Room EQ Measuring Equipment

[alanB]

It seems my questions were starting to hijack the sub-woofer discussion below. 

As suggested I've started a separate thread on the subject although I've only just learnt about the subject.

My requirement is that I would like to be able to quantify improvements to my listening room (and maybe others) after adding sound treatments such as absorbers etc.  Clearly I dont want to spend a lot of cash on this, because that would make the project a non starter.

I understand that to do this I need:
1) Freeware Room EQ software, which I have now downloaded http://www.hometheatershack.com/roomeq/
2) A USB Sound card (bit worried about the cost of this)
3) An SPL meter (it seems the Radio Sahck meter is relatively inexpensive.

Once I have this I understand that I would get the software to generate a sweep signal, which I play through my system via the sound card and RCA intelinks connected to my amp.
The SPL meter records the response at my listening position and I can then compare before and after plots to evaluate any changes made to the room.

Any other tips and hints would be appreciated.

[Timber_MG]

Often times the frequency response of cheap electret mics below 50Hz is unpredictable (and some have significant production variations). A calibration is the best answer to this question, but that is a problem here in South Africa where the only calibration facilities cater for mega-buck externally polarized microphones (and those generally only require a measurement point per octave or so and thus one doesn't get the calibration curve one needs for consumer type mics)

A couple of  forum members use the Behringer ECM-8000 with various software packages, but even there we have problems and we will likely go for a calibration exercise in the near future as both sub and HF are in doubt.

[GearSlave]

OK, I have elaboratd about my setup in fragmented fashion all over the place, but I'll outline it here for completeness.

My main goal was having a platform that can be used to measure various things, not just room modes. So, I did go all out when purchasing hardware, so it'll maybe exlude it from your list Alan, but may be of use to others. Anyway, I wanted the ability to measure speakers, amplifiers and other bits of audio hardware. So, this setup needed to be quite universal and semi-portable as well.

Hardware:
- M-Audi Firewire 410 Audio Interface. This can also supply phantom power to the mic
- ECM8000 measurement Mic
- Breakout box for speaker measurement with all the loopback capabilities for the measurement reference
- Breakout box for audio hardware measurements with attenuators

Software:
- Sound Easy V14 - Measuring speakers and Rooms
- RMAA Pro for doing distortion measurements

With the rig above I'm pretty much set to measure anything audio related and with the software/hardware I'm pretty much guaranteed of making distortion measurements down to around -110dB.

Just a note on measurement software. I understand that a lot of the freebie stuff uses swept sinewaves to perform measurements. This is the easy way, but is prone to contamination from adjacent signal sources. A better way is to use MLS based systems. This limits contamination of other signal sources affecting the measurements.

Like I said before, this configuration may not apply to you Alan, as it is pretty expensive. But, I noted it here for reference. Also, If you just want to make casual measurements on your room, you welcome to borrow it and this offer stands for anybody else in the JHB are.

Cheers

Gert

[alanB]

Thanks guys

As I understand it I would be looking for two types of improvements in the graphs:

1) A "flattening" of the peaks and valleys
2) A decrease in reverberation times

Do I need a highly calibrated mike/spl meter for that? 

Also as an aside what does EQ stand for?  I've worked SPL stands for Sound Pressure Level.  If I'm going to throwing these terms around I better know what they mean 

[Vaughan]

Also as an aside what does EQ stand for?  I've worked SPL stands for Sound Pressure Level.  If I'm going to throwing these terms around I better know what they mean

EQ stands or equalization. In simple terms equalization is the process of boosting or cutting the level of certain audio frequencies or tones compared to other frequencies.

Regards,

[Timber_MG]

Here the results of a German company that calibrates these mics against a known reference. Scroll down for the graphs.

http://diylautsprecher.de/index.php?option=com_content&task=view&id=94&Itemid=35

Especially low frequency measurement is not always without its problems as many devices have filters of their own in the measurement chain. The better designed systems have a reference input (even voltage dividers of the amp signal feeding the DUT="device under test"). One also often requires a sufficient signal to noise ratio in the measurement or averaging (not the same as smoothing) to provide one with reliable results.

