[sdiy] digital delay, was ...

[Glen]

At 12:33 AM 9/12/03 , harrybissell wrote:
>This idea has some of the merits of using the old oberheim
>case to plant geraniums :^P

Of course, I was the one that suggested using it to sort germaniums.


>(otoh I don't think the sections of the same BBD would track well
>enough to get any noise reduction in this manner...)

I realize that it wouldn't be a perfect match, and the potential noise
reduction would suffer *at least* some. However, I have experimented with
analyzing an "unused" channel of an otherwise "stereo" digital recording to
remove the residual hiss caused by my soundcard's electronics, from the
channel that actually carried my desired signal. 

It actually did work fairly well, at least in that case. It certainly beat
trying to use tone controls, or using a dynamically adjusted low-pass
filter in a single-ended noise reduction scheme ala National
Semiconductor's approach. It was also better than the pumping of a
compander approach.

I realize that a BBD is a much harsher environment to try this with, but I
wouldn't be surprised if it actually worked at least fairly well. Maybe
I'll eventually get around to trying it on a breadboard, and use my
computer and my (new and improved) soundcard for the processing power.


>One twilight zone idea of mine would be to separate the audio into
>several different bands (low mid high) and set the gains so that each
>BBD wouls see maximum signal (the highs presumably would be
>lower amplitude than the lows)... then when summing them back together
>you would account for the different gains to get a delayed signal with
>higher signal amplitude (better S/N)
>
>twilight zone enough ???   :^P

That's not so "twilight zone" in my opinion. Using different bands would
probably also help with intermodulation issues--if you had planned to add
companders to each frequency band. 

How many bands were you thinking of using?

later,
Glen