Odd treble response graphs

**hifisponge** · **01-07-09, 06:51 PM**

I recently finished full-range measurements on my 5 channel HT and I decided to take some near-field measurements. The graphs produced all show a rather steep roll-off above roughly 14KHz. I expect to see a gradual roll-off in the treble at my listening seat 12 feet away, but all of the measurements below were taken just 3 feet from the left speaker. My speakers use a Scan Speak silk dome tweeter that is supposed to have response out to at least 25KHz.

Is there a setting in REW that would limit my high-frequency reading? My starting frequency is 20Hz and the ending frequency is 21,999.

All of the measurements above were taken with the mic tip pointed up (as I believe it should be), but I also tried pointing the mic at the tweeter.

I'm not sure what to make of that graph other than there is still a sharp drop off above 14KHz.

It could be a problem with the speakers, but I'm trying to rule out testing errors first.

Thanks,

- Tim

**brucek** · **01-07-09, 08:52 PM**

What mic and calibration file are you using?................

brucek

**hifisponge** · **01-07-09, 09:00 PM**

I'm using the DBX RTA-M mic.

The cal file looks like this:
10.00 -5.64
10.29 -5.24
10.58 -4.88
10.88 -4.56
11.19 -4.27
11.52 -4.04
11.84 -3.87
12.18 -3.72
12.53 -3.58
12.89 -3.45
13.26 -3.33
13.64 -3.21
14.03 -3.10
14.43 -2.99
14.84 -2.89
15.27 -2.79
15.70 -2.69
16.15 -2.59
16.62 -2.49
17.09 -2.38
17.58 -2.27
18.08 -2.15
18.60 -2.02
19.13 -1.89
19.68 -1.74
20.24 -1.62
20.82 -1.54
21.42 -1.48
22.03 -1.44
22.66 -1.39
23.31 -1.35
23.98 -1.30
24.66 -1.25
25.37 -1.21
26.09 -1.16
26.84 -1.11
27.61 -1.07
28.40 -1.02
29.21 -0.97
30.04 -0.92
30.90 -0.88
31.79 -0.83
32.70 -0.78
33.63 -0.74
34.59 -0.69
35.58 -0.64
36.60 -0.60
37.65 -0.55
38.73 -0.50
39.83 -0.45
40.97 -0.41
42.15 -0.36
43.35 -0.31
44.59 -0.27
45.87 -0.22
47.18 -0.17
48.53 -0.13
49.92 -0.08
51.35 -0.03
52.81 0.02
54.33 0.06
55.88 0.09
57.48 0.11
59.12 0.12
60.81 0.11
62.55 0.10
64.34 0.09
66.18 0.07
68.08 0.06
70.03 0.04
72.03 0.03
74.09 0.02
76.21 0.01
78.39 0.01
80.63 0.01
82.94 0.01
85.31 0.01
87.75 0.01
90.26 0.00
92.84 0.00
95.50 0.00
98.23 -0.01
101.04 -0.02
103.93 -0.02
106.91 -0.03
109.96 -0.04
113.11 -0.04
116.35 -0.04
119.67 -0.04
123.10 -0.02
126.62 -0.01
130.24 -0.01
133.97 0.00
137.80 0.01
141.74 0.02
145.80 0.03
149.97 0.04
154.26 0.05
158.67 0.05
163.21 0.06
167.88 0.06
172.68 0.06
177.62 0.06
182.71 0.05
187.93 0.05
193.31 0.05
198.84 0.05
204.53 0.06
210.38 0.06
216.40 0.07
222.59 0.07
228.96 0.07
235.51 0.07
242.24 0.07
249.17 0.07
256.30 0.07
263.63 0.07
271.18 0.06
278.94 0.06
286.91 0.07
