MM5837? Scott rider and PIC noise!

[Magnus Danielson]

From: Tim Ressel <Tim_R1 at verifone.com>
Subject: RE: MM5837? Scott rider and PIC noise!
Date: Tue, 5 Sep 2000 15:55:10 -0700 

> So like, what if you took a 7-stage noise gen, and fed it in as the clock for an
> 8-stage noise gen? I think this would have a dithering effect. Would it sound
> like a 56-stage? or a 15-stage?

It would certainly not sound like a 56-stager, and not like the 15-stager.

A 7-stager has the following critical statistic:

Shortest sequence of 0's: 1
Shortest sequence of 1's: 1
Longest sequence of 0's: 6
Longest sequence of 1's: 7
Shortest period: 2
Longest period: 13

What would happend is that the 7-stager would cause a varying delay of the
8-stager's transitions. While this migth sound mind-boggeling, it doesn't make
the noise THAT much more noisier. You will prolong the periods for the
8-stager's 255 phase positions. It is the number of cycles contained in the
7-stager's that will decide on the length of the full sequence. Also, such a
design will have less high-frequency content. I haven't made a FFT thought.

Then, someone proposes, let's XOR the output of the 8'stager with the output
of the 7-stager... well, it does not change the length, but it will cause more
transitions, it will double them infact.

You make much better use of those stages by forming a 15-stager. Now, slowly,
it migth dawn on you why they call them Maximum Length Sequence.

Cheers,
Magnus