Measurements on loudspeakers aimed at designing the crossover network

[2slow4u]

I need to measure the individual speakers of an acoustic speaker, already mounted on the baffle, to be able to simulate and design the crossover network. To do this I need to measure the correct phase relationships and delays between the various speakers. I have no experience with REW, is it possible to do what I just explained? If so, what is the correct procedure for making these measurements? If I understand correctly, REW sets t = 0 on the maximum pulse peak, or on the first sample. But in these cases I lose the delay information due to the different speaker offsets, so what I would like to do is not possible. Am I wrong? Thank you!

 

[John Mulcahy]

You just need to measure with a timing reference.

 

[2slow4u]

Hi there!
I don't think I understand, English is not my language.
If I use a loopback, isn't it just that all measures will have the same delay? I, on the other hand, need to grasp the delay between tweeter, midrange and woofer. I need the system to include the acoustic offsets of the various speakers in the measurements. I would need step-by-step instructions on how to do this if possible. Thanks again.

 

[John Mulcahy]

If I use a loopback, isn't it just that all measures will have the same delay?

No. Measurement with a timing reference does exactly what you need.

 

[2slow4u]

Thanks a lot John!

 

[2slow4u]

Hi, I'm going back to the topic because I can't find a solution. By measuring the speakers using loopback, I get a noticeable excess of phase:

It's normal?
Then trying to follow this tutorial ...

...I visualize the minimum phase in Virtuix Cad and by introducing a negative delay in the appropriate field, to show measured phase without excess delay as described in the tutorial, I cannot really copy the shape of the minimum phase curve. This is the best result I can get:

Measured phase and minimum phase do not coincide. Where am I wrong?
Thanks for your patience.

 

[John Mulcahy]

The timing reference maintains relative timing between your measurements. There will be a delay between the loopback connection and the measurement captures mainly due to the distance from driver to microphone, as explained in the second paragraph of the help I linked:

If a loopback is selected the reference channel signal must be looped back from output to input on the interface and measurements will be relative to the loopback timing. Usually this means measurements will have a time delay that corresponds to the time it takes sound to travel from the speaker being measured to the microphone, which can be removed for subsequent measurements using the Timing offset.

REW can estimate the delay for you using the "Estimate IR delay" button in the graph controls. That would typically be used on the tweeter measurement, the delay that is found can then be applied as a timing offset for further measurements so it is removed when the measurement is made. For existing measurements the delay could be removed using the "Offset t=0" feature in the Impulse graph controls, the same delay must be removed from each measurement to preserve their relative timing. That is much the same as you were doing manually in VituixCAD, though your measurement looks to be inverted, hence the 180 degree phase difference between measured and minimum phase.

 

[2slow4u]

For existing measurements the delay could be removed using the "Offset t=0" feature in the Impulse graph controls, the same delay must be removed from each measurement to preserve their relative timing. That is much the same as you were doing manually in VituixCAD, though your measurement looks to be inverted, hence the 180 degree phase difference between measured and minimum phase.

Forgive me if I come back to the subject but I continue to have difficulties. I have two measures, one on the tweeter and one on the midrange, using the loopback and keeping the mic in the same position. Starting with tweeter I have this situation:

I then ask REW to estimate the delay, take note (2.617 ms, to be applied later to the midrange measure), Then I click on "Shift IR" to remove the delay.

I get this:

... but returning to the SPL screen, I find a totally altered FR curve:

Please, where am I wrong?
Thank you.

 

[John Mulcahy]

The problem is the left IR window you have applied. Set it to Tukey 0.25 and set the width to at least 1 ms. What were your reasons for changing it and setting it so narrow?

 

[John Mulcahy]

Note also that to have REW automatically apply the estimated delay as a timing offset for subsequent measurements you should use a V5.20 build.

 

[2slow4u]

The problem is the left IR window you have applied. Set it to Tukey 0.25 and set the width to at least 1 ms. What were your reasons for changing it and setting it so narrow?

There is no particular reason for setting the left window so narrow. Rather I preferred to set it to Blackman-Harris 7 because in this way I obtained cleaner and more legible curves. Are there any problems if I use Tukey 0.25 on the left and Blackman-Harris 7 on the right? With these settings I can actually successfully apply your suggestions. Thanks again for your kind advice!

