Subwoofer EQ test cds (1 Viewer)

[Kevin C Brown]

Hey Brian- Here's something that might mess you up a little:

If I add 10% to 90, I get 99. But if I subtract 10% from 99, I get 89.1, not 90. The error is always less than 1.0 dB or so though. See what I'm getting at?

 

[PaulT]

KCB,

I would expect the numbers you got.

10% of 90 is 9. >>>> 90 + 9 = 99
10% of 99 is 9.9 >>>>> 99 - 9.9 = 89.1

It's all relative ;

 

[Brian L]

Not a clue

I am OK with your math, but am not seeing what that has to do with the effort to create a compensated set of test tones.

I think I see what your saying WRT Goldwave....you can not set the level of the wav in dbFs...you have an arbitrary scale of 0 to 1.0, right?

In any event, although I will have inherent errors of .5 to .3 depending on frequency (since I can not do fractional levels), I suspect that the meter itself has enough sample to sample variance, not to mention the likely hood that I will not be able to place the meter in the exact same place from session to session would mean that what I will have in these tones at the end of the day will probably fit under the category of "close enough for my purposes".

BGL

 

[Kevin Deacon]

Can someone confirm the 1/3 octave tones from 5hz up to 250hz. Thanks for your help

I've got:

5
6.25
8
10
12.5
16
20
25
31.5
40
50
63
Is there one or more in between 63 and 125
125
Is there one or more in between 125 and 250
250

 

[Brian L]

1/3rd Octave between 80 and 800 are:

80
100
125
160
200
250
315
400
500
630
800

If you are planning to do sub work, you need better than 1/3 octave...most call for minimum of 1/6. So, to give you a start on that, 1/6 octave from 20 to 80 are:

20
22.4
25
28
31.5
35.5
40
45
50
56
63
71
80

BGL

 

[Kevin Deacon]

I guess I should have been more clear. I am doing sub work and using the Behringer Feedback Destroyer, which is a 1/3 octave parametric eq. Can't do 1/6 octave.

Thank you very much for filling in the blanks on the 1/3 octave steps.

 

[Brian L]

Uhhhhh no, the BFD is not 1/3 ocatve. Its base frequency centers are 1/3rd octave, but those can be changed to cover a much tighter range with the Fine control. Take a look at the chart on page 20 of the manual, which shows all possible frequency centers..

IIRC the examples at snapbug are tweaked for 1/6 octave, IIRC.

BGL

 

[Kevin Deacon]

Uhhhhh, sorry Brian. It's been a while since I messed with this. I guess I need to recreate my test cd with 1/6 octave tones. Thanks for clearing this up for me.

 

[Brian L]

No problem at all.....FWIW, I think the BFD is God's gift to those of us that obsess over our Subs and getting the best possible in room bass response.

BGL

 

[BruceD]

Just to add a couple of insights to this discussion:

1) The BFD supports a "Q" (bandwidth narrowness) of 1/60th of an octave, not the 1/3 octave stated above.

2) When measuring and adjusting for bass frequencies, do you take the Fletchur-Munson effect into account. i.e. a relatively higher absolute SPL level is required for bass vs. higher frequencies to sound like they are at same level (the reason for "C" wieghted curves).

For this reason, I would not trust many of the RS SPL meter correction tables, because many of them correct the "C" weighting to "U" weighting (unweighted).

I would trust the RS SPL meter correction curve published by ETF5 software, but that correction curve is designed for use with the MLS-based test tones in the ETF5 software, so I don't how useful it will be for sine-wave test tones. Also, I don't think the old correction file (2000 vintage?) applied corrections to any frequencies below about 400Hz.

 

[Kevin Deacon]

Brian, when I first turned on my newly built IB sub, it sounded like crap. I thought I made a mistake somewhere, but after a quick calibration my fears subsided. Without the BFD it was unlistenable in my room. Now having put base traps (bags of insulation stacked floor to ceiling) in the front corners the reverb boominess is gone and it sounds great.

Bruce, I think many of us use the published correction tables for the RS SPL meter and hope that we're close enough. I had read somewhere that if we calibrated without using the corrections we would have a nice "house curve". It's really hard for me to tell what the correct settings for the sub should be, flat vs curve, as they just sound a little different. I think my brain compensates if something isn't quite right. Take care.

