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

[PhilS]

Hey, I have not read all the pages and posts, but I was wondering, with the weather turning for the worse, I wanted to do some measurements of my subwoofers indoors. Can I use the same method outlined here for nearfield bass measurements?

I live in the UK, the weather is never suitable for outdoor measurements , nearfield measurements are my only option and this process is what I use

 

[TwoEyedJack]

I suppose the reason my measurement doesn't look so clean and I only get combfilter-free FR below 2ms is because of the space I am using for the measurements? The speaker was 1.5m from the floor and 1.7+ from nearby items/furniture/walls. I will need to get outside to increase my resolution?

 

[fluid]

1.5m from the floor should give you 6.29ms of reflection free time. You also need to consider reflections from the microphone to the stand or clip or other objects near to the microphone. Zoom in the impulse response to see just where the reflections are, then you can work out where they might be coming from. If it is less than 6ms then it is not the floor, but it could be the speaker stand, mic clip etc etc.

 

[Dave Bullet]

How big is the baffle you are measuring? A rule of thumb assuming a baffle barely wider than the largest driver is mic distance at least 3 times that width. So for a nominal 200mm wide baffle for a 2 way 6.5'" woofer design. You can probably get down to 600mm mic distance without masking baffle diffraction or driver summation accuracy. You can see tools like vituixcad to model and optimise at a listening distance e.g 2.5m

The above will gain you some resolution. For a 3 way this doesn't really help if you are crossing below the gate as is often the case

 

[Newman]

1.5m from the floor should give you 6.29ms of reflection free time.

Also check ceiling height must be at least 3m for the above to work.

 

[TwoEyedJack]

First reflection is at the 1.8ms. The width of driver is 7cm and the baffle is 15cm. The boom of the mic is coaxial with the mic but the base is a bit bulky, however the mic extended well past the base, around 20cm ( placed 1m from the center of the driver)
Indeed the celling is at a Hight of 2.6m so the speaker being at 1.5m isn't in the midpoint between the celling and the floor...silly of me. Would I get better results if i placed the speaker at 1.3 meters for the floor?
In 1.8ms sound travels 0.6m thats a shorter distance than even the mic was placed from the driver how is that even possible ?

 

[fluid]

The whole impulse response looks messy and unsettled, that is not a single reflection from a specific object. Some pictures of the speaker and setup may be worth more than 1000 words of explanation. Putting the speaker at the centre of the room height will give you the longest reflection free period so long as no other surfaces are any nearer. But given the impulse above something else is happening.

 

[Newman]

In 1.8ms sound travels 0.6m thats a shorter distance than even the mic was placed from the driver how is that even possible ?

It's best if you post a photo of your setup for measurements. It's certainly possible for first reflection to be 1.8ms/0.6m delayed, after all, that is the additional path length, for a total path length of 1.6m.

 

[Dave Bullet]

Make sure you have highly reflective surfaces. Yes that's right. Don't try and soften the reflections otherwise they become harder to see in the impulse.

Also put your mic on a long boom. Ideally the stand and any clip will be outside the measurement sphere. Otherwise you'll see high frequency ripple above about 8Khz that are reflection artefacts from the mic mount and stand

 

[No. 5]

First reflection is at the 1.8ms. The width of driver is 7cm and the baffle is 15cm. The boom of the mic is coaxial with the mic but the base is a bit bulky, however the mic extended well past the base, around 20cm ( placed 1m from the center of the driver)
Indeed the celling is at a Hight of 2.6m so the speaker being at 1.5m isn't in the midpoint between the celling and the floor...silly of me. Would I get better results if i placed the speaker at 1.3 meters for the floor?
In 1.8ms sound travels 0.6m thats a shorter distance than even the mic was placed from the driver how is that even possible ?
View attachment 412792

Like Fluid said, that IR does not look right. It could be environmental or electrical noise, but sometimes when I see an IR like that, restarting REW or the computer clears it up for me.

 

[TwoEyedJack]

Well, my setup is very junky... the mic is a dayton imm-6, and I don't currently have any audio interface, only some dongles and a schiit fulla E so I have to boost the mic signal a bit to get a working measuremt. The mic cant be mounted on a normal mic stand so i made one.

Driver and mic are both 1.4 meters from the floor celling is 2.4 meters. ( speaker upside down so that driver is closer to the middle point between ceiling and floor)
Bookshelf is 1.7m away, objects on the other side of the camera are 2m+ away.
Speaker in this case is a badly designed DIY sealed cabinet using the PS95 driver that I wish to EQ ( front baffle 15cm W x 28 cm H driver not flush mounted)

Measurement looks a bit better than the previous maybe because i improved the stand (now first strong reflection is at around 5ms but there is still "stuff" before.)

 

[Newman]

Thanks for that.

Might be worth pointing the mic directly up straight at the ceiling and using its 90º correction curve...being an omni mic.

A speaker like that would probably be sending some baffle edge diffraction towards the mic, and also some baffle step diffraction, where some lower frequencies diffract off the rear edge of the speaker (which would be not far from the 0.6m added path length arrival seen in your plot). Neither of which would be due to measurement technique, but an actual property of the speaker.

 

[Fledermaus]

FWIW : I experienced once randomly fuzzy impulse traces like those, spent time troubleshooting every possible cause including PC latency, to find out eventually that it was the audiophiloid USB cable I was using that was defective

 

[TwoEyedJack]

Thanks for that.

