minimum vs real (unwrap) phase in ARTA

[crazyhub]

Hi all,

Who can explain me the differences between these two sorts of phases?
And therefore what sort is more appropriate to measure drivers mounted in an enclosure before designing the passive filter?
And where?: in-axis of each driver or at ear height (usually tweeter height)? (I usually measure at 1 meter / 39")

Thanks (I have no scientific background, so please no or little math)

[jerome69]

Hi,

See http://www.winmls.com/hjelpfiler/phasetype1.htm

Wrapped phase
Restricts the phase range to –180 to +180 degrees, or –pi to +pi if radians is chosen as y-axis. The phase of a measurement will usually exceed this range, which results in jumps in the phase.

Unwrapped phase
The unwrapped phase is not restricted to the wrapped range, and has no jumps. The unwrapped phase is computed from the wrapped phase by changing absolute jumps greater than 180 degrees to their 2*180 degrees complement. Note that if the phase curve is not smooth, the computation of unwrapped phase may give a wrong result. Because of this, it is not recommended to select unwrapped phase when displaying the phase of a room measurement. View the phase as wrapped instead.

Minimum phase
The minimum phase is computed by taking the Hilbert transform of the log magnitude frequency response. If a system is minimum phase this means that for a given frequency response, the input energy will be transferred to the output in the least amount of time.


Excess phase
The excess phase is the difference between the actual phase and the minimum phase, the phase shift in excess of the minimum phase shift.
For excess phase measurements the phase influence of the microphone can usually be ignored since a microphone usually has a nearly minimum phase response.

For the other question it depends on the simulation software, your measurement place and the desired response... There is no better method.

Cheers.

[crazyhub]

Hi Jerome,

Thanks for this link.
Usefull to give me what I think the best way to proceed with for my measurements (exporting with minimal phase, 50cm in-axis of each driver mounted on the enclosure).

Cheers

[teodorom]

Bonne nuit Jerome,
It seems that you know the topic, so I ask for some help.
Me, I'm trying to reproduce the results of LEAP or FRC (the Excel spreadsheet), but I get quite different results. Even after the unwrapping.
Can you help me please ?
Thanks
Teodoro

[jerome69]

Hi Teodoro,
I think I can't help a lot

This a phase curve (in red) of a measured loudspeaker with ARTA. I extract the minimum phase with ARTA (black) and Response modeler. With response modeler I extract the phase with all the points (2Hz-40kHz - blue) and with a part of these points (300Hz-20kHz - purple).

There are serious deviation in the phase curves. I can't help because I haven't have any idea of what's happened in these software.

I recommend to use your measurement software to extract all minimum phase if you want precise simulations. Don't forget phase is something relative, you should have the same reference.

Regards.