 

[Brian L]

Here is a link to the corrections that I found in our own SW archive.

http://www.hometheaterforum.com/htfo...hack+SPL+Meter

I noted in another thread that these factors are set to apply only to the test tones used to derive them (filtered pink noise, IIRC). That said, it does say that the lowest frequencies were in fact measured with sines, so we do have that to go with.

But, as Bruce pointed out, they may or may not be accurate with sines across the board.

Of course, the question then becomes, How inaccurate are they?

I don't know about the house curve thing....I know that Guy Kuo suggests setting the sub with Avia to 83 dB vs. 85 dB for the mains to account for inaccuracies with the RS meter, but even going flat 85 dB at all channels would not really give you what I think a house curve provides, which is a gradual upward slope from somewhere around 100 Hz to about 30 Hz.

Having said that, after EQ'ing for as flat as possible response from my sub, I then rebalance its level with my center. At that point, I know what reference is, but I am not hesitant to dial in some additional cut or boost to suit me.

BGL

 

[Ryan Tsang]

I thought of that when I was calibrating my sub. But then I thought I'll just fine tune the level to my taste after. I want to know at least what flat is like and leave it up to the sound engineers to record what ever they want.

 

[DaveHo]

I downloaded the NCH Tone Generator. Anyone have a suggestion on how long of a duration one should use for each tone? Don't want to overheat the voice coil running sines through for an extended period of time.

-Dave

 

[Brian L]

I currently have my TTG generated tones at 30 seconds, but in practice that is way too long, and I get a stable meter reading much sooner. I usually advance to the next track as soon as I get a good reading.

I am thinking between 5 and 10 seocnds is more than enough for the reading to lock in with the RS meter, and when I burn my new disc with pre-compensated tones, I will probably go with 10 seconds.

BGL

 

[Wayne A. Pflughaupt]

Bhagi,
That method will get you a very poor house curve, as the meter only starts deviating drastically below 40Hz or so. As Brian mentioned, a good house curve usually starts rising at around 100Hz and levels at 32Hz. Rely in the RS meter’s “free” curve and the rise will continue upwards, which is probably what you don’t want.

Regards,
Wayne A. Pflughaupt

 

[BruceD]

To know what a flat bass curve sounds like, you typically still need to measure the levels with a "C" weighted curve (the "C" setting on the RS SPL meter).

If instead you measure with a "U" weighted curve (corrected values), you won't be adjusting the bass to flat, far from it actually.

 

[Kevin C Brown]

Brian and Paul- Actually two thoughts came to mind,
that "pre-compensating" is actually slightly different than "post". Plus, that if the RS adjustments values are a definite dB value, then somehow it should be referenced to the tone value. I.e., if I play a 40 Hz tone at 75 dB, then the RS adjustment value is "x". But "x" will probably be different if I play that same 40 Hz tone at 85 dB. ??

Bruce- Actually, if you read through the ETF info, the reason why there is no compensation below 400 Hz for the RS meter, is that he says the RS meter is "quite accurate" at low freqs. But ... when using ETF, you are simply using it as a microphone passthrough to the PC. I.e., what the meter is actually reading doesn't matter. It just passes the actual sound back to the sound card where the PC does the measuring of the volume. I.e., the adjustment values for ETF are completely different for a different reason than what we use when we use a RS meter manually. If that makes sense.

At one point, I had two sets of adjustment values for the RS meter. One was a simple calibration for C weighting, and the other was from a dude who actually calibrated his meter with a known standard. Someday I need to dig up both sets and compare again. (On my PC at home...)

 

[PaulT]

They may be refering to the microphone itself rather than the meter as it infers in the later part of the info.
Programs like TrueRTA apply the corrections for the RS Meter (if that is what you select for your input microphone), so they may be doing something different than the 'pass through' that ETF does.

I've got a good list of corrections collected from various websites and this forum, consisiting of 1/3, 1/6, 1/12th Octave, plus 10-80Hz in 1 Hz increments.

Edit - posted below

DaveHO - I also burn test tones at 10 second times, you could go to 15 but I find that 10 has been good enough.

