Why the +6dB gain boost to a subwoofer when level matching speakers?

[kexodusc]

I'm pretty sure I've read this before and use to know, but I'm even more sure that I've long since forgotten.

When using an SPL meter to level match speakers, we always here the sub should be set +6 dB (give or take) higher than the speakers. Can someone please explain why?

Thanks.

[Sir Terrence the Terrible]

I'm pretty sure I've read this before and use to know, but I'm even more sure that I've long since forgotten.

When using an SPL meter to level match speakers, we always here the sub should be set +6 dB (give or take) higher than the speakers. Can someone please explain why?

Thanks.

There are a couple of factors that make the 3-6db boost necessary. The first is that our hearing is less sensitive at lower frequencies than at midrange frequencies. They are also less sensitive to phase and amplitude at higher frequencies(above 8khz). So to make all frequencies flat to the ears, you have to increase amplitude as you drop in frequency.

The second reason lies in the fact that the subwoofer only handles two ocataves of information(20-80hz) and our mains handle the rest (80-20khz). If you were to splice the subs with the mains output equally, the system would be out of balance to the ears. The balance would tip to the speakers that handle the most frequencies, at the most sensitive area of our hearing. To counter this, and to create balance between the sub and mains, you increase the output of the speaker that produces the fewest frequencies, in the less sensitive areas of our hearing.

I hope this is easy for you to understand so you can explain it to me.

[kexodusc]

Well, I pretty much guessed the first reason had something to do with it...the 2nd one is wee bit baffling to me...but I guess that's why so many people have recommended I have a downward sloping curve from 20 Hz to 80 Hz when I get the BFD all figured out...

Thanks again, Sir T.

[Sir Terrence the Terrible]

Well, I pretty much guessed the first reason had something to do with it...the 2nd one is wee bit baffling to me...but I guess that's why so many people have recommended I have a downward sloping curve from 20 Hz to 80 Hz when I get the BFD all figured out...

Thanks again, Sir T.

Let's try and de-baffle you. Here is what the sub handles[......] Here is what the mains handle[...................................] As you can see, the mains handle three of four times as many frequencies as the sub. Not only do they handle more information, but that information is going to be louder to the ears. So in order for there to be audible balance you must turn up the channel with the least amount of information, in the least sensitive area that your ears can hear. That would be the sub.

I hope you have been sucessfully de-baffled.

[Woochifer]

Here's yet another simple reason -- the Radio Shack SPL meter that most of us on this board use decreases in sensitivity as you go further down into the bass range. Since you're talking about an inaccuracy that typically ranges from 1.5 db around 80 Hz up to 4.0+ db below 30 Hz, a wideband test tone would likely require an adjustment of at least 2.0 db just to account for the SPL meter sensitivity.

[kexodusc]

Let's try and de-baffle you. Here is what the sub handles[......] Here is what the mains handle[...................................] As you can see, the mains handle three of four times as many frequencies as the sub. Not only do they handle more information, but that information is going to be louder to the ears. So in order for there to be audible balance you must turn up the channel with the least amount of information, in the least sensitive area that your ears can hear. That would be the sub.

I hope you have been sucessfully de-baffled.

I think ya made it worse!
Let's re-baffle me and take it a step further with these cool graphics to show who handles what:

Sub: 2 octaves or less: [........]
Midwoofers: 5 octaves or less: [....................]
Tweeters: 3 octaves or less: [............]
(give or take)
Clearly the midrange is the most important area (though is useless by itself). Yet when I build a crossover in a speaker I make darn sure the midwoofer and tweeter are level matched as close as I can get them...By this logic I should be replacing 5 ohm resistors with 1 ohm resistors or so on the xo to drive the bejeezus out of the tweeters too. Unless, by George, you're gonna tell me we're more sensitive to the highest 2 octaves (5kHz-20kHz or so)...I don't think I'd believe that - too many people are deaf above 13kHz.

Still, I understand the fact that we are least sensitve to sub 100 Hz frequencies, so maybe that's all I'll go by for now.

[toenail]

I found that the +6db theory varies with how flat the sub curve is. When I had peaks at 46hz and 63hz in my sub curve and tried +6db or so I was still audibly missing really low frequencies during payback. When I altered my LF output curve with the BFD to slope slightly downward from 20hz to 80hz for HT use I didn't need nearly the +6db gain to account for the sub. If I recall I ended up setting it at 0bd or very close. This is weak for TV watching with analog signal but excellent for DVD/digital output.

[Sir Terrence the Terrible]

I found that the +6db theory varies with how flat the sub curve is. When I had peaks at 46hz and 63hz in my sub curve and tried +6db or so I was still audibly missing really low frequencies during payback. When I altered my LF output curve with the BFD to slope slightly downward from 20hz to 80hz for HT use I didn't need nearly the +6db gain to account for the sub. If I recall I ended up setting it at 0bd or very close. This is weak for TV watching with analog signal but excellent for DVD/digital output.

Actually the theory doesn't vary, it';s everyone individual taste that varies. Actually if you have a downward slope house curve from 20-80hz, then your boost is in the house curve already. A downward curve from 20hz to 80hz has to have some boost at 20hz just to start a curve downward.

My room is pretty unusual because it is a open floor plan, and bass peaks where no higher than +6db before equalization. After eq my subs were +6db at 20hz, and flat at 60hz, and down 3db at 80hz. It is still set +5db over my center channel volume.

While the boost is good for getting a flat curve to the ears, this still boils down to how much bass you really like. +6db boost would represent the high side of the boost, some people may only want +3db, others maybe +2db. It's all about your taste.

