Rational Acoustics



Dr. J
July 6th, 2011, 04:04 PM
I wasn't going to post this but I cannot find the discussion where I believe Arthur made the comment.

The conversation was about how measurement mics (even good ones) tend to have a bump in the upper frequencies that don't show up in the magnitude response. I want to say a 4kHz-ish bump???

Anyway I think it was Arthur (forgive me if it wasn't) stated that all you have to do to see if it is valid was to purposely point the measurement mic STRAIGHT up into the air from the original OnAX position. I guess this displays a different or possible flaw in the mic's response so you can take that into consideration.

I have a DBX mic and no calibration file or anything like that to go off of.

What made me think of this was -- I was measuring my Tops trying to find a good crossover point area using the Spectrograph. After following Arthur's advice on searching for the smoothest coverage edges (I still need practice with this) I noticed at 4k on the Spectrograph had a hint of orange and red in it. The Magnitude response was perfectly flat right there.

So here is another question on Spectrograph:

I have a perfectly FLAT magnitude response at 4k BUT my spectrograph shows some Hotness there. I don't remember what the phase trace demonstrated or even the coherence trace.

Is it reasonable to use the Spectrograph to further fine tune Hot spots? I do get feedback in the 4k area a lot of times & it is usually the first frequency to go.

Is this due to my measurement mic not being calibrated very well throughout this area?

Thanks for any ideas!

Kip Conner
July 10th, 2011, 07:55 PM
If there's no calibration file to know what the mic is actually doing then... well you know. You could create a poor man's calibration file but doing a simple mic compare to a mic that has a know calibration. Know anyone with an earthworks measurement mic? It was shipped to them with a piece of paper that shows the frequency response. You can apply the same live/reference principles to see what your RTA420 mic is actually doing.

Second fun experiment would be to make yourself a poor man's polar plot. You'll need a nice day to be outdoors, protractor, a meaurement mic and piece of this rope/twin about 8-10 feet in length. Take your string and attach it to the top rear of the box (this is real easy if it's a speaker on stick) with some gaff and tie the other end to the mic stand. Start with the mic on axis and call it zero. Now have someone keep the string taught and walk the mic in a semi circle slowly so you can save the trace every 10 degrees or so until you get to 180. This will tell you the polar response of any frequency in question.

Maybe it's not your frequency response per se, but the placement of the vocal mics?

I only throw this out there because your mic would have to have a dip in the 4kHz range and cause you to put more in the PA to get your described results. For the RTA 420 mics- the roll off is a little higher than 4k.

I'll try to get you a screen shot of an earthworks and some 420's that I own in the next day or so.

Ferrit37
July 11th, 2011, 06:50 AM
Hey Kip,
your fun experiment will measure the polar of the speaker and not the mic.
To measure the polar of the mic you need to rotate the mic on a fixed point.
Simple way is to get a lazy-susan, mark it's center point and position the mic capsule over this point, use your protractor to set 5 or 10 degree increments. :

Kip Conner
July 11th, 2011, 08:50 AM
Hey Kip,
your fun experiment will measure the polar of the speaker and not the mic.
To measure the polar of the mic you need to rotate the mic on a fixed point.
Simple way is to get a lazy-susan, mark it's center point and position the mic capsule over this point, use your protractor to set 5 or 10 degree increments. :

That was what I was intending if I wasn't clear, sorry about that!

There used to be a great free program in the late 90's for PC that allowed you to do this process and then it would combine the data and give you the polar plots for your speakers.

Anyone remember this? It was very basic but worked.

Dr. J
July 11th, 2011, 10:50 AM
Hey Kip,
your fun experiment will measure the polar of the speaker and not the mic.
To measure the polar of the mic you need to rotate the mic on a fixed point.
Simple way is to get a lazy-susan, mark it's center point and position the mic capsule over this point, use your protractor to set 5 or 10 degree increments. :

Hey Ferrit, Thanks for the input. I am not quite sure what you mean but I think you are saying to take a rotational device like the lazy susan (my wife won't mind me taking one from the kitchen for a few hours :D), find the center of it & place the mic straight down at the center point? From there rotate the device?

OR mic ONAX over the center point and rotate from there???

Any clarification for me would be great as I have never done this before. I also could just buy a nicer measurement mic with an established response...:D

There was a discussion out there in which I someone on the forum here stated that pointing the measurement mic straight up could reveal some flaws in the mic in the upper frequencies. I can't find the discussion or who actually stated it. I think it was Arthur but I am not sure.

Thanks again!

Kip Conner
July 11th, 2011, 05:54 PM
The idea would be to put the speaker on the lazy susan and leave the mic in a fixed position. As you rotate the speaker off axis you would save traces. What you should see in your screen shots is the high frequency roll off a little each time you make a small movement of the cabinet.

If the speaker you want to play with can be mounted a pole, even better. At that point you can just spin the speaker on the pole.

The reason I even brought this up was so you could add another dimension to your knowledge of your system. It's easy to repeat something you read in the specs section of your manual... if you we're lucky enough to get one. However, if it says your high freq drivers are 90x60degrees it's often not real clear what that means. 90 degress in the horizontal coverage and 60 in the vertical- but at what frequency?