[Timber_MG]

And here another link off that page outlining results from over 200 mic calibrations (including cheap Skype phone type capsules)

http://hifi-selbstbau.de/index.php?option=com_content&task=view&id=138&Itemid=1

I think I should start a seperate thread on calibration as there was some interrest from Gert and Gerhard to calibrate our ECM8000s. I had mine calibrated over 16 points (with very good results) but some recent measurements have cast some doubts on the long term stability of the results and we are looking at getting a suitable mic for calibration purposes.

[Vaughan]

Often times the frequency response of cheap electret mics below 50Hz is unpredictable (and some have significant production variations).

The frequency response inaccuracies down low can be corrected with a compensation file. There is a thread comparing a calibrated Behringer ECM 8000 versus a cheapie Radioshack analog model down to 20 Hz (with correction applied of course) and the results are very interesting (about a 1 dB variation at worst). I'll try to dig up the thread a bit later.

Regards,

[Timber_MG]

The linked pages contain statistical analyses of accuracy of a couple of measurement mics (including the Panasonic capsules, ECM8000 and some cheap stand-mics) against a source calibrated against a LAB grade calibrated microphone.

An ECM8000 is not a reference, but it does very well for most crossover-design tasks faced by loudspeaker designers and when an individual calibration is performed it allows for some fairly accurate measurements.

The issue with a general correction curve is that it may well be representative for a given batch of mics (or even only a single mic), but one needs to be wary of the statistical variance in microphones to arrive at an uncertainty of measurement, a factor that can be reduced drastically by verifying (calibrating) the response against a known reference.

[Vaughan]

An ECM8000 is not a reference, but it does very well for most crossover-design tasks faces by loudspeaker designers and when an individual calibration is performed it allows for some accurate measurements.

I own one.  But I understand. For those who want reliable results you don't need super expensive industry-grade equipment.

Even simple SPL meters that cost R500.00 (like the RS meter) with the correct filters applied can give very accurate results within the intended bandwidth of the design. High frequency measurements will be a problem however but then it is a cheap meter.

The Radio shack models do have slight variances from model to model (like 1-2 dB's) within the low frequency range which won't pose too much of an issue in my opinion.

Regards,

[Shonver]

I own one. 

Me, too. But I'm kinda frustrated by the ever-looming uncertainty factor. I want an Earthworks M30!!

[Timber_MG]

I own one.  But I understand. For those who want reliable results you don't need super expensive industry-grade equipment.

Even simple SPL meters that cost R500.00 (like the RS meter) with the correct filters applied can give very accurate results within the intended bandwidth of the design. High frequency measurements will be a problem however but then it is a cheap meter.

The Radio shack models do have slight variances from model to model (like 1-2 dB's) within the low frequency range which won't pose too much of an issue in my opinion.

No need for a very expensive mic for measurement, just a calibration of an inexpensive mic. For many uses such devices are adequate, but they do have their quirks (even the Panasonic capsules such as the old colony sound labs modified as per S.Linkwitz for DLSSA iirc). Becasue here in SA there are no calibration services who cater for DIYers I may soon offer such a service in collaboration with Gert and Gerhard.

I have personally had some problems with cheaper mics. I remember an audio install I did in the Uk where the system response didn't measure as expected. Hence we decided to hire a B&K 1/4" capsule for a week end (hire costing well more than a ECM-8000 in itself) and ended up with a difference on the order of magnitude of ~5dB @16Hz.

Gerhard and I did a difference measurement on our two ECM8000s and one of his panasonic capsules and came up with some very interesting results (and they are unfortunately not in the order of 1-2dB). Funny enough I am looking at a 5" driver to perform the calibration measurement for the LF from 10Hz and up.

[Shonver]

Martin

I'd like to also see a low-noise mod for ECM8000. Perhaps a SL + new front end. That should be killer!

[Timber_MG]

The mod was for the panasonic capsules iirc (this is going back 10 years or more), the name Mitey Mike rung a bell and google brought up this internet history gem:

http://www.speakerbuilder.net/web_files/Articles/diymic/diymicmain.htm

Gerhard uses Rod Elliott's pre-amp for such a capsule with good results.

Modding the ECM8000 will likely be difficult and I wager not really worth anybody's while. If you really want a low noise floor, unfortunately the better capsules in the more expensive mics (especially with larger diameter capsules) might be worth a look.

[Shonver]

My first mic also has the SL mod. It used the Jason Neal preamp, which I'm sure is similar to the Mighty Mike. Only problem I really had with it was the battery running down between tests (naturally I often forgot to switch it off!).