[kinku]

I am wondering what are the softwares avaialble for measuring room acoustic problems. I am considering REW with lot of freedom with hardware to use. I am wondering how accurate REW is? Is there any other available with the same freedom of REW to use custom hardware?

[Ethan Winer]

REW is extremely accurate. It's pretty much the standard these days. As for hardware, this might help:

Room Measuring Primer

--Ethan

[kinku]

Thanks Ethen. I am deciding on a mic.ECM 8000 VS Dayton EMM-6.I know they have to be calibrated. Dayton comes with calibration which they claim as very good. Calibration is more or as expensive as mic.Any suggestions?

[Ethan Winer]

You don't really need a calibrated microphone for room measuring. You might need that if you measure rooms for a living, visiting people and certifying their systems. Even budget microphones of the correct type are plenty flat enough. A microphone that's off by 1 or 2 dB is insignificant compared to the 30+ dB variations added by your room! More here:

Comparison of Ten Measuring Microphones

--Ethan

[Alex Reynolds]

Thanks Ethen. I am deciding on a mic.ECM 8000 VS Dayton EMM-6.I know they have to be calibrated. Dayton comes with calibration which they claim as very good. Calibration is more or as expensive as mic.Any suggestions?

Either would work fine. The cal file for the Dayton is nice to have but not really necessary for home measurement as Ethan has stated.
In either case, once you do get the mic the testing procedure is quite simple. You can check our video that goes over everything from install to taking your first measurement with REW here: http://gikacoustics.com/room-eq-wizard-tutorial/

[kinku]

Just a stupid question, what exactly the calibration do to the plots in REW? iF THE GRAPHS ARE NOT ACCURATE how we can use it with an room correction device like equalizer?
Are you guys suggesting that the room measurements plots are useful for comparison with each other than using the frequency with an elctronic correction system?
I  have attached the plot I got from my analysis.Any suggestions?

[JohnR]

The calibration removes inaccuracies due to the test equipment itself. REW has two, the soundcard calibration and the mic calibration. The first one you do yourself, the second one provided by the mic manufacturer or you can send the mic to be calibrated. (The rest of your electronics are not typically calibrated, although they could be if you were finicky enough.)

So, with the test equipment calibrated, you are then measuring what you want to measure, and not measuring the test equipment.

What the measurement means, and what you do with that information, is another story. The Acoustical Measurement standard is a good guide worth working through:

http://www.acousticfrontiers.com/whats-new/2011/10/13/acoustic-measurement-standards-for-stereo-listening-rooms-pu.html

[Alex Reynolds]

Just a stupid question, what exactly the calibration do to the plots in REW? iF THE GRAPHS ARE NOT ACCURATE how we can use it with an room correction device like equalizer?
Are you guys suggesting that the room measurements plots are useful for comparison with each other than using the frequency with an elctronic correction system?
I  have attached the plot I got from my analysis.Any suggestions?

Yes, exactly.
I wouldn't suggest using the measurements to provide EQ curves and such without calibration, but when you're comparing two measurements made with the same equipment, it will work fine. Also, many calibration files and the like won't really change how the measurements pick up reflections and decay, which are all important parts of the measurement as well instead of just the frequency response.

[JohnR]

Um, the calibration WILL affect how measurements and decay etc are picked up.

I fail to see how it could not...

Anyway..........

[Ethan Winer]

the calibration WILL affect how measurements and decay etc are picked up. I fail to see how it could not.

Reflections and ringing are both relative measurements. So the ratio of reflections versus direct sound, and the rate of decay for a waterfall plot, will be the same even if the absolute frequency response is off by a dB or 2.

--Ethan

[JohnR]

Reflections and ringing are both relative measurements

Only sometimes. An ETC or impulse response is not "relative".

And, consider the cost of calibration - $15 to $30 over getting an uncalibrated mic. You and Alex would happily sell thousands of $$ of acoustic treatment to anyone reading this thread yet advise that it's not worth spending $15-30 extra to get a calibrated microphone?? Sorry but that makes absolutely no sense whatsoever to me.

[Ethan Winer]

An ETC or impulse response is not "relative".

Sure it is. It shows the relative dB level of the direct versus reflected sounds. Now, there are frequency components to reflections, but that's mostly relative too. I say "mostly" because an omni mic's response might be different off-axis. However, that's a different issue, and knowing (calibrating) the on-axis response doesn't tell what happens off-axis unless that is specifically measured too.

You and Alex would happily sell thousands of $$ of acoustic treatment to anyone reading this thread yet advise that it's not worth spending $15-30 extra to get a calibrated microphone?

I never advise against more accuracy! It's just that for most applications - in my opinion! - it's farther down the list of importance. An investment in acoustic treatment always yields a positive result. An investment in knowing your microphone's precise response is more a matter of personal satisfaction. If you look at the microphone comparison article I linked above, the difference between the Earthworks "reference" and the cheapest Nady is only a few dB between 30 Hz and 20 KHz. For most of the audible range they match within 1 dB. That said, I spent a lot of money on my DPA 4090 "measuring" microphone! It cost about 20 times what I could have spent for similar results, but I need the accuracy more than most because I do this for a living.

--Ethan

[JohnR]

If you look at the microphone comparison article I linked above, the difference between the Earthworks "reference" and the cheapest Nady is only a few dB between 30 Hz and 20 KHz. For most of the audible range they match within 1 dB.

You can't generalize from a sample size of one. The graph below shows measured responses from ECM8000s measured by Herb at Cross-Spectrum Labs. That was posted a few years ago, recently he says they've gotten worse, to the point where he's stopped selling them at all.



http://www.cross-spectrum.com/weblog/2009/07/

[Ethan Winer]

You can't generalize from a sample size of one. The graph below shows measured responses from ECM8000s measured by Herb at Cross-Spectrum Labs.

Well, I can't argue with that! Seriously. If that's really representative of ECM8000 microphone variance, it's pretty poor. I haven't seen nearly that much change, but I haven't seen as many mics as Herb.

(But the measuring options mentioned are still relative. <ducking, running, grinning>)

--Ethan

[JohnR]

(But the measuring options mentioned are still relative. <ducking, running, grinning>)

I had to think about this... on the face of it, I would want to disagree, as after all you cannot actually claim that the time domain is independent of the frequency domain (multiplication in the frequency domain == convolution in the time domain). However, if looking at a time scale of tens of ms, an impulse response that long does imply a pretty wild response in the frequency domain...  so in practice it would take a pretty wild mic frequency response to make any difference at all when looked at on that time scale.

Long and short, in the end I accept what you're saying there. What do you think?

Thanks

[Ethan Winer]

What do you think?

I think we are now both right.

Seriously, even with a totally screwed up microphone response, the room's relative decay time at a given frequency falls off at the same rate. This applies to both modal ringing decays and ETC decays of individual reflections. Now, reflections have a response of their own, so a mic error will skew that. But the rate of decay shouldn't be affected.

--Ethan