This was measured with a RS analog meter with corrections applied (80 db scale C-weighted Slow). Anything greater than +/- 10 db I didnt change the scale. This occurs at 10KHz and above and down in the 25 - 40 Hz range. The speakers are two H3's and a Sunfire True Sub. 

Any comments would be sincerely appreciated.

Hard to tell, but I believe it shows that all the drivers in HIII and the sub (wow) are working.

I looked at the Rives site and it shows their 1/3 octave sources on their "II" version.  The x-axis looks to be set out out 1/3rd points.

IIRC, 1/3 octave noise is distinctly hissy.  1/3 octave warble tones sound like a Star Trek phasor. BTW.

The use of 1/3rd octave sources is a way of averaging out peaks and valleys in the measurement.  It is not really averaging out what is measured (the speaker and the room) but rather what is fed to the speaker -- which amounts to pretty much the same thing.

Your graph looks ragged at first.  But much of it fits in a plus or minus 3 dB window.  I'd say that is not too bad. 

The smoothing is supposed to remove (average out) the reflections from the room walls.  But I think you might still be measuring the effects of floor, ceiling,and wall bounces.

Why don't you tell us more about where the speaker and microphone were located. 

One technique for testing is to move the speaker outdoors and set it on a hard surface, like a driveway, away from buildings. The speaker box is placed upside down so the tweeter is near the ground.  Then put the microphone about 10 feet away and cant the speaker front to aim at the microphone.  Maybe more than you want to do. 

It is difficult to convince people just how much the room reflections affect measurements.  One time I used a single tone to a speaker (say 200 Hz).  At some places in the room, the tone could not be heard at all.  People have reported something similar with higher freqs (maybe 1000 Hz) and walking around the room with an RS meter.  Meter readings (more obvious if you have a moving needle type) vary widely with movements of one foot. 

WMcD

.

There is a sofa drawn to the left; the microphone "M" on tripod is located near it. The speakers are located on the right wall labeled "H3" on either side of the entertainment unit. The sub is in the top right corner with a diagonal orientation (labeled "S"). The room is 15 x 15'. Each grid is one foot. Let me know if you need anything else, thanks.

glennconti:

Let me know if you need anything else

So this dimension (the one blue)  is ~3 ft?  The mic is four feet from the wall?  What's the height of the ceiling?

 Hi QH,

Yes, the blue dimension is 3 feet. The mic was four feet from the wall and the ceiling is 9 feet tall.

--Glenn

This is the recomended sub location in the owner's manual. My top wall (horizontal) is an outside wall and the verticle wall in my house is a load bearing wall. The driver on the left as you face the sub is the active driver and the driver on the right is the passive driver.

As far as the chart being "ok", that's a matter or your personal tastes. Sub bass looks a little hot (you cut the graph off at 10 dB), but it's at the frequencies that need the most lift.

Using the link provided you can see:

1. Minor attenuation from the ceiling of the room at 9' above the sub.

2. Big cancellation from the wall behind the sofa affecting the mic. (try measuring closer or further from the wall and watch for this dip to coincide with the change in position)

3. Cancellation from the flank of the entertainment center and the wall...very likely affecting both the sub and the Heresy on that side of the room.

 http://www.padrick.net/LiveSound/CancellationMode.htm

Thanks for the input. I love the way my stereo sounds and it looks like the frequency response graph is not showing much of a problem with room acoustics. I did another frequency chart. This time I normalized the settings so the bass response was not getting clipped. Basically it showed, if you average out the peaks and valleys, that my subwoofer is set about 6 db "hotter" than the H3's. This gives me just about the right amount of bass to my taste. My bass tone control is set at 12:00 (as are my other tone controls) at the receiver and it's all good. Thanks again for looking at my chart and room layout. If anybody else wants to chime in feel free.

 After seeing Artto's post:

http://community.klipsch.com/forums/t/162245.aspx

I am wondering if I am missing something? I don't want to pop the bucks ($330) for a Behringer DEQ2496 (plus it looks like it uses only balanced inputs and outputs - which I don't have) but I am wondering if I would hear any difference by using a 20 band (10 Left and 10 Right) equalizer? I can pick one up cheap on craigslist. I can get one for less than $50.00. I may just get it and experiment around with it today. 

glennconti:

 After seeing Artto's post:

http://community.klipsch.com/forums/t/162245.aspx

I am wondering if I am missing something? I don't want to pop the bucks ($330) for a Behringer DEQ2496 (plus it looks like it uses only balanced inputs and outputs - which I don't have) but I am wondering if I would hear any difference by using a 20 band (10 Left and 10 Right) equalizer? I can pick one up cheap on craigslist. I can get one for less than $50.00. I may just get it and experiment around with it today. 

The Behringer DEQ2496 can be had for less than $330. The first one I bought locally at Gand Music so I could easily take it back if I didn't like it ($299). The second one I bought new (open box) on Ebay for $200 w/free shipping. http://www.gand.com/

Personally, I don't think a 10 band graphic EQ which is going to give you  the selectivity you need at low frequencies. At these low frequencies individual musical notes are only about 1Hz apart. The DEQ2496 has filters that go as sharp as 1/60 octave. And the time delay capability really helps make things sound more coherent. You can even adjust time delay by meters, feet and inches or milliseconds.

