Measuring Frequency Response of your setup

[Emjay]

Hi again

 

First off, apologies if this, too, has already been covered.

 

I did a quick search and didn't see anything on the first 3 pages...

 

Do any of you have any experience measuring the frequency response of your setup?

 

Preferably something that doesn't cost $x,000s as this is purely idle curiosity

 

I downloaded 2 (free) aps from the Google Play store which purport to offer such measurements, but there's no way in the world that they're accurate.

 

First of all, I'm sure the mic on a mobile phone isn't calibrated all that carefully for the full frequency range, given that it's primary job is to capture the human voice in spoken conversation.

 

Secondly, just look at the screen-caps I got from them - there's no way in the world that my RF-83s (with no sub) are producing single-digit frequencies while playing "MTV Dance" channel on cable (Dolby Digital encoding).

 

Alright, so my (carpeted) floor is chip-board, so may provide some resonance (and may be why I have entirely done away with a sub), but there's still no way these are right

 

I got a second opinion, so to speak, from another ap which actually allows a 'pinch-to-zoom' which changes the scale the graph is shown in, but, if anything, it's even worse!

 

So, can anyone suggest something which is relatively cost-effective? I really don't want to spend all that much, so it may well be that my curiosity will have to go unsatisfied (i.e. I really don't even want to spend 3 figures, let alone 4!)

 

These are the screen-caps from the aps:

[attachment=97471:Screenshot_2014-10-18-13-11-19.png][attachment=97472:Screenshot_2014-10-18-13-12-16.png][attachment=97473:Screenshot_2014-10-18-13-17-50.png]

[derrickdj1]

Depending on the avr or pre pro a frequency response graph may appear in the GUI screen.  You most likely are asking this question and that is not an option.  Tool such as REW can be downloaded for free and will do the job.  This one is very popular and very useful.  Real Traps test tone can also be downloaded and a graph can be plotted.

[Emjay]

Tool such as REW can be downloaded for free and will do the job.  This one is very popular and very useful.  Real Traps test tone can also be downloaded and a graph can be plotted.

 

Thanks for the info on REW, it looks pretty comprehensive, and the price is right

 

What about the mic, though? I would think that a fairly professional (read: expensive) microphone would be needed to accurately capture the full audio range?

 

Actually, having RTFM, it appears that this is actually designed to measure the accuracy of reproduction of the frequency?

 

My curiosity was more along the lines of "what frequencies are currently being reproduced?"

Edited October 18, 2014 by Emjay

[Ski Bum]

The recommended mic is less than a hundred bucks.  Spectrum Labs, I think, check hometheatershack to verify.

[WMcD]

I was surprised to find this very frugal microphone.  http://www.parts-express.com/dayton-audio-imm-6-calibrated-measurement-microphone-for-iphone-ipad-tablet-and-android--390-810

 

From what I see, measurement microphones are not totally accurate as made.  They can be compared to a known flat microphone and then the correction data is supplied.

 

The other part of this is how and where to measure your speaker.  You can just put the microphone at your listening spot and gather your data.  You'll get the response at that spot which is likely not what you'll get at another spot in the room.  The effect of the room is influential.

 

Anther way is to take you speaker amp and measuring equipment outdoors.  it is good if you have a driveway and a quiet neighborhood.

 

Put the speaker box upside down (on a towel to prevent damage) so the tweeter is near the pavement.  Tip the box somewhat so the face of the box faces the microphone, which you place on the ground about two meters in front of the box.  You may have to play with tilt and distance.  Then run your curve.

 

This is pretty much a poor-man's anechoic (no-echo) chamber equivalent.  This minimizes ground reflections since the tweeter is close to the ground, and the microphone is on the ground.

 

People also will put the speaker up on a pole or big ladder so the ground reflections are reduced.  Or they will put the speaker in a hole in the ground so the front panel is flush with the ground surface and there are therefore no reflections.

 

Let me also suggest that you report what you find.  All here are interested.

 

WMcD

[garyrc]

When you buy a mic, be sure to get a calibrated one.  You then load the calibration into REW. 

 

See also the long, long thread on AVS forum titled something like "REW with a USB mic."   Actually, see this first, since it is really comprehensive.

 

As WMcD said, moving the mic just a bit will make a difference in the result.  No speaker and room combo is flat, although a lot can be done with bass traps, absorbers and diffusers.  Audyssey uses 8 microphone positions, instead of just one, then rather than just averaging them, uses "fuzzy logic" (a good thing) to assess the speaker/room response.  They do start with the main listening position (one head position and width), though.

[WMcD]

Of course I agree with Gary in principle.  I don't know what goes on with fuzzy logic.

 

There is the issue that maybe be want to measure the speaker alone, and we can do that outdoors.

 

But then we really want to measure the speaker in the room.  We can do that at the listening spot, but small changes in location make a big difference. Or you can average.  Or use this fuzzy logic idea -- and I don't know what fuzzy logic tells us.

 

No matter what you do, there is always some other way.

 

It is like the man and the boy and the donkey which is a tale from antiquity.

 

http://aesopfables.com/cgi/aesop1.cgi?sel&TheMantheBoyandtheDonkey  In an alternate version the man and the boy simply carry the donkey and he remains in good health -- I like that better.

 

There is a revealing test you can run about the effect of the room.  I did it and was impressed.  Start by playing a 200 Hz tone into one of your speakers if you have the capability to do so with a computer or signal generator.  Then walk around the room.  You will find that in some spots, the tone just about completely goes away as if someone shut down the amp and speaker.  These are the nodes of standing waves.  But if you change the freq to 300 Hz, the nodes are in other locations. 

 

We obviate these nodes by averaging measurements from many locations, or use 1/3rd octave averaging; or warble tones, or band limited noise as test signals.    

 

The resulting data looks smoother but I'm not quite sure what it tells us.

 

WMcD

[Emjay]

hmmmm....

 

I fear that this way madness lies!

 

Let's say I'm leaping around the front yard with my speaker buried in a hole(?), and I use REW to produce a 200Hz test tone which is not measured at 200Hz.

 

Is it:

a) The microphone?

The speaker?

c) The speaker wire?

d) The amplifier?

e) The DAC?

f) The transport?

g) The PC?

h) Any combination of any of the above?

 

It might be best not to start down this slippery-slope!

[garyrc]

It might be best not to start down this slippery-slope!

 

Indeed.

 

Take comfort in the fact that, while frequency response is important, it is probably not as important as low distortion, wide dynamic range, good transient response, etc.  At least I'm inclined to think so.  So did PWK, I think.