Might be worth pointing the mic directly up straight at the ceiling and using its 90º correction curve...being an omni mic.

A speaker like that would probably be sending some baffle edge diffraction towards the mic, and also some baffle step diffraction, where some lower frequencies diffract off the rear edge of the speaker (which would be not far from the 0.6m added path length arrival seen in your plot). Neither of which would be due to measurement technique, but an actual property of the speaker.

I don't think the mic has a 90º correction curve... What would be the point of pointing it straight up?

So there are no issues with the setup? The soundcard?Amp?Wiring?

I have modeled the Baffle step and edge diffractions in VituixCad and Jeff Bagby's excel sheets. But would this affect the impulse response ?
(Diffraction even from the rear corner ? )
From other measurements of this driver (like the dayton audio copperhead, as well as the manufacturers data ) the respond seem to miss some characteristic features like the spikes after 10khz? Or I shouldn't even compare them?

Never the less, if this measurement encompasses baffle step and diffraction, should I consider it usable for my purpose of EQ? what about the dip around 8-10khz?
Should I use the same measurement setup for other speakers?

 

[TwoEyedJack]

I measured 2 other speakers to see if the behavior persists.

PS95 Speaker
Something going on around 1.5ms-2.5ms probably diffractions. Strong Reflections at 5.5ms

Second speaker. Still diffraction issues but to lower extend First strong reflection still at 5.5ms.

Last speaker. The reflection is at the same point, 5.5ms.

All the examples have issues with diffractions.

 

[mwmkravchenko]

Try moving something around in the room. You will quickly figure out the true source of your reflections. Sometimes it's not what you would expect.

 

[No. 5]

I don't think the mic has a 90º correction curve... What would be the point of pointing it straight up?

If it is oriented that way, any potential reflective surfaces on the mic holder will not be behind the microphone to be reflected back to it. Since the mic body is so small, there will probably only be around 2dB or so of difference to on axis, and only at very high frequency.

So there are no issues with the setup? The soundcard?Amp?Wiring?

The impulse itself is very messy, as is the time after it. Therefore, I strongly suspect that there is electrical noise being injected. First thing to examine would be wiring. But it could also be the dongle, I have a dongle that has awful crosstalk, like, it makes its own loopback level of crosstalk. Does your laptop have a TRRS mic/headphone jack?

Just to prove that electrical noise can look like that, here is what an IR response should look like when measuring a speaker indoors:

And this is one with a bad cable:

Everything else was the same between these two, and the furnace was even running for both for some background acoustic noise.

You could see what the RTA looks like with no signal coming out of the speakers, or you could do a measurement without the speaker connected and one with.

I have modeled the Baffle step and edge diffractions in VituixCad and Jeff Bagby's excel sheets. But would this affect the impulse response ?
(Diffraction even from the rear corner ? )

Nope.

From other measurements of this driver (like the dayton audio copperhead, as well as the manufacturers data ) the respond seem to miss some characteristic features like the spikes after 10khz? Or I shouldn't even compare them?

I wouldn't be too surprised if there are differences above 10kHz, it could be unit variation or just the the mic isn't in the same place relative to the driver; the sound field that it's producing up there will be a touch chaotic.

Never the less, if this measurement encompasses baffle step and diffraction, should I consider it usable for my purpose of EQ? what about the dip around 8-10khz?

There is a lot of frequency response ripple and noise, I wouldn't trust it to do much EQ.

Should I use the same measurement setup for other speakers?

Once you get it working properly.

 

[fluid]

Your physical measurement setup of the speaker looks OK. Diffraction off the speaker is part of the speakers response and is normally done within 1ms unless it is very big or particularly diffractive. Diffraction off the speaker itself has nothing do do with the measurement problem.

The pre impulse response is clean and flat and the usual linear phase DAC filter pre ring is present, the problem occurs after the speaker plays, it corrupts the speakers impulse and the reflections impulse. This could be either an electrical problem in the wiring setup, some kind of feedback, a computer audio setting, such as some kind of surround sound or windows enhancer type setting, or mechanical from the mic stand contraption.

There is a regularly repeating frequency across all of the interference that presents as ripple but there is also something else.

You may want to plug it straight into a phone and use an app to test what kind of response you get. That will help to narrow down where the problem is.

 

[Newman]

I don't think the mic has a 90º correction curve... What would be the point of pointing it straight up?

Check the Dayton download files for your mic, just to be sure. My UMIK came with 0º (blue) and 90º (green) curves, shown below.

Using the mic at 90º avoids axial reflections off the shoulder of the mic (A) and the stand (B) that you put it on, see photo below.

So there are no issues with the setup? The soundcard?Amp?Wiring?

I have modeled the Baffle step and edge diffractions in VituixCad and Jeff Bagby's excel sheets. But would this affect the impulse response ?
(Diffraction even from the rear corner ? )

We are not trying to eliminate the diffraction effects from the measurement, since they are part of the speaker's imperfection.

But they can contribute delayed sound arrivals at the mic, so if we want to understand ripples on the impulse response, they are pertinent. Hence I mention them, C and D on the photo above.

In particular, we are looking for a sound source for 0.6m additional path length. The only one I can see in your setup that looks like 0.6m, however insignificant, is D. That is why I mention it.

Never the less, if this measurement encompasses baffle step and diffraction, should I consider it usable for my purpose of EQ? what about the dip around 8-10khz?

Regarding the dip, are you sure that your mic calibration curve is being applied?