Minimum phase for a speaker

ctviggen · « **on:** 16 Sep 2005, 11:13 am »

Can anyone point me to a site that explains (particularly, defines) minimum phase for a speaker? ETF will show minimum phase but does not describe what minimum phase is or how it can be determined. When I search on google, I get a ton of hits, but many of them are for minimum phase digital correction.

I assume that minimum phase is the smallest possible phase change from input to output, but that doesn't seem correct, as the minimum phase ETF changes relative to frequency.

ctviggen · « **Reply #1 on:** 16 Sep 2005, 11:15 am »

Never mind! JoshK had already pointed out a site to me.

konut · « **Reply #2 on:** 16 Sep 2005, 11:45 am »

Can you post the link please?

ctviggen · « **Reply #3 on:** 16 Sep 2005, 12:38 pm »

Actually, the link does not describe minimum phase:

http://www.uaudio.com/webzine/2003/december/text/content2.html

It describes linear phase (denoted by "d" in the equations). The page discusses minimum phase but does not state what minimum phase is.

So, I still do not know what minimum phase is for either a speaker or a crossover. Since the link says that analog designs cannot produce linear phase but can produce minimum phase, perhaps minimum phase is some quantification of distortion (i.e., nonlinear phase).

KJ · « **Reply #4 on:** 16 Sep 2005, 12:44 pm »

Found this site. However, it doesn't necessarily indicate how it would be applied.

-KJ

ctviggen · « **Reply #5 on:** 16 Sep 2005, 12:45 pm »

Again, I'm wrong! Dangit! The third page (in terms of printed pages) does define minimum phase: minimum phase filters have the least latency for a given magnitude response.

ctviggen · « **Reply #6 on:** 16 Sep 2005, 12:47 pm »

Thanks, KJ. Now, that's my kind of site! Get right to the equations.

JoshK · « **Reply #7 on:** 16 Sep 2005, 02:13 pm »

Wow, thanks KJ, that is a fantastic reference! This speaks my language.

What I wouldn't give to go back to grad school there!

_scotty_ · « **Reply #8 on:** 16 Sep 2005, 02:26 pm »

ctviggen said

Quote

I assume that minimum phase is the smallest possible phase change from input to output, but that doesn't seem correct, as the minimum phase ETF changes relative to frequency.

I think for the purposes of realworld loudspeakers and their performance
with respect to phaseshift you will have to ask the loudspeaker to reproduce a recognizable squarewave at a specified frequency. Given the bandwidth
limitations of realworld loudspeakers you will probably have to use Dunlavey's criteria and require that the speaker pass a 1Khz squarewave
with some limits on how far the speaker can deviate from perfection.
In the absence of any criteria defining the phase accuracy of a "minimum phase loudspeaker" the terminology degenerates into just so much marketing hyperbole.
Scotty