How to make quasi-anechoic speaker measurements/spinoramas with REW and VituixCAD

[sebackman]

Hi all,

Frequency measurements are well covered in this thread.

But I do from time-to-time struggle with finding a good repeatable practice to determine time alignment and phase correction to correct in the DSP. And the order thereof.

Genelec often receive good reviews and I think this not only their very nice FR but also that phase and time alignment is excellent. I use a pair of 8340 as reference when building, which is humbling.

I typically use 2-channel time-windowed MLS pulses (reasonable noise resistant) to measure and more or less follow the description in the @napilopez description to find & correct the FR. So far, all good.

When happy with FR and correction on each driver I would like to find a “best practice” for finding the flight time correction and phase correction needed for a 2- or multi-way system.

I guess flight time correction, ie drivers not being in a vertical plane can be done by “ruler”, ie measuring the physical distance, or by SW, ie measuring the difference in flight time at one spot with a certain distance from the speaker baffle.

When using the “ruler” method how do you determine the measuring point on a cone speaker? Or on a larger dome for that matter? Or driver /wave guide?

If you use SW (REW/ARTA or other?), at what distance do you measure and do you focus the mic on the tweeter or “in-between”?

How do you go about this?

In a DSP (non-FIR) all EQ’s, PEQ’s, crossovers and potential passive components after the amp will alter the phase for each driver. A way of finding the phase at XO can be to track phase for the woofer and tweeter when measuring and then apply phase shift on either to “catch-up” to the other. Not easy. This will create different outcome dependent on distance and position of the mic.

How do you go about this? What is your procedure and what settings do you use?

Maybe this is described i practical manors somewhere and if so a link would be appreciated.

Kind regards

//Rob

 

[kimmosto]

These are master documents when designing with VituixCAD or any other simulator with simple 3D geometry system using mechanical/rotation center for each radiator.
Quick manual to produce measurement data with ARTA: VituixCAD_Measurement_ARTA.pdf
Quick manual to produce measurement data with CLIO: VituixCAD_Measurement_CLIO.pdf
Quick manual to produce measurement data with REW revision V5.20.14ea27 or later: VituixCAD_Measurement_REW.pdf
Quick manual to produce measurement data with SoundEasy: VituixCAD_Measurement_SoundEasy.pdf
Other instructions in this thread for measuring complete speaker are not usable.

 

[dasdoing]

talking about a single driver, why do meassuerements in the vicinity of the dustcap have restrictions at high frequencies?

 

[mwmkravchenko]

talking about a single driver, why do meassuerements in the vicinity of the dustcap have restrictions at high frequencies?

You are finding areas that have reflections and then enigmodes, or cancellations. positive and negative air pressure fluctuations. Sound is air pressure after all.

 

[dc655321]

talking about a single driver, why do meassuerements in the vicinity of the dustcap have restrictions at high frequencies?

Diffraction will occur when waves encounter objects of a size similar to their wavelength (like a microphone and/or its support).

 

[dasdoing]

You are finding areas that have reflections and then enigmodes, or cancellations. positive and negative air pressure fluctuations. Sound is air pressure after all.

Diffraction will occur when waves encounter objects of a size similar to their wavelength (like a microphone and/or its support).

thanks guys,

I found this document: https://www.ap.com/wp-content/uploads/2021/09/AppNote-Loudspeaker-EA-Measurements.pdf

as far as I understand the following

The far field quasi-anechoic measurement is conducted at some distance greater than 3M, where M is the most significant dimension of the loudspeaker Application Note: 18 Loudspeaker Electroacoustic Measurements enclosure.

to meassure my mini single driver speaker in the "far field" the minimum distance would be 3 times it's height? 14.5cm*3 = 43.5cm?

 

[mwmkravchenko]

thanks guys,

I found this document: https://www.ap.com/wp-content/uploads/2021/09/AppNote-Loudspeaker-EA-Measurements.pdf

as far as I understand the following



to meassure my mini single driver speaker in the "far field" the minimum distance would be 3 times it's height? 14.5cm*3 = 43.5cm?