[Richard Greene]

When using an SPL meter to level match speakers, we always here the sub should be set +6 dB (give or take) higher than the speakers. Can someone please explain why?

Our ears don't hear bass well unless the bass is so loud (95-100dB) that the SPL could permanently damage our hearing.

The C-Weighting used on most sound meters assumes a gradual roll-off of hearing ability that reaches -6.2dB at 20 Hz. -- that means the average person listening to a 20Hz. tone needs +6.2dB more SPL than a 1000Hz. tone for both tones to sound equally loud.

If you listen at an average SPL of 75 to 80dB, bass frequencies that measure flat using the C-weighting will sound flat (in reality "flat" with C-Weighting is really +6.2dB at 20Hz. when using an unweighted SPL, sometimes called U-Weighting).

The C-Weighting is a "House Curve" for the general population.

For really loud sound averaging 85dB or more the U-Weighting may more accurately represent what you hear.

In plain English this means that after you find a subwoofer level that sounds balanced to your ears with most movies or music, the average SPL of the subwoofer measured with pink noise is likely to be +3dB to +6dB louder than the main speakers.
+6dB is more likely if you prefer below average volume (perhaps 75dB and lower)
+3dB is more likely if you prefer above average volume (perhaps 85dB.and higher)
Most people prefer average SPL's between 75 and 85dB

If you do not use parametric EQ or lots of bass traps to control the bass peaks (typically +3dB to +6dB) caused by standing waves, you'll find some bass notes will sound too loud and some too weak even though the average SPL (measured with pink noise that does not excite standing waves) seems "right".

If you eliminate bass peaks with a parametric EQ, you will be able to listen to the bass louder than before without interfering with the other frequencies. Loud bass peaks at specific frequencies can be very distracting -- EQ of the specific frequencies that are too loud makes more sense than turning down the overall level of the subwoofer.

[kexodusc]

Thanks for the thorough explanation, Dr.Greene, if I may ask a few quick follow up questions:

Can we safely assume that this gradual roll-off in response at lower frequencies isn't compensated for by the engineers and mixers in the studio when the sound track is being made?

And, how well, or accurately, do the microphones in the original recording process pick up the low frequency effects? Is there any additional loss in response contributed at this end?

I just find it odd that in a world of international standards and format-forums that these experts would knowingly allow Dolby Digital and DTS formats to deliver low-frequency signals that would be heard at such a significantly lower volume than the rest of the spectrum when all channels are SPL matched. But then again, perhaps it's easier to let the user customize the response levels to their preferences.

[Sir Terrence the Terrible]

Thanks for the thorough explanation, Dr.Greene, if I may ask a few quick follow up questions:

I am not Dr. Green, but I am his bald assistant. Maybe I can be of some service.

Can we safely assume that this gradual roll-off in response at lower frequencies isn't compensated for by the engineers and mixers in the studio when the sound track is being made?

You cannot safely assume this. The standards we use to calibrate our hometheaters are derived from the same standards used to calibrate 5.1 mixing studios. Since our hearing insensitivies in the low bass range is well understood, it is compensated for just like we do in our hometheaters.

And, how well, or accurately, do the microphones in the original recording process pick up the low frequency effects? Is there any additional loss in response contributed at this end?

It depends on if you are referring to the LFE(which is not recorded through a mike, but constructed using various signal processing devices), or low bass in recordings. Microphones used for recording(depending on their type) are either flat to 20hz(PZM and other large diaphram mikes) or they may roll off like a speaker would.

I

just find it odd that in a world of international standards and format-forums that these experts would knowingly allow Dolby Digital and DTS formats to deliver low-frequency signals that would be heard at such a significantly lower volume than the rest of the spectrum when all channels are SPL matched. But then again, perhaps it's easier to let the user customize the response levels to their preferences.

Keep in mind that the LFE content in the film formats is boosted 10db during playback. So they are not at a low level at all.

[toenail]

From Sir Terrence- "Actually the theory doesn't vary, it';s everyone individual taste that varies. Actually if you have a downward slope house curve from 20-80hz, then your boost is in the house curve already. A downward curve from 20hz to 80hz has to have some boost at 20hz just to start a curve downward."

As usual Sir Terrence, your observation is 100% accurate. Whether I removed signal from towards the 80hz end of the curve and turned up the amplifier to level match the midbass, or boosted the 20hz end of the curve and left the amplifier level alone, the net result is still a "boost" at the 20hz end of the curve and increased overall output levels.

Kex- Got that BFD in yet?

[Geoffcin]

I'm pretty sure I've read this before and use to know, but I'm even more sure that I've long since forgotten.

When using an SPL meter to level match speakers, we always here the sub should be set +6 dB (give or take) higher than the speakers. Can someone please explain why?

Thanks.

I run the mains full up, and my subs filter hinges @ 40hz. With the sub set +3db @ 40hz I seem to get the best (transparent) match. Responce of the sub is down 3db (0) @ 20hz.....hey that's almost as good as an AN standmount!

[drseid]

Responce of the sub is down 3db (0) @ 20hz.....hey that's almost as good as an AN standmount!

Too funny... ;-)

---Dave

[kexodusc]

I Responce of the sub is down 3db (0) @ 20hz.....hey that's almost as good as an AN standmount!

...Almost

Toenail: Don't have the BFD yet...PE is holding it for me while I wait for some back-ordered parts to arrive...This will save me $40 in brokerage fees at the border (lousy shippers) but it's taking forever.
Not to worry, I don't have my sub built yet either, so I couldn't do much with it anyway...
Maybe next week.

[toenail]

Kex- Border? I didn't know you were a Hab from up nord. How 'bout the hockey season, eh?