Since we know that frequencies become more directional the higher they go it's important to know the design of the box and what might be leaking past the 180 degree mark. You can bet you'll have all sorts of stuff below 500Hz.

Why is this important? It will help you point your boxes in the right direction. Pointing them away from open microphones and away from the walls.

If you want to get real specific you can do the same test, but instead of the mic moving in the horizontal position, you can move the mic stand in the vertical position give you the answer to "Should I flip my boxes upside down because the horns are too high on the lip of this stage?"

Ferrit37
July 13th, 2011, 03:48 AM
Hey Dr J,
sorry for the late reply, I've been in class. You need to position the center of the capsule over the center of the turntable.
Keeping the speaker still and rotating the mic gives you the polar response of the mic (the original point of the this thread) :)
Placing the Speaker on the turntable and rotating it will give you the polar response of the speaker. This is a bit more accurate than the string technique as in that technique the mic is in a different space for each measurement. You have to be careful to keep the point of rotation at the apparent acoustic source center to avoid apex error.
Measuring polars in spectrograph mode can give you very interesting results, moving mics near boundaries can be an eye-opener too (thanks josh for this technique)
Another technique is to place the speaker on the floor and also the microphone (ground plane) then its easy to move the mic around in small increments ( remember to tilt the speaker so it focusses on the mic) you can get horizontal polars and if you place the speaker on its side you can then measure Verticals.
Placing two mics. a known reference and an unknown one, as close together as possible near a small, reasonably flat source and doing a transfer measurement between them can show you the difference between them. You could use this technique to check your mics. * I think there used to be a smaart app note on this somwhere? I check my Earthworks M30 (travel mics) back to a B&K ref mic (that never leaves home $$$) and against each other with this technique. ( I've also built IEC test baffles and calibration chambers, but I'm confirmed crazy) :)
Remember "always test the testers"

ps * it's a case study #7 on the resources page here

Dr. J
July 13th, 2011, 11:30 AM
Thank you Ferrit!

Arthur Skudra
July 13th, 2011, 05:01 PM
I wasn't going to post this but I cannot find the discussion where I believe Arthur made the comment.

The conversation was about how measurement mics (even good ones) tend to have a bump in the upper frequencies that don't show up in the magnitude response. I want to say a 4kHz-ish bump???

Anyway I think it was Arthur (forgive me if it wasn't) stated that all you have to do to see if it is valid was to purposely point the measurement mic STRAIGHT up into the air from the original OnAX position. I guess this displays a different or possible flaw in the mic's response so you can take that into consideration.

I have a DBX mic and no calibration file or anything like that to go off of.

What made me think of this was -- I was measuring my Tops trying to find a good crossover point area using the Spectrograph. After following Arthur's advice on searching for the smoothest coverage edges (I still need practice with this) I noticed at 4k on the Spectrograph had a hint of orange and red in it. The Magnitude response was perfectly flat right there.

So here is another question on Spectrograph:

I have a perfectly FLAT magnitude response at 4k BUT my spectrograph shows some Hotness there. I don't remember what the phase trace demonstrated or even the coherence trace.

Is it reasonable to use the Spectrograph to further fine tune Hot spots? I do get feedback in the 4k area a lot of times & it is usually the first frequency to go.

Is this due to my measurement mic not being calibrated very well throughout this area?

Thanks for any ideas!
I can't remember off hand where that conversation was, but I have found that the RTA-420 exhibits a flatter HF response if the source is 90 degrees from the capsule, in other words, I have my Earthworks M30 pointing straight at the source, but my RTA420's pointing away from the source 90 degrees. Do your own experiment using a known "good" mic as described by the other excellent posts above and you will see the effects of angle on the mic. Does it really matter? You decide. Normally I don't do much eq'ing by looking at the magnitude response in Smaart in the upper frequencies where these "cheap test mic anomalies" occur, unless it's an obvious problem that I can hear with my ear. Get a bunch of cheap test mics doing a live average, and those HF anomalies average out a bit.

The few DBX mics I have evaluated are pretty good for smooth response. As to the 4K bump you see on the spectrograph, could be just about anything including mic response anomalies. I do find the spectrograph really useful to investigate resonances.

Ferrit37
July 13th, 2011, 06:43 PM
Hey Guys,
I seem to remember that the bump in the response might be due to mic diameter, I think the B&K stuff mentioned this. there was also a Syn-Aud-Con paper on this. RH Campbell maybe?

Ferrit37
July 15th, 2011, 01:55 PM
Errrmmm....
Mic diameter, 4KHz bump? :o rubbish - a half inch is good to about 24KHz
There's a Syn-Aud-Con paper by Pat Brown 349

Arthur Skudra
July 15th, 2011, 03:03 PM
Hey Ferrit,

Maybe you were thinking of the information published in the old Genrad manual that Mr. Rayburn scanned and posted on his website?

See here:
http://www.soundfirst.com/GenRad.html

Look at "Sound Fields page 3" and "Microphones page 4" in particular.

Arthur

Ferrit37
July 15th, 2011, 06:51 PM
Arthur,
Brilliant that's the one!
this oldtimers disease is...............what was I saying? :)