 

[John Mulcahy]

If you wish to use BH-7 on the left set it the same width as the right window. There isn't any problem with having different window types on left and right. The left window has little influence provided the window reference position is correct and the window it is not so narrow that it truncates the response, Tukey 0.25 is a sensible choice for it.

 

[sm52]

John
Is it possible to see how the graphs change when changing the type of the left and right windows without a new measurement? The measurement is done, I change the window in the 'Analysis' section and I want to see what has changed. Need a new measurement?

 

[John Mulcahy]

The Analysis defaults only apply to new measurements. For existing measurements you can change the selections in the IR Windows dialog and use the Apply Windows button.

Choosing window types for an impulse response differs from choosing a window for signal analysis (e.g. in the RTA). Tapering of the IR window reduces the levels of reflections and also reduces the frequency resolution a little (compared to a rectangular window of the same width). The default choice of Tukey 0.25 was made to preserve the response including its reflections and their influence out to the point where noise begins to dominate, then taper it from there.

 

[2slow4u]

Note also that to have REW automatically apply the estimated delay as a timing offset for subsequent measurements you should use a V5.20 build.

I have 5.20 beta6 but I didn't understand how to automatically apply the estimated delay as a timing offset for subsequent measurements

 

[John Mulcahy]

There have been more than 60 V5. 20 builds since beta 6, if you use a current build you will see the button on the Estimate IR Delay dialog to set the timing offset.

 

[2slow4u]

There have been more than 60 V5. 20 builds since beta 6, if you use a current build you will see the button on the Estimate IR Delay dialog to set the timing offset.

Opsss!
If I install the latest build will i lose all software settings?
Thank you.

 

[John Mulcahy]

No, settings are kept.

 

[2slow4u]

Thank you, nice feature!
But this is new to me:

Please, what is it about?

 

[John Mulcahy]

Modal frequencies, based on the room simulator dimensions or dimensions entered in the modal analysis panel of the EQ window. A selection in the graph controls turns them on/off.

 

[2slow4u]

Hi John, forgive me but I don't understand anything anymore!
Before installing the latest build (before I used the old 5.20 beta 6), I opened REW, made for example an axis response to the tweeter and found the pulse centered at "x" ms. ;

If I wanted to eliminate the delay, I had to use the "estimate delay" button.
Well, at your suggestion, I downloaded an updated release of REW because it allows me, when I need it, to eliminate the delay in the tweeter measure and use this delay as a reference for subsequent measures on the other speakers. I thought that this feature was activated only by clicking on "estimate IR delay" and immediately after on "shift and update timing offset". Instead today I opened REW, first measured the tweeter, then the midrange and to my surprise I realized that the IR peak was automatically positioned at 0ms on the tweeter and at 0.21ms on the midrange.

I thought I should be the one to ask the software to do this for me when I needed it, but it all happened automatically. Why did this happen? Why did I click on "shift and update timing offset" days ago, and from that moment REW does it automatically for me? I am a little confused.
Thanks!

 

[John Mulcahy]

Difficult for me to say what you did or didn't do days ago. If you have selected to measure with a timing reference on the Measure dialog and a timing offset figure is on the dialog, it will be used.

 

[2slow4u]

In the previous session, two days ago, I simply measured the tweeter, then clicked on "estimate IR delay" and immediately after on "shift and update timing offset". So I measured the midrange. I thought that this was only valid for that session, and that upon reopening the program, this would work as I used to see until I clicked again on "estimate IR delay" and immediately after on "shift and update timing offset". Obviously I was wrong, in fact only now I noticed that "timing offset: 2.94 ms" appears in the measurement dialog.

I didn't set that value in the dialog, I found it like this today. Is this a value that REW entered two days ago when I made those measurements?
Thank you!

 

[John Mulcahy]

Yes, that is the timing offset that is updated when "Shift and update timing offset" is clicked.

 

[2slow4u]

So to reset everything is it enough to set to "0" that value? Thank you so much, your support is really great!