Re: Re: 1 last question about filters

[Guest guest]

Pam,

1. You can set the delay lower than 250 milliseconds if you wish, but don't go

above that.

2. The problem with FIR filters, which are time-based and very accurate across

the range in terms of delay, is that they are not very accurate in the frequency

domain. If you look at the filter graph, you'll see that the very central

frequency of the filter is very accurate, but the accuracy rolls off very

quickly, so the edges are not very accurate. I still prefer Elliptic filters

for inhibits and Butterworths for reward except for coherence, phase and

synchrony, which are timing-based relationships.

3. not that I can think of.

4. Yes, I have another batch that will be released before I leave for Europe,

Australia and Korea later this month. I'll mention it on the list, and those

who have joined the Yahoo group for those who have purchased the protocols

subscription will get a notification from that as well.

Pete

Re: 1 last question about filters

>

>Thanks for the answers on the FIR filters, Pete. Of

>course, that only led me to other questions:

>

>1. Would the " delay " then remain at what we've always

>set all designs---.25, or does that change to

>something else too?

>

>2. Would you use FIR filters for asymmetry and

>balance designs or only the coherence up and down

>designs?

>

>3. Are there any other designs I should change to FIR

>filters other than the ones mentioned? And, are there

>any designs that need the " delay " changed from

>something other than .25?

>

>4. Will there be an updated set of designs with

>corrected filters, delay settings, more design notes,

>etc. coming in the near future?

>

>Thanks again for your help!

>

>Pam

>

>__________________________________________________

>