Home Theater Forum and Systems banner

1 - 8 of 8 Posts

·
Registered
Joined
·
34 Posts
Discussion Starter #1
I managed to get hold of a B&K 2270 for the weekend and have been seeing which of my budget microphones gives the most similar frequency response. Here's what I got: -



The audio interface was an RME Multiface. My other microphones were all over the place, but the built in microphone on the little Pocketrak surprised me. It's good to know that once I give the B&K back I'll have something reasonably reliable to use in assessing rooms.

I also made impulse responses using both REW and DIRAC, and they were pretty much identical - which is good to know, since I only have use of that copy of DIRAC for a few more weeks.

Satisfaction at last! Now to the pub.

:clap:
 

·
Premium Member
Joined
·
1,400 Posts
I'd suggest taking a few more measurements ( including a variation of distance, as well as changing the mic-capsule orientation, maybe 45 deg away from the target ) to give you an idea if the relationships ( between the 2 curves ) remain constant .

Note; if you turn one mic 90 degree away from the target / you must duplicate position & orientation with the other mic being compared .

<> EarlK



> You'll need to put in a lot of custom calibration points ( between 1K & 20K ) to make those 2 curves eventually matchup .
 

·
Registered
Joined
·
34 Posts
Discussion Starter #3
Yeah I think the directivity of the the Yamaha microphone is the main problem. There are two microphones on it (left and right). Would it be an idea to take two measurements - one from each microphone and then make an average of those two responses? Or even do that, then turn the unit by 90 degrees, and take two more measurements, then average all four?

In simpler language, if only a regular microphone is available, is it feasible to take multiple measurements - each with the microphone in a different 90 degree position (6 in total) and then to average those 6 measurements?
 

·
Premium Member
Joined
·
1,400 Posts
> One would want to measure using only a single capsule of a double mic system .

> Measuring with 2 capsules together introduces HF cancellations ( giving a binuaral effect ).
- Perhaps that's what's seen in your first posted measurement .


> I think your asking; if it's a valid approach ( when using a directional mic ) to take multi direction readings ( & then average them ) to simulate the captured response of a single omni-directional mic .

> Short answer ; I don't know .
- I believe there some validity to the method / though pratically it makes little sense when one compares the low cost of a decent omni to the huge hassle of doing multiple data captures ( to simulate omni from uni-directional ) .
- To simulate a spherical pickup I believe one should do at least 36 samples . :spend:

<> EarlK :sn:
 

·
Registered
Joined
·
34 Posts
Discussion Starter #5
Okay well maybe I should look at an omnidirectional microphone then. Are you talking about buying just an omnidirectional microphone, or a sound level meter with an omnidirectional microphone attached?

Can you recommend a few?
 

·
Premium Member
Joined
·
1,400 Posts
Can you recommend a few?

The standard recommendation ( from this site ) still holds true .

Buy a calbrated omni mic from Cross Spectrum Labs . They ship overseas .

I believe the Dayton model ( lately ) has the edge over the Behringer ( though with a custom calibration, it doesn't really matter ) .

I wouldn't bother with the Galaxy SPL meter . Just get a cheap Radio Shack meter ( or clone ) for doing straight-generic SPL measurements .

<> EarlK :sweat:

PS : This RS clone seems like it would do the job ( of establishing a ballpark SPL level ) .



I'd be very wary of using this SilverLine for aything more than just establishing a ballpark SPL level (ie; to be used when calibrating with a better mic ) .

Here's some interesting info about this Silverline model ;

clicky_in_1/20/2011 said:
I have recently purchased one of these, mine was labelled "Silverline" but they are sold under a variety of different labels. Mine was faulty when recieved, as it was very cheap and a mail order buy then I decided to repair it rather than return it. The problem it had was that on all ranges the needle would not drop below the -4 position although it was working as far as detecting sound was concerned. After re-soldering the badly made joints of the link wires on the PCB and cleaning the gold PCB switch contacts then it all appeared to be working well.

At this stage I decided to compare the circuit with the Radio shack version and while most of it is very similar the RS use a single BA301 op-amp and this new meter uses a C4558 dual op-amp and quite a few different component values, so NO it is not the same meter. It appears to be updated from the original circuit by having larger value coupling capacitors so it should be better at low frequencies than the old one.

I remove the microphone insert (9mm dia marked EPE) and checked it's frequency response to find that while it was reasonably flat it showed quite a drastic loss above about 7kHz, so that wasn't good.

I next checked the frequency response of the meter without a mic attached with a spectrum analyzer (with the meter set to C weighting) to find a very large dip in the response around 8kHz, again not good.

I cured the dip by putting a 47pf capacitor across resistor 2R10 (the 750k near the A-C switch) and although this caused a very slight loss in the response at low frequencies it was now virtually flat from 70Hz to 20kHz.

I have since changed the mic insert to a Panasonic WM61A which has a much flatter response than the original and all tests so far show that it is all now greatly improved.

Both of these modifications will affect the calibration so if you are planning to do the same then you either need a calibrator or some sort of very stable source of sound where you can make before and after comparisons that enable you to reset the meter to the correct level.

Be warned that if you you decide to start playing about with the component values on the board then some of the components originally fitted do not correspond with the board markings so the circuit has already been modified from the original design.

Otherwise, Yes it is a good meter and well worth the very low price. If you are trying to use it unmodified to
measure sound levels at different frequencies then it isn't so good but if you are just using it to compare different sound levels at the same frequency then you would not notice it's deficiencies.
Quote taken from here !
 

·
Registered
Joined
·
34 Posts
Discussion Starter #7
Thanks for all that, I think I will go for one of those omnidirectional microphones.

As far a an SPL goes, what's the actual use of having one in this context?
 

·
Premium Member
Joined
·
1,400 Posts
As far a an SPL goes, what's the actual use of having one in this context?
- Simply to give you a more realistic reference of where 75 - 80 db actually is when you're measuring .

- "Guessing" doesn't really cut it ( especially if you're asking others for their opinions on a FR curve ) .

<> EarlK
 
1 - 8 of 8 Posts
Top