It's more a question of what you are trying to measure. If you are looking for a pretty anechoic measurement find your 3 or 4mm best position along the cone and have at it. It looks pretty, and is basically a measurement. If you want truth as to what your hearing do a listening position measurement and try a moving mic type of an averaging. If you are using REW there is enough information here to do it well already. That will give you a reasonable idea of what you have at your listening position. As a point of interest. Most mastering work is done at or around 85db. Our ears are the most evenly sensitive across the widest frequency range at and around this point. It is also a reasonably loud point to check out your system. Spikes and sharp dips are not what is the issue when you are listening. Our ear/brain system is very forgiving of those. It's the broad bumps particularly from 1 to 4 kilohertz that our ears are the most sensitive. In the bass region it is a similar thing. Below about 500 hertz you can rapidly change a voice, and below 200 hertz you will change the timbre of many instruments. Accuracy is most easily determined by recording familiar sounds like a family member, your car door slamming etc. You know these sounds intimately. A little food for thought.

Mark

 

[dasdoing]

It's more a question of what you are trying to measure. If you are looking for a pretty anechoic measurement find your 3 or 4mm best position along the cone and have at it. It looks pretty, and is basically a measurement. If you want truth as to what your hearing do a listening position measurement and try a moving mic type of an averaging. If you are using REW there is enough information here to do it well already. That will give you a reasonable idea of what you have at your listening position. As a point of interest. Most mastering work is done at or around 85db. Our ears are the most evenly sensitive across the widest frequency range at and around this point. It is also a reasonably loud point to check out your system. Spikes and sharp dips are not what is the issue when you are listening. Our ear/brain system is very forgiving of those. It's the broad bumps particularly from 1 to 4 kilohertz that our ears are the most sensitive. In the bass region it is a similar thing. Below about 500 hertz you can rapidly change a voice, and below 200 hertz you will change the timbre of many instruments. Accuracy is most easily determined by recording familiar sounds like a family member, your car door slamming etc. You know these sounds intimately. A little food for thought.

Mark

I want to take the quase anechoic of the speaker sans bass (it doesn't have any anyways) so I can study what my room does to it in LP distance.
Guess I just play around with distances and windows and try to guess which is the most accurate

 

[mwmkravchenko]

I want to take the quase anechoic of the speaker sans bass (it doesn't have any anyways) so I can study what my room does to it in LP distance.
Guess I just play around with distances and windows and try to guess which is the most accurate

No guessing required. Use the tools I suggested. Sounds across the frequency range that you are accustomed to. Heck a cell phone can make reasonably good recordings now.

Mark

 

[dasdoing]

I want to take the quase anechoic of the speaker sans bass (it doesn't have any anyways) so I can study what my room does to it in LP distance.
Guess I just play around with distances and windows and try to guess which is the most accurate

Constructed professional anechoic chamber haha

as you can see i didn't put much effort into this. this tiny speaker is for experiments (also as rears for the rare occasions I watch a movie).

A few FWDed measurements: numbers indicate approximate driver distance.

as far as I can tell the 22 and 34 would be in the far field, since 2 and 10 seam to have a diffraction bumb at the bottom end?


EDIT:

here is the 22 with a groundplane-ish one I took the other day with even less afford. I guess the ground one looks more realistic? gona retry that

 

[Jon AA]

To be able to measure down to 20Hz, the time until the arrival of the first reflection must be at least >1/20 seconds, i.e. 50ms, which means that the first reflecting surface should be at least around 10m away - but even then a gate is still needed and the result is strongly smoothed.

I'd like to revisit this apparent contradiction and ask if anybody has practical experience with measurements on this threshold. You certainly won't go wrong as far as accuracy in mid to high frequencies by gating before the first reflection. But the Arta excerpt certainly appears to be implying, if the first reflection comes from 5 or more times the measurement distance, the reflection will be so low in level it will have a negligible impact on the overall response such that it doesn't necessarily need to be gated out.

Has anybody taken ground plane measurements at around the 5:1 suggested limit (10m for a 2m measurement or 5m for a 1m measurement to the nearest wall) to see the effect that first reflection actually has? Is it large enough to give up better bass resolution by not including it? Even if 1 dB errors occurred (I assume mainly in the bass region?) may that still not be more accurate than nearfield sums and a calculated baffle step (all of which can be prone to errors, time consuming and very difficult to do, depending upon the speaker)?

Why am I asking? Here's a pic of the "open end" of my new shop, currently under construction:

It'll be quite a while before I'm able to take measurements in there and test things for myself, but the curiosity is killing me!