295.12 0.07
303.57 0.08
312.25 0.08
321.18 0.08
330.37 0.09
339.82 0.08
349.55 0.08
359.55 0.09
369.83 0.09
380.41 0.08
391.29 0.08
402.49 0.08
414.00 0.08
425.85 0.08
438.03 0.08
450.56 0.07
463.45 0.07
476.71 0.07
490.35 0.07
504.38 0.07
518.80 0.07
533.65 0.06
548.91 0.06
564.62 0.05
580.77 0.06
597.38 0.05
614.47 0.05
632.05 0.05
650.14 0.04
668.73 0.05
687.87 0.05
707.54 0.04
727.79 0.04
748.61 0.04
770.02 0.03
792.05 0.03
814.71 0.03
838.02 0.03
861.99 0.03
886.65 0.02
912.02 0.02
938.11 0.01
964.95 0.01
992.55 0.00
1020.90 -0.01
1050.20 -0.01
1080.20 -0.02
1111.10 -0.03
1142.90 -0.04
1175.60 -0.05
1209.20 -0.06
1243.80 -0.06
1279.40 -0.07
1316.00 -0.08
1353.60 -0.09
1392.40 -0.10
1432.20 -0.11
1473.20 -0.13
1515.30 -0.14
1558.70 -0.16
1603.30 -0.18
1649.10 -0.20
1696.30 -0.21
1744.80 -0.23
1794.80 -0.24
1846.10 -0.25
1898.90 -0.26
1953.20 -0.26
2009.10 -0.27
2066.60 -0.27
2125.70 -0.27
2186.50 -0.27
2249.10 -0.27
2313.40 -0.27
2379.60 -0.27
2447.70 -0.27
2517.70 -0.27
2589.70 -0.26
2663.80 -0.25
2740.00 -0.22
2818.40 -0.18
2899.00 -0.13
2982.00 -0.09
3067.30 -0.04
3155.00 -0.01
3245.30 0.01
3338.10 0.02
3433.60 0.02
3531.90 0.02
3632.90 0.04
3736.80 0.08
3843.80 0.15
3953.70 0.22
4066.80 0.30
4183.20 0.36
4302.80 0.42
4425.90 0.48
4552.60 0.57
4682.80 0.68
4816.80 0.77
4954.60 0.80
5096.30 0.79
5242.10 0.77
5392.10 0.78
5546.30 0.81
5705.00 0.80
5868.20 0.72
6036.10 0.59
6208.80 0.48
6386.40 0.45
6569.10 0.53
6757.00 0.66
6950.30 0.71
7149.20 0.68
7353.70 0.68
7564.10 0.80
7780.50 0.96
8003.10 1.01
8232.00 0.97
8467.50 0.97
8709.80 1.03
8958.90 1.03
9215.20 1.00
9478.90 1.00
9750.00 0.99
10029.00 0.93
10316.00 0.99
10611.00 1.08
10915.00 1.00
11227.00 1.00
11548.00 1.12
11878.00 1.12
12218.00 1.10
12568.00 1.19
12927.00 1.22
13297.00 1.21
13677.00 1.26
14069.00 1.32
14471.00 1.36
14885.00 1.37
15311.00 1.40
15749.00 1.40
16200.00 1.42
16663.00 1.39
17140.00 1.36
17630.00 1.34
18135.00 1.29
18653.00 1.25
19187.00 1.22
19736.00 1.23
20301.00 1.20
20881.00 1.27
21479.00 1.29
22093.00 1.31
22725.00 1.29
23375.00 1.21
24044.00 1.05
24732.00 0.83
25439.00 0.57
26167.00 0.27
26916.00 -0.05
27686.00 -0.45
28478.00 -0.89
29293.00 -1.37
30131.00 -1.94
30993.00 -2.55
31879.00 -3.19
32791.00 -3.84
33729.00 -4.49
34694.00 -5.14
35687.00 -5.82
36708.00 -6.55
37758.00 -7.32
38838.00 -8.16
39949.00 -9.03
41092.00 -9.86
42268.00 -10.62
43477.00 -11.28
44721.00 -12.00
46000.00 -12.75

Here's the spec sheet:
http://www.mediafire.com/?w0yjzyzjmmn

**hifisponge** · **01-08-09, 02:32 AM**

BTW, and this may have nothing to do with my original question, but what are the "replay buffer" and "record buffer" for?

