Measuring my DIY Speakers for Crossover Design

[Nordo]

That guide is fine for complete speakers, however for individual drivers for design purpose, timing reference needs to be included to combine the measurements of multiple drivers. There’s a perfectly good guide written by the VituixCAD developer here, Nordo just needs to figure out the difference between acoustic timing reference and loopback reference so he can follow the guide using his USB mic.

I have used Kimmo's REW guide for all my driver testing measurements, but as he (understandably) has an aversion to USB mics, his guide only includes instructions for loopback reference, which makes it hard for me to realise what settings/instructions are relevant when using acoustic timing reference.

 

[dcibel]

I guess placing the mic in front of the driver being measured wasn't obvious...

 

[dcibel]

That's great.
Thanks John.
I've read everything I could find relating to acoustic timing reference that you have written, but didn't come across that one.

You won't find information on timing reference there. Instruction for evaluating completed speakers only. Funny that it's a VituixCAD oriented instruction but uses an old spreadhseet for Diffraction instead of the VituixCAD tool.

 

[Nordo]

You won't find information on timing reference there. Instruction for evaluating completed speakers only. Funny that it's a VituixCAD oriented instruction but uses an old spreadhseet for Diffraction instead of the VituixCAD tool.

I haven't read it yet, but I understand what you are saying.
VituixCAD has a lot of built-in apps that are as good any stand alone program.
John has developed REW to be fully featured as well, going well beyond is original intention of evaluating in-room responses.

I think my main problem now with VituixCAD and REW is making sure I don't double up on a setting.
For instance there is a setting in REW's Preferences/Analysis where I can add a second pulse from my reference speaker at the end of the IR sweep. If I use that, do I still have to add the timing offset in the Measurements window?

 

[dcibel]

I think my main problem now with VituixCAD and REW is making sure I don't double up on a setting.
For instance there is a setting in REW's Preferences/Analysis where I can add a second pulse from my reference speaker at the end of the IR sweep. If I use that, do I still have to add the timing offset in the Measurements window?

Did you try it? How did it affect timing?

Answer on second pulse is here:

Impulse "chirp x 2 ?

This has been bothering me for a while when using acoustic timing I seem to be getting 2 "chirps" one from the target speaker as expected at the start of the sweep, but I then hear a second "chirp" near the end of the sweep from my right side surround spealer I am sure it never used to be like...

www.avnirvana.com

Main thing to get right is location of mic. Timing offset is less important whether you include it in the measurement or not, t=0 can be offset after measurement is completed via the impulse response tab, you don’t necessarily need to remeasure anything if you forget it or set it incorrectly.

Here’s basic instruction one last time. Keep distance from acoustic reference to mic constant for all measurements, keep distance from mic to speaker baffle constant for all measurements. Measure each driver straight on. Adjust t=0 (either using timing offset when measuring, or offset after measuring) so that impulse is near t=0, I would use 0 deg tweeter measurement as point of reference. Any adjustment made here is applied to all measurements, relative timing must be maintained. Use kimmo’s guide as reference for all other aspects.

 

[Nordo]

Did you try it? How did it affect timing?

Answer on second pulse is here:

Impulse "chirp x 2 ?

This has been bothering me for a while when using acoustic timing I seem to be getting 2 "chirps" one from the target speaker as expected at the start of the sweep, but I then hear a second "chirp" near the end of the sweep from my right side surround spealer I am sure it never used to be like...

www.avnirvana.com

Main thing to get right is location of mic. Timing offset is less important whether you include it in the measurement or not, t=0 can be offset after measurement is completed via the impulse response tab, you don’t necessarily need to remeasure anything if you forget it or set it incorrectly.

Here’s basic instruction one last time. Keep distance from acoustic reference to mic constant for all measurements, keep distance from mic to speaker baffle constant for all measurements. Measure each driver straight on. Adjust t=0 (either using timing offset when measuring, or offset after measuring) so that impulse is near t=0, I would use 0 deg tweeter measurement as point of reference. Any adjustment made here is applied to all measurements, relative timing must be maintained. Use kimmo’s guide as reference for all other aspects.

Thanks again.
I've added your comments to the notes I have in front of me when I'm measuring.
I just want to be 100% certain of all the correct settings (particularly when using my USB mic) before I re-measure.
I'll have to re-measure instead of just adjusting my current IR results, as I'm not sure whether I had something else set or forgotten in REW that would upset the IR data.
I definitely had the mic and speakers set up properly in my last measuring set - although I will move the reference speaker closer this time.

Thanks for the link. John made a comment regarding the REW setting of "Adjust clock with acoustic ref" that explained the setting in simplistic terms.

 

[dcibel]

If you've got timing information and mic is on-axis with drivers, why have you decided to re-measure? Or how have you decided the measurements are no good and need to be re-done?

