Re: Stress filter parameters

New Message Reply Date view Thread view Subject view Author view

Veraart (rioj7++at++fel.tno.nl)
Tue, 13 Aug 1996 14:10:27 +0200 (MET DST)

> What are reasonable numbers to use in a pfChannel's stress filter? While
> the Performer documentation explains what all the parameters are (sort of),
> it doesn't really provide any guidelines on how to go about selecting them.

I haven't succeeded in using the Performer Stress filter algorithm.
It always gives a oscilating behaviour of the stress value. This is due to
the fact that the magnitude of the control value (S[i]) is used to compute
a new stress value.
You must use the difference of the wanted and the measured
value, they are resp 1.0 and 'load', to calculate a new S[i] value.
It's best to use a proportional integrating algorithm to get a smooth
respons of the stress value. It's possible to use the hysteresis of
the load/high value to disable a change to S[i].

It comes down to writing your own stress filter and making a call to
pfChanStress(). I think the right place to do it is between pfSync() and
pfFrame(), because it's a 'time critical' setting of the channel.

Maybe the next release of Performer can have a change of the stress algorithm,
or a choice between different algorithms.

Mario Veraart
=======================================================================
List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
            Submissions: info-performer++at++sgi.com
        Admin. requests: info-performer-request++at++sgi.com

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:53:21 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.