**JohnM** · **01-08-09, 04:21 AM**

There are no REW settings that limit bandwidth, the primary factors are the soundcard and the microphone. Looks like you are on a Mac so the soundcard runs at 44.1kHz, which sets 22.05kHz as the upper limit. Your soundcard cal will show what the soundcard is doing at the upper end and saving the cal file will compensate for it.

Re the buffers, scroll down a little in the help panel on the soundcard settings, all explained there.

**hifisponge** · **01-08-09, 01:36 PM**

Thanks John. What you've said makes sense, and I have created a cal file for the sound card, so I'm not sure what's going on with my HF measurements. Is there anything I can do with the length of the sampling window in REW that may change the measurement results?

**JohnM** · **01-08-09, 02:12 PM**

Altering the impulse response window only affects the low frequency extension of the measurement, it doesn't contribute to any HF roll-off. Without knowing the speaker's crossover response its difficult to know where the roll-off may originate, worth measuring the other speaker to see if it is the same. Probably all fairly academic mind, unless you are a teenager

**hifisponge** · **01-08-09, 02:50 PM**

At 40 years of age, and hearing just barely up to 17K, yes this is mainly academic. I just want to know why I'm measuring HF roll-off at 14K considering the tweeter in my speakers is supposed to have flat response out to 30KHz.

I have five speakers that all use this same tweeter (a Scan Speak 2905/9500) and all of them show the same roll-off in the HF.

http://www.tymphany.com/files/produc...905_950000.pdf

Next up is to have my mic professionally calibrated I guess.

**brucek** · **01-08-09, 06:00 PM**

Quote:

Next up is to have my mic professionally calibrated I guess

Yeah, I would agree....

You may want to consider if the calibration lab accounts for baffle reflection. Read this thread on that subject.

brucek

**hifisponge** · **01-08-09, 07:16 PM**

Bruce -

It's not allowing me to access the link you posted.

- Tim

PS - I was planning to use the West Caldwell Labs for the mic calibration.

**brucek** · **01-08-09, 08:51 PM**

Quote:

It's not allowing me to access the link you posted.

Sorry, I mistakenly referenced a post in the moderators section.

Anyway, we do recommend West Caldwell labs as the place to go... Be sure that you use the mic in a vertical position when measuring. The calibration will be the most accurate that way..

brucek

**hifisponge** · **01-08-09, 09:29 PM**

Quote:

brucek wrote:

Sorry, I mistakenly referenced a post in the moderators section.

Anyway, we do recommend West Caldwell labs as the place to go... Be sure that you use the mic in a vertical position when measuring. The calibration will be the most accurate that way..

brucek

I have to say that you have piqued my interest with your baffle reflection link. What was the jist?

**brucek** · **01-08-09, 10:53 PM**

Well, the jist was that we had some discussion regarding the method of calibration that WCCL was using (pressure testing) and whether they compensated for baffle interaction of the microphone which would cause treble anomalies. WCCL does an excellent job and is very professional, but the concern was whether at high frequencies there might be inaccuracies because baffle step compensation wasn't considered. A basic explaination of baffle step with regard to speakers can be found here. It applies to microphones just as readily as speakers, so the small 0.5" baffle of a measurement mic could begin to have a problem at upper frequencies.

I'm no expert in this area, so someone else would have to explain it deeper for you....

brucek

**hifisponge** · **01-08-09, 11:30 PM**

Bruce -

That jist is good enough for me.

Back onto the thread topic, here's some more data to consider.

Below is a reading of a small Monitor Audio Radius speaker I use in a bedroom system. This is using the same test equipment that I used to measure the speakers shown in the top graph in this thread.

