How to measure direct response of speaker ?

[buzztiger]

Can REW be used to measure direct response of speaker w/o influence of room or reflection ? Or do i need a anechoic chamber to do that ?

[JohnM]

That depends on how far away from the room's surfaces (including floor) you can mount the speaker and how low you need the frequency response to extend. For example, suppose the nearest surface is 5 feet away. You can measure the speaker with REW, then go into the impulse response (IR) view and adjust the IR window to only include the time period from the initial peak of the response to about 10ms later, which is when the first reflection from that surface 5 feet away would show up. The drawback is the low frequency response only extends to 1/(time width of the IR window), so for a 10ms window that is 1/0.01 = 100Hz. To get more low frequency info you would either need to get the speaker further away from the surfaces or take it to an anechoic chamber, but even anechoic chambers have limits on the usable low frequency range depending on how large the chamber is and how much absorption the walls have.

[Anthony]

John,
Has REW always had this feature?

Gating was the one reason I used Speaker Workshop over REW for my speaker tests. I guess not anymore!

Anthony

[brucek]

Has REW always had this feature?

Yes..........

[Anthony]

I must have missed that then. I though I asked about gating a while back and was told that it was a future feature, but never checked back again.

Unfortunately my room is pretty small even for gated measurements I can only get useful stuff down to 300 Hz or so when I get rid of the first reflection.

[buzztiger]

I am actually interested in the frequency response of the speaker from 100Hz onwards. However i have something within 3 feet of the speaker in my room. Can i do a makeshift anechoic chamber using a cardboard box stuffed with either dacron or fibreglass and make the measurement ? While searching i came across 2 methods of placing the mic. One is near the speaker driver almost touching and the other is 1m away from the speaker. Which would be better ?

[JohnM]

buzztiger wrote:

Can i do a makeshift anechoic chamber using a cardboard box stuffed with either dacron or fibreglass and make the measurement ?

No, total waste of time. Why not just move the speaker into the middle of the room?

buzztiger wrote:

While searching i came across 2 methods of placing the mic. One is near the speaker driver almost touching and the other is 1m away from the speaker. Which would be better ?

Depends on what you want to measure. If you are only interested in the response of one drive unit then you can measure very close up. If you want to see how the drive units integrate in a multiway speaker you usually need to be at least 1m away (sometimes further, depending on the separation of the drive units and how the designer has set up the crossover). For more detailed info and tips on measuring speakers you are probably best to post in the DIY Speakers forum.

[Ricci]

You could always take it outside if it's not too much of a hassle.

[DrWho]

Going outside is by far the best approach. A large room is about equally as good provided the reflections are far enough away.

Gating indoors certainly works, but you have to be very aware of the frequency resolution changes. Being able to compare to anechoic (or more anechoic) measurements makes it very apparent, which I've found to be a useful learning tool.

A helpful tip when doing ground planes outdoors...put the mic on the ground and tilt the speaker forward to aim at the mic. This will prevent reflections off the ground from introducing comb-filtering into the measurement. I usually do this with in-room measurements too if the environment is conducive.

Btw, absorbing a reflection will technically work and is actually something I did back in college, but I had the benefit of some anechoic measurements to verify when that reflection was properly addressed. One of the problems is that perfectly absorptive material needs to be at least 1/4 wavelength thick to be effective...1/4 wavelength of 100Hz is like 3 feet, and it's gonna have to be thicker with real materials - and then the partially reflective behavior is going to affect the higher frequencies...