Group Delay Question

[Oldmics]

Gentlemen

When referring to the term "group delay",should the reference be made to the longest time delay regardless of frequency band (lows or hi"s)?Would that then be the "group delay" of that particular crossover network regardless of which frequency is delayed the longest?

It seems more precise to refer to the time delay referenced to the particular frequencys.

I"m just not sure how it is spoken of when passive crossover networks are discussed.

Oldmics

[whgeiger]

OM,
Analog signal filters introduce time delay between their input and output terminals. This delay can be represented as a phase shift when a sine wave passes through the filter. For most filters, such shifts exhibit frequency dependence that is imbedded in the filter's transfer function. When the phase factor [t=f(w)] is isolated, group delay [Dg] may be characterized by taking the derivative with repect to frequency [w]; i.e., Dg = dt/dw. This is the slope of the phase vs. frequency curve when plotted linearly.
For more details see:
Filter Design Guide
http://frequencydevices.com/guide/fullguide.html#analog
Regards,
WHG

[hector.murray]

I was wondering what group delay ment at all when I searched for and found this thread. This explanation is probably too simplistic an explanation and was found in the bassbox pro users manual.

Ideally, there should be no delay from one frequency to the next. However, a uniform delay is usually not a problem as long as it is modest or, if it is significant, the speakers are not used for live sound reinforcement or lipsync to a video image. The most important thing to avoid is a sudden large change in group delay since this can be very audible. An ideal plot line would be flat or have only a modest slope.
Copyright 1992-2002 by Harris Technologies, Inc.

So I'm thinking that the graph below would represent a bad idea

[B&KMan]

Gentlemen

When referring to the term "group delay",should the reference be made to the longest time delay regardless of frequency band (lows or hi"s)?Would that then be the "group delay" of that particular crossover network regardless of which frequency is delayed the longest?

It seems more precise to refer to the time delay referenced to the particular frequencys.

I"m just not sure how it is spoken of when passive crossover networks are discussed.

Oldmics

Many other info if you search time, delay,

look this thread too (it is my problem.)


http://www.audioheritage.org/vbullet...ght=2405+delay


I hope this help you little bit....

jean...