As far as the balanced connections go, you can use a cable that has regular RCA type connectors on one end and XLR on the other. I had some made in Mogami Neglex Quad 2534 in 3 foot lengths with Neutrik connectors, color coded from Upscale Music for $22.30 each. You can also get adaptors to do the same but I prefer as few breaks in the signal path as possible. And you may find that the adaptors are almost the same price as cables depending on the cable length you need.

 http://www.upscalemusic.com/categories.php?cat=19

 Artto,

Thanks for the advice. I dont want to spend much for the experiment. So I got a used 62 band (31 left and 31 right) equalizer on ebay. A Samson E62 for $56 plus freight. This 1/3 octive setup corresponds to my Rives Audio test tone cd. The Sampson E62 also has 1/4" inputs and outputs so I got a bag of 10 1/4" male to RCA female adapters for another $8 on Amazon. This setup is nowhere near what you have but, as per your suggestion, it will be much better than the 10 band EQ I was originally considering. If the experiment goes well I may get a DEQ2496 and sell the Samson E62 for what I paid for it. Thanks again for your input. With this forum's help I don't feel like I'm gropping around in the dark as much. 

artto:

At these low frequencies individual musical notes are only about 1Hz apart.

After I get the Samsom E62 setup and a flat frequenct response for the 1/3 octaves, I am going to retest the frequency response with new higher resolution low frequency test tones. 

This is a quote from Real Traps website which offers a free low frequency test tone cd:

"There are a lot of commercial test tone CDs available, but all of the CDs we've seen suffer from the same fatal problem: The tones are spaced at either 1/3 or 1/6 octave intervals, which is far too coarse to assess the low frequency response in rooms the size you'll find in most homes."  This agrees with what Artto says.

"Therefore, we supply only low frequency sine waves, in 1 Hz increments"

This will be interesting. However, they also don't recommend using correction factors for Radio Shack SPL Meters  So I am a little confused.

"Some test CDs "calibrate" the level of each tone to match the known inaccuracies of the Radio Shack SPL meter, but we didn't do that for two reasons. First, meter calibration offsets are not that useful because all Radio Shack meters are not the same. More to the point, low frequencies are usually accurate enough even with inexpensive microphones. We compared our Radio Shack SPL meter to our own expensive AKG calibrated microphone, and they were within 1 dB of each other below 600 Hz. The main point of these tones is to find the major peaks and nulls at low frequencies, and any inexpensive meter is fine for that, calibrated or not."

Should I correct for the C-weighted drop off or not? Any comments would be appreciated. Thanks.

 There is a free audio test CD, Bink Audio. Also NCH Tone Generator. They have a free trial period and has the actual individual musical notes as sine wave test tones ($59).

You should use the "C" weighting on an SPL meter, no compensation needed, thats the flatter response of the two (A or C). I don't use calibration files for the Radio Shack SPL meter because there is too much variation in their product. Most "omni" directional microphones produce a pretty flat response curve, it's the nature of the beast. It really boils down to how accurate you want to get and how much you want to know in how much detail (resolution).

Behringer and Dayton Audio makes an inexpensive omni microphone. With the Dayton you can download the actual calibration file for the specific mic you purchased using its serial number. The file can be read by any number of Room EQ software applications (it's just a basic text file) and automatically correct for deviations in the response of your specific mic. In most of these programs you can also take it one step further and calibrate the sound card too. But below 20Hz pretty much everything seems to get a little dicey and has to be taken with a grain of salt.

1/3 octave is IMHO pretty much useless for subwoofers. The frequencies I show SPL levels for on my charts are the actual musical tones and as you can see they are about 1Hz apart, 12 notes to an octave.

Reviews

the absolute sound

February 2003

If you have a subwoofer--and nowadays who doesn't--this BEHRINGER digital EQ device is an opportunity to make a remarkable and remarkably cheap sonic improvement in your system, be it in home theater or two-channel music. Think of it this way: The world is full of subwoofers that will produce deep, low-distortion bass. Thunder is no problem. What is a problem is low bass with smooth frequency response. Few subwoofer/room combinations are free of annoying peaks and dips--especially peaks, those infuriating one-note booms. [...]

But with BEHRINGER, you can always make it work out, and most systems need help in this area. [...]

I believe that every audiophile who does not own a digital EQ device ought to think seriously about buying the BEHRINGER just as an experiment. [...]

 

 

 

Looks like it's time to test my room frequency response in the lower frequencies and, if needed, save my money for a Behringer experiment.

The DEQ 2496 is very handy device...especially when you buy the ECM8000 microphone to go with it. The RTA function is very quick to setup...

 However, after a year or so of using manual EQ, I stepped up to subwoofer Auto EQ and time alignment in the form of a AS-EQ1 and haven't looked back since.

The AS-EQ1 is no longer available...but a similar product, the Antimode 8033 might just be the ticket for your sub. Very quick results. To do what it does in a matter of minutes can take hours of measuring, data analysis, and tweaking otherwise.