Here's Home Theater Mag's measurement of that same Monitor Audio speaker.

So I am getting response up to 20KHz from my equipment, but it does look like it may not be flat, since the spike at 20Khz in the HTMag graph is higher than the one in mine. Thoughts on the differences between the two?

**brucek** · **01-09-09, 08:22 AM**

Can we see a graph your soundcard calibration file?

Specifically show us a blow up of the 10Khz to upper range...

What microphone preamp are you using and did you include it in the soundcard calibration routine when testing? Is so, did you include the mic preamp or just the line-in?

brucek

**hifisponge** · **01-09-09, 02:05 PM**

Quote:

brucek wrote:

Can we see a graph your soundcard calibration file?

Specifically show us a blow up of the 10Khz to upper range...

What microphone preamp are you using and did you include it in the soundcard calibration routine when testing? Is so, did you include the mic preamp or just the line-in?

brucek

I will post the soundcard cal file later tonight, as I don't have access to the program at the moment, but I can tell you from memory that the FR plot for the cal file was flat to about 18-19K and then rolled off sharply from there.

I'm using the M-Audio Mobile pre and while I am using the cal file for that soundcard/mic pre, I created the cal file by going line-out to line-in rather than from line-out to mic-in. Could that be a potential problem?

If I recalibrate the soundcard/pre using the mic-in, should I engage the phantom power button?

**brucek** · **01-09-09, 02:52 PM**

Quote:

...... created the cal file by going line-out to line-in rather than from line-out to mic-in. Could that be a potential problem?

If I recalibrate the soundcard/pre using the mic-in, should I engage the phantom power button?

No, the line-out to line-in will be fine. If the soundcard cal file drops off quickly at 18K or is jagged, then that's as high as you should measure. If it tails off smoothly and slowly, then you can go higher. (read this thread through on mic preamp calibration methods).

I'd like to see the graph though.

Really though, when you start to get serious about measurements above 15KHz in speakers, you should get a professional calibration done on the microphone. The generic files we supply are fine for sub woofers and general mid range work, but you're trying to get up to 20KHz without a properly calibrated mic....

brucek

**hifisponge** · **01-09-09, 10:36 PM**

Bruce -

Per your request, here are the soundcard cal graphs. The cal file shows -2.28dB at 20KHz.

**brucek** · **01-11-09, 07:47 AM**

The soundcard cal files look good, so that isn't contributing to the problem.

Quote:

If I recalibrate the soundcard/pre using the mic-in, should I engage the phantom power button?

No, you don't engage the phantom power button. It is power for a microphone circuitry.
With a high quality card such as you have, I can tell you that the mic preamp will have a good response and won't be part of the problem. You can include it in the soundcard cal routine if you want, but I don't think it will be the source of the trouble.

The only thing left is the microphone calibration.

brucek

**hifisponge** · **01-11-09, 02:44 PM**

Bruce -

Thanks for the confirmation that the sound card is functionin properly.

I'm making arrangements to have the mic cal'd.

I found another thread here in which someone tested the ECM8000 (which I think is very similar to my DBX) and there was roll-off in the high treble of his graphs too, though not as early as the roll-off in my readings. It surprises me that either of these mics would roll-off early when the generic cal files show lifted response from about 10KHz on up.

**brucek** · **01-11-09, 03:01 PM**

Quote:

It surprises me that either of these mics would roll-off early when the generic cal files show lifted response from about 10KHz on up.

Please explain? I don't see that in the generic files? The cal files all show negative values as you get to the extremes of bandwidth...

brucek

**hifisponge** · **01-11-09, 05:48 PM**

This is a close-up of the top-end of my mic meter cal graph.

According to this file, the response of the mic rises by roughly 1.5dB from 3KHz to above 20KHz, no?

And I see no compensation at all in the cal file for a sharp roll-off above 14KHz like I'm getting in my speaker readings. Could my mic be that far off of the generic cal?

Advertisement