 

[PaulT]

What the heck - here goes:

Radio Shack SPL Meter Corrections from Various Websites (1/3rd Octave, (include -xx) 1/6th Octave))

10.00 Hz+20.5 db
-11.20 Hz+18.0 db
12.50 Hz+16.5 db
-14.00 Hz+14.0 db
16.00 Hz+11.5 db
-18.00 Hz +9.5 db
20.00 Hz +7.5 db
-22.00 Hz +6.5 db
25.00 Hz +5.0 db
-28.00 Hz +4.0 db
31.50 Hz +3.0 db
-36.00 Hz +2.5 db
40.00 Hz +2.5 db
-45.00 Hz +2.0 db
50.00 Hz +1.5 db
-56.00 Hz +1.5 db
63.00 Hz +1.5 db
-71.00 Hz +1.5 db
80.00 Hz +1.5 db
-89.00 Hz +1.5 db
100.00 Hz +2.0 db


111.00 Hz +1.0 db
125.00 Hz +0.5 db
142.50 Hz +0.5 db
160.00 Hz -0.5 db
180.00 Hz -0.5 db
200.00 Hz -0.5 db
250.00 Hz +0.5 db
315.00 Hz -0.5 db
400.00 Hz 0.0 db
500.00 Hz -0.5 db
630.00 Hz 0.0 db
800.00 Hz 0.0 db
1.00 kHz 0.0 db
1.25 kHz 0.0 db
1.60 kHz -0.5 db
2.00 kHz -1.5 db
2.50 kHz -1.5 db
3.15 kHz -1.5 db
4.00 kHz -2.0 db
5.00 kHz -2.0 db
6.30 kHz -2.0 db
8.00 kHz -2.0 db
10.00 kHz -1.0 db
12.50 kHz +0.5 db
16.00 kHz 0.0 db
20.00 kHz +1.0 db

Radio Shack SPL Meter Corrections from Anthony Gomez’s PEQ Program (1/12th Octave)
Interpolated with 4th order Polynomial (confirmed by PM from Anthony)


FreqRS Calib.
10.0+20.0
10.6+19.0
11.2+18.0
11.9+17.0
12.6+16.5
13.3+15.0
14.1+14.0
15.0+13.0
15.8+11.5
16.8+10.5
17.8+10.0
18.8 +8.5
20.0 +7.5
21.1 +6.5
22.4 +6.0
23.7 +5.5
25.1 +5.0
26.6 +4.5
28.2 +4.0
29.9 +3.5
31.6 +3.0
33.5 +2.85
35.5 +2.75
37.6 +2.75
39.8 +2.5
42.2 +2.4
44.7 +2.0
47.3 +1.75
50.1 +1.5
53.1 +1.5
56.2 +1.5
59.6 +1.5
63.1 +1.5
66.8 +1.5
70.8 +1.5
75.0 +1.5
79.4 +1.5
84.1 +1.5
89.1 +1.75
94.4 +1.8
100.0 +2.0

HTF Jeremy Stockwell Interpolations 10Hz – 80Hz in 1Hz Steps

10+20.0
11+18.25
12+16.5
13+15.25
14+14.0
15+12.75
16+11.5
17+10.5
18+9.5
19+8.5
20+7.5
21+7.0
22+6.5
23+6.0
24+5.5
25+5.0
26+4.66
27+4.33
28+4.0
29+3.66
30+3.33
31+3.0
32+3.0
33+3.0
34+3.0
35+3.0
36+3.0
37+2.5
38+2.5
39+2.5
40+2.5
41+2.5
42+2.5
43+2.5
44+2.5
45+2.0
46+2.0
47+2.0
48+2.0
49+1.5
50+1.5
51+1.5
52+1.5
53+1.5
54+1.5
55+1.5
56+1.5
57+1.5
58+1.5
59+1.5
60+1.5
61+1.5
62+1.5
63+1.5
64+1.5
65+1.5
66+1.5
67+1.5
68+1.5
69+1.5
70+1.5
71+1.5
72+1.5
73+1.5
74+1.5
75+1.5
76+1.5
77+1.5
78+1.5
79+1.5
80+1.5

As you can see, depending on how the interpolation is done from the 'normal' published values, you can find frequencies that could be .25 - .5dB different anyways.

For those trying to EQ with a BFD, burning single Hz tones can find your exact centre frequency that you want to dial in as your filter.