 diagonal1x: It's used in everywhere that involves signal processing but more often in hardware analogue styles, it seems I had tried to bring in one more SW based layer of that.
07:53 diagonal1x: But at Christmas time when I have finished my other work, I look more deep how they present such in mathematics. But in my current vectors it's just matters how many times the common value is added and multiplied indeed.
08:13 diagonal1x: It's still that de ja vu thing, it's very special how it tracks the order of elements, I do not fully understand how to explain this some black box hw magic even maybe.
09:08 diagonal1x: Very weird, so we store 511 and 2, their sum is 513, but we add 512 twice, so 1537-511-2 then 1024-511-2 is 511, as well as 511+
09:08 diagonal1x: 514 is 1023
09:09 diagonal1x: Err 1025
09:12 diagonal1x: So if you add three times 512 it gets you 1023
09:33 diagonal1x: And subtract one more time 511-2 it gets 510 which is 2 from 512
09:34 diagonal1x: Maybe that is some FFT
11:36 diagonal1x: Seems to matter to access even or odd numbers, or you'd be off by one, it's possible to only do even numbers.
15:43 diagonal1x: Yeah discard that, I saw delusions, on mobile phone I tested what I wrote last time again, so there's no difference if it's odd or even, befolllllllllllllllllllllllllll
15:44 diagonal1x: I just looked at the logs, but whatever, I am so often wrong
15:45 diagonal1x: The trick seems fairly simple
15:46 diagonal1x: Tried at work with 511 and 510 , two element vector
17:58 diagonal1x: Tried with 511 and 2 as well as 524 and 511 two element vectors works all the way, can select the element in logloglog