You can always post the measurement set here, or to some could storage like Google Drive if it's too large to post.

 

[Nordo]

If you've got timing information and mic is on-axis with drivers, why have you decided to re-measure? Or how have you decided the measurements are no good and need to be re-done?

You can always post the measurement set here, or to some could storage like Google Drive if it's too large to post.

I've had a play around with my measurements in REW and can see how altering the Windows Ref Time in the IR Windows (diff between Ref and Test distances) aligns the t=0 with the peak of the impulse.

My only concern with my measurements was that I didn't use the Timing Offset in the Measurements Window.
But if I can take care of it afterwards in the IR Windows, then as you suggested, I don't need to re-measure.

When I took the measurements, I had used "Adjust clock with acoustic ref" (i.e. the second ref pulse), so the phase data should also be in alignment.

 

[dcibel]

My only concern with my measurements was that I didn't use the Timing Offset in the Measurements Window.
But if I can take care of it afterwards in the IR Windows, then as you suggested, I don't need to re-measure.

In impulse tab, "offset t=0". Same offset value must be applied to all measurements to maintain relative timing information, so be careful with this procedure. Would be great to see an "apply to all" button here @John Mulcahy.

 

[John Mulcahy]

In impulse tab, "offset t=0". Same offset value must be applied to all measurements to maintain relative timing information, so be careful with this procedure. Would be great to see an "apply to all" button here @John Mulcahy.

You can use "Set t=0 at cursor" on the Impulse overlay graph to quickly adjust t=0 for a selected set of measurements.

 

[dcibel]

Thanks John !

 

[Nordo]

Thanks guys.
Can you check my IR images below. I think I might already have the correct t=0 offset.
The first image is "as measured", and the 2nd image is after I went into Control and applied "Offset t=0".
I can't see any difference.

 

[dcibel]

You can’t read x axis to see what you did?

Fwiw translating distance to time is easy. Sound travels 344m/s, so 0.8m * 344 m/s = 2.74ms

If you’ve applied that offset to all far field measurements you are done. Proceed to windowing and export to vituixcad.

 

[Nordo]

Oops, sorry everyone.
The difference is obvious. Not sure why I didn't see it before???

 

[Nordo]

You can’t read x axis to see what you did?

Fwiw translating distance to time is easy. Sound travels 344m/s, so 0.8m * 344 m/s = 2.74ms

If you’ve applied that offset to all far field measurements you are done. Proceed to windowing and export to vituixcad.

Sorry to be so naive, but are you saying that the first image does not have the correct offset, and the second image does?
The notes that REW added to the original measurement (first image) states that it recognised a 2.4ms delay - and that's what's confusing me.

 

[dcibel]

The offset image shows t=0 just to the left of the impulse, isn’t that where you want it to be ?

(Smile and nod)

 

[Nordo]

The offset image shows t=0 just to the left of the impulse, isn’t that where you want it to be ?

(Smile and nod)

Sorry - I just confuse myself when I over think things.
Thanks again.

 

[dcibel]

The notes that REW added to the original measurement (first image) states that it recognised a 2.4ms delay - and that's what's confusing me.

Sorry, typo above, of course simple math of 0.8m at 344 m/s is 2.325ms. But, these specifics aren't critical, all that you are doing with t=0 offset is to move the reference time near the impulse so that excess phase is minimal, making your phase plots easy to read. Really you can use the original measurements as-is, but with a very busy phase plot which will only further the confusion in design process.

FWIW default speed of sound in REW preferences is 343 m/s, minor detail but I would change to 344 m/s for more accurate distances at room temperature.

Speed of Sound Calculator

Check with our speed of sound calculator how the speed of sound in water or air changes with temperature.

www.omnicalculator.com

 

[Nordo]

Sorry, typo above, of course simple math of 0.8m at 344 m/s is 2.325ms. But, these specifics aren't critical, all that you are doing with t=0 offset is to move the reference time near the impulse so that excess phase is minimal, making your phase plots easy to read. Really you can use the original measurements as-is, but with a very busy phase plot which will only further the confusion in design process.

FWIW default speed of sound in REW preferences is 343 m/s, minor detail but I would change to 344 m/s for more accurate distances at room temperature.

Speed of Sound Calculator

Check with our speed of sound calculator how the speed of sound in water or air changes with temperature.

www.omnicalculator.com

I think at one stage I researched exactly what the speed of sound would be in my situation, but ended up just adopting 344m/sec.
I'm just about to offset all my t=0 locations and export to VituixCAD.
Even if the t=0 offset doesn't effect the essential part of my design data, doing the offset helps me understand things that are in play when measuring and designing (I'm still a long way from understanding everything).

I want to thank you again, as well as John M, AllenB, and many others who have painstakingly and very patiently taken me through this journey. Hopefully I'm going to end up with a half decent set of speakers.