Phase correction

[filio]

Hello from France,

Maybe you are going to be of help, I measure my one way loudspeakers at sweep point (R+L together) and I found a phase going from +600° to -400° from 40Hz to 300Hz and then it is plus or minus flat up to 20 000Hz.
Does it makes sense please? And how could I correct this?

Thank you for your help

 

[filio]

Please see hereunder

 

[Dan Twomey]

Have a look at my thread.
Phase is a out by a mile?

Perhaps we have something in common?

Regards,
Dan

 

[jtalden]

Please see hereunder

Post the mdat file here and I will take a look at it. It is likely just the normal effect of the room influence.
[FYI: It is best to measure the channels separately rather than together when we are looking at mid and high frequency as well as the bass range. I still will be able to answer this question using this measurement however.]

 

[filio]

Hello Dan, Hello JT,

Yes, Dan, I had a look ! It is very similar indeed! The first one who will find will help the other! Your amplitude graph is flatter I also see... For your information after too much correction, I came back to a lighter correction and it was more musical. But I know I will have to work on it later...

JT, please find hereunder 4 files. The initial file is the last one (mesure du 10 mai sweet point).
The 3 others are for R / L / R+L measurement.

The micro was at the same place for all measurements (sweet point) and was directed toward the center of the 2 loudspeakers. The distance between the micro and each loudspeaker was adjusted with a wire and should be very close.
As I said, the loudspeakers are 1 full range drive unit types.

Thank you for your help and staying at your disposal,

Regards,

Filio

 

[jtalden]

I just opened the last one (see attached). You used acoustic timing and it resulted in a -45ms delay of the measurement impulse. That is unusual. Normally the L reference sound would arrive with the same delay at the mic as the R measurement channel sound which would result in the impulse being near 0 ms rather than at -45 ms. Perhaps you have 45 ms 'timing offset' set into the measurement panel.
Anyway, I first adjusted the impulse to 0 ms. and then applied 7 cycle FDW to allow viewing the direct sound phase. The direct sound phase trace now looks correct. You should be able to do this for other files as well.

 

[Dan Twomey]

JT,

That's awesome. Now let's see if I can do the same with mine! What did I learn from your example here.
Using Shift IR is the first step to a better phase reading. Then apply FDW as low as 2 cycles as you suggested. Here's my interim results. Can I take anything away from this such as my two LR crossovers show the most mismatch (phase wise) right at the 2.5Khz crossover point?

Regards,
Dan

 

[filio]

Thank you JT for this!!!

I tried to get the same result but with no success till now. How did you please adjusted the impulse to 0ms?! You are right, I saw a timing offset of something like 22ms on the measurement panel, but not 45ms if I remember well.

Further than this, I would have 2 questions regarding the result you did get:
+ In low frequencies, we see a deviation toward the +, can this be corrected, does it have a consequence on sound quality please?
+ And also, we see a +/- 40° result on the phase, would you say this is acceptable or does the correction of this will have an impact on sound quality please?

Thank you for your help again,

Regards,

Filio

 

[jtalden]

How did you please adjusted the impulse to 0ms?!

In this case I located the cursor near the impulse and selected 'Set t=0 at cursor'.

+ In low frequencies, we see a deviation toward the +, can this be corrected, does it have a consequence on sound quality please?

We are very sensitive to SPL changes. There is no significant phase impact on sound quality when talking about a single driver. It can be very significant when different drivers are playing the same frequency range. Examples are; a crossover range, or multiple SWs. That is why we we are careful to create a favorable phase relationship between the drivers in those cases.

The room effects are the primary reason the phase looks chaotic when measured from the LP, hence the FDW was used to help see the direct sound phase. A close mic position is another way to see the direct sound phase of a driver.

We can use an FIR filter to remove some of all of the phase rotation of an audio system including that from a SW if we have the needed equipment.

+ And also, we see a +/- 40° result on the phase, would you say this is acceptable or does the correction of this will have an impact on sound quality please?

No impact.

 

[Dan Twomey]

Using JT's approach I've achieved great success in straightening the phase of my point source speaker as seen below. However, although I can achieve dramatic improvements with my line array measurements. Getting close to minimum phase is still out of reach.

Regards,
Dan