 

[ChargedCapacitor]

View attachment 292489


I'd like to revisit this apparent contradiction and ask if anybody has practical experience with measurements on this threshold. You certainly won't go wrong as far as accuracy in mid to high frequencies by gating before the first reflection. But the Arta excerpt certainly appears to be implying, if the first reflection comes from 5 or more times the measurement distance, the reflection will be so low in level it will have a negligible impact on the overall response such that it doesn't necessarily need to be gated out.

Has anybody taken ground plane measurements at around the 5:1 suggested limit (10m for a 2m measurement or 5m for a 1m measurement to the nearest wall) to see the effect that first reflection actually has? Is it large enough to give up better bass resolution by not including it? Even if 1 dB errors occurred (I assume mainly in the bass region?) may that still not be more accurate than nearfield sums and a calculated baffle step (all of which can be prone to errors, time consuming and very difficult to do, depending upon the speaker)?

Why am I asking? Here's a pic of the "open end" of my new shop, currently under construction:

View attachment 292490



It'll be quite a while before I'm able to take measurements in there and test things for myself, but the curiosity is killing me!

Awesome place to take a measurement!

It's a bit overkill; you can get a great 0-180 degree GPM in a standard two car garage down to 200 hz, gated. Then, the single 10-200 hz measurement can be taken anywhere outside with now boundaries within ~50 feet.

 

[Jon AA]

Awesome place to take a measurement!

Thanks!

It's a bit overkill; you can get a great 0-180 degree GPM in a standard two car garage down to 200 hz, gated. Then, the single 10-200 hz measurement can be taken anywhere outside with now boundaries within ~50 feet.

Yeah, but I'm trying to get away from splicing. There aren't any large, flat, concrete/asphalt places in the area for separate bass measurements anyway (rural area).

Don't get me wrong, I'm not having delusions about being able to measure large subs in this place. But bookshelves, even towers that only go down to the typical 30-40 Hz, is a completely different ballgame than trying to get precise measurements down to 10 Hz. I'm hoping it can be done, but we'll have to see.

 

[Newman]

@Jon AA great discussion, thanks.

I can’t answer your question (“why isn’t GPM the recommended approach taken by this thread”), but I have a related question:-

Why not turn the omni mic in the photo (extract below) 90º and use its 90º calibration curve? The way it’s shown could have reflection problems from the stem of the mic back to the capsule, yes?

Cheers

 

[fluid]

Why not turn the omni mic in the photo (extract below) 90º and use its 90º calibration curve? The way it’s shown could have reflection problems from the stem of the mic back to the capsule, yes?

It is a consideration for a full frequency ground plane measurement but for low frequencies it is not an issue.

http://www.excelsior-audio.com/Publications/AES137_PD16_Measuring_Loudspeaker_Systems_rev06a.pdf

 

[Newman]

I was actually thinking 90º sideways, not upwards. And yes, for full range measurement.

 

[fluid]

It is the distance of the centre of the mic capsule to the ground that determines the influence of the reflection, so 90 degrees with the mic buried in the floor would eliminate it, any other orientation it will still be a factor. It is so easy to make good gated high frequency measurements and splice them I can't see much value in trying to make a truly full range ground plane measurement.

 

[Newman]

I see; you are thinking about reflections off the floor to the centre of the mic capsule.

I am thinking about reflections off the shoulder of the microphone's body. I believe it can come into play, or am I fussing unnecessarily? And turning the body 90º to one side (and using the 90º calibration) will be a good solution?

The path length would equate to interference at about 1 kHz.

cheers

 

[fluid]

The floor to mic capsule limits the upper end bandwidth. In a ground plane the mic tends to be on the floor or held pointing down by a small fixture which limits the interaction.
Reflections from the mic body are usually low enough to not be easily seen, but any mic holder can make a real difference if it's not well thought out and then you get ripple.

From one of B&K's old papers

 

[Jon AA]

It is a consideration for a full frequency ground plane measurement but for low frequencies it is not an issue.

http://www.excelsior-audio.com/Publications/AES137_PD16_Measuring_Loudspeaker_Systems_rev06a.pdf

I'm not sure I buy that chart at face value. There was a formula published way back when that apparently had an error which made the HF error much larger than it should be. I'm not sure everybody got the memo. Erin certainly didn't find that much error: