• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

How do speaker manufacturers calculate the -6dB frequency cutoff?

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,109
Likes
8,424
Location
NYC
Here's something I've long wondered but somehow haven't found a clear answer for.... Where do speaker manufacturers get there -6dB point from? As in-6dB relative to what? 1K? 200Hz? The average dB within a predetermined range?

I've seen a few different methods, but was wondering if there was a standard methodology or if someone had insight into how some manufacturers go about establishing it.
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,523
Likes
37,056
That is a good question. You'd think usually vs 1 khz with electronics. With speakers it is a bit less clear.

Manufacturers play games. I know way back in the days of cassette machines a common spec was something like 50hz to 15,000 hz +/- 3 db. When in fact cassette was pretty flat in the mid-band and it was really 50hz-15,000 hz +0db/-6 db which is actually correct either way, but one is a very different picture of what you'd expect vs the other.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,202
Likes
16,982
Location
Riverview FL
Apply a sweep tone of any steady amplitude and note at what frequency the output level changes by 6db.

Filters may be more commonly named for their 3dB cutoff, so, an example:

1572905071178.png


Vpeak may be any voltage. The 3dB cutoff is at 0.707 x Vpeak.

A filter named for its 6dB cutoff would be at 0.5Vpeak (or half the input voltage)
 
OP
N

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,109
Likes
8,424
Location
NYC
That is a good question. You'd think usually vs 1 khz with electronics. With speakers it is a bit less clear.

Manufacturers play games. I know way back in the days of cassette machines a common spec was something like 50hz to 15,000 hz +/- 3 db. When in fact cassette was pretty flat in the mid-band and it was really 50hz-15,000 hz +0db/-6 db which is actually correct either way, but one is a very different picture of what you'd expect vs the other.

Yeah, jsometimes it seems like manufacturer's anechoic -6dB figures are highly inflated. And that's not to mention manufacturers that only give you in-room figures (and don't even clarify at that).

Apply a sweep tone of any steady amplitude and note at what point the output changes by 6db.

Filters may be more commonly named for their 3dB cutoff, so, an example:

View attachment 37880

Thanks - it's pretty clear with filters and electronics, but I should have made clear in my original posts I was talking about speakers(I've now edited that for clarity). For example, a manufacturer could claim a much lower -6dB point if they're measuring it relative to the lowest dip in the rest of the frequency range rather than a peak or average.
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
Does "-6dB" not typically refer to only the low end of the spectrum?
i.e. it is a reference to the generally high-pass filter characteristics of a driver or system of drivers.
 

digitalfrost

Major Contributor
Joined
Jul 22, 2018
Messages
1,521
Likes
3,086
Location
Palatinate, Germany
When they design the enclosure, it's really easy to know the -6dB point of the woofer based on Thiele-Small parameters. I'd like to think honest manufacturers will state this value. Dishonest manufacturers might calculate some form of room gain on top.
 

leonroy

Member
Forum Donor
Joined
Feb 15, 2018
Messages
80
Likes
96
Location
London, UK
If two speakers are rated at say 40Hz at -6db how do you compare whether a speaker is delivering 40Hz at a higher or lower SPL than the other?

Also are +/-3b ratings measured at the speaker’s max SPL or at some sort of agreed standard dB?
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
If two speakers are rated at say 40Hz at -6db how do you compare whether a speaker is delivering 40Hz at a higher or lower SPL than the other?

You look at sensitivity and max power specs.

Also are +/-3b ratings measured at the speaker’s max SPL or at some sort of agreed standard dB?

It shouldn't matter as long as you don't hit compression or maximum extension limits.
 
OP
N

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,109
Likes
8,424
Location
NYC
When they design the enclosure, it's really easy to know the -6dB point of the woofer based on Thiele-Small parameters. I'd like to think honest manufacturers will state this value. Dishonest manufacturers might calculate some form of room gain on top.

It's become more common for manufacurers to give both anechoic and estimated in-room figures, which I think is the best approach. Anechoic figures tend to make people think they won't hear anything beyond the stated frequency, which is almost never the case in a regular room, while in room figures can obviously inflate a speaker's performance (especially with no standardization as to how they are calculated).

If two speakers are rated at say 40Hz at -6db how do you compare whether a speaker is delivering 40Hz at a higher or lower SPL than the other?

Also are +/-3b ratings measured at the speaker’s max SPL or at some sort of agreed standard dB?
In my experience, while this is rarely an issue for passive speakers, it is a significant one for active speakers.

Take the recent case where I tested teh Bowers and Wilkins formation Duo. B&W claims a -6dB point of 25Hz, which.... might actually be conservative. This is how I measured the speaker at its default EQ setting with neafield data spliced at 300hz and compensated for baffle step via simulation.

Duo-Horizontal.png


Keeping in mind potential error from my part - splicing nearfield is always kind of messy, especially on a speaker with permanent grilles - if you're measuring from 200Hz, you could say the -6dB point is 25Hz. From 3KHz, you could get away with claiming the -6dB point is 20.

However, those are measurements from 1m away at about 85dB. These in-room measurements I took show how from my listening position about 8-9 feet away (bottom graph is KEF R3 and can be ignored, this is an old image), compression starts to set in the bass just before 95dB. This is more than loud enough for regular listening, but does mean you likely won't get big dynamic peaks in the sub-bass.

Which brings up another point: While I think active speakers are the future, I do think manufacturers need to specify frequency ranges at at least two different SPL levels to indicate how bass compression is a factor.
 

digitalfrost

Major Contributor
Joined
Jul 22, 2018
Messages
1,521
Likes
3,086
Location
Palatinate, Germany
If two speakers are rated at say 40Hz at -6db how do you compare whether a speaker is delivering 40Hz at a higher or lower SPL than the other?

Also are +/-3b ratings measured at the speaker’s max SPL or at some sort of agreed standard dB?
-x dB should be always be relative to "full scale", whatever that may be. Does it really matter in this case? The speaker's efficiency rating in dB/W/m should tell you how loud it goes.
 
OP
N

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,109
Likes
8,424
Location
NYC
-6dB relative to the SPL rating, which typically is something like the average of 300Hz-3000Hz.

Bonus: -6dB and +/-3dB are pretty much the same thing as far as marketing is concerned for most companies.

Yeah, I really don't get some try to use +/- 3dB the same as -6dB unless the speaker has a perfect on-axis frequency response or listening window. Well, maybe with 1/1 octave smoothing...
 

Zion

New Member
Joined
Aug 3, 2020
Messages
1
Likes
0
It's actually quite simple, the -6dB point can be at any point, the point that a speakers starts loosing effectiveness, and is dictated by the Thielle Small parameters (https://en.wikipedia.org/wiki/Thiele/Small_parameters, https://www.epanorama.net/documents/audio/speaker_parameters.html)

Furthermore it doesn't matter at what SPL it is, the cut-off is always the same.
-6dB at 100dB 100Hz will be 94dB 150Hz, 88dB at 225Hz, 82dB at 337,5Hz etc.... (-6dB is -6dB per octave, and an octave is half the current Hz)
-6dB at 50dB 100Hz will be 44dB 150Hz, 38dB 225Hz, 32dB 337,5hz, etc...
-12dB at 100dB 100hz will be 88db 150Hz, 76dB 225Hz, 64dB 337.5Hz etc...

Also the SPL (Sound Pressure Level) is not rated in Hz, but in dB (sensitivity), so if you have a 50W set of the same quality and one is rated at 89dB, and the other at 92dB, the one with 92dB will be louder. (+3dB at the same input power).

In the end the input is more important than the speaker, i've ran 2KW through 1200W (P/A) speakers with ease, as long as you make sure you have a stable sinus, and there's no clipping (block-forming, where you're sinus becomes a square) then your speakers (if they're from any respectable manufacturer) will most likely survive. on the other hand, if you connect € 10.000,- speakers to a Schneider amplifier, you'll most likely blow your speakers.
 

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
In the end the input is more important than the speaker, i've ran 2KW through 1200W (P/A) speakers with ease, as long as you make sure you have a stable sinus, and there's no clipping (block-forming, where you're sinus becomes a square) then your speakers (if they're from any respectable manufacturer) will most likely survive. on the other hand, if you connect € 10.000,- speakers to a Schneider amplifier, you'll most likely blow your speakers.

Oh no, not that old one again. Suggested reading: Speaker Failure: The Truth

To quote:

8.3 - The 'Smaller Amps are More Likely to Damage Speakers than Bigger Ones' Myth

This naïve corollary to the clipping myth is also utterly wrong (see Figure 3). Average power over time is what heats and burns a voice coil - so a larger amp is always potentially more hazardous to a speaker. The popularity of 1000 watt plus per channel amplifiers relies on this complete nonsense being believed by a great many folk in audio.

It should hardly require any explanation - if you replace a smaller amp that is clipping with a bigger one that eliminates that clipping, then average power output must go up. This is shown clearly in Figure 3.

Likewise, we would hope that no-one would believe that a 50W amp driven hard into clipping (with a full-range signal) will damage a 500W speaker. Unfortunately, there appear to be people who do believe that is the case. In theory it might possible, but extremely unlikely. According to the myth (if taken to extremes) a 10W amplifier should be able to destroy any speaker ever made if it clips hard enough. The idea of 'headroom' as provided by a very high-power amp is fine, provided those operating the system know what they are doing.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
This should be nuanced a bit.

underpowered amps will not destroy woofers.
underpowered amps will not destroy tweeters.

Severely clipping underpowered amps (clipping is the key here) will not destroy woofers.
Severely clipping underpowered amps (clipping is the key here) can destroy (low power rated) tweeters. Not all tweeters have an equally low power rating.

overpowered amps can destroy low power rated speakers when driven far beyond the speakers max. power rating for a prolonged time period.

Clipping amps (in general) can potentially kill certain tweeters. This depends on how severe the clipping is and how often this happens.
In other words if the energy of the clipping+music content power exceeds that of the tweeter rating for a prolongued time.

Oscillations in amplifiers can kill both amps and tweeters.
Oscillation is rare but possible under certain conditions (exotic cables can sometimes be the cause)

DC and too much power can kill woofers.

Too much power kills drivers.
Too little power never kills drivers, only severe clipping could potentially kill some tweeters.

Tweeters usually have a MUCH lower power rating than woofers.
Midrange speakers usually sit inbetween.
 
Last edited:

Julf

Major Contributor
Forum Donor
Joined
Mar 1, 2016
Messages
3,004
Likes
3,998
Location
Amsterdam, The Netherlands
Clipping amps (in general) can potentially kill certain tweeters.

Sure. But if a 50 W amp clipping kills a tweeter, so does a 100 W amp not clipping, but playing heavily distorted electric guitar or synthesizer at full blast.

Too much power kills drivers.

Exactly.

only severe clipping could potentially kill some tweeters.

And only because it causes too much power to go through the tweeter.
 

Mnyb

Major Contributor
Forum Donor
Joined
Aug 14, 2019
Messages
2,641
Likes
3,614
Location
Sweden, Västerås
I

Which brings up another point: While I think active speakers are the future, I do think manufacturers need to specify frequency ranges at at least two different SPL levels to indicate how bass compression is a factor.

This +1 :) . I think speaker data is so disjointed ? should it not be tied to a piont of operation somehow for example max SPL and frequency response they are obviusly not valid at the same time . But thats not obvius to most people bying speakers , it sure sounds nice with a small two way with -6dB at 35hz and max spl at 110dB, but thats not real . ( I work a lot with electrical motors there most of the rated data is valid at the same time , ie its the same point of operation . rpm power and torque etc is "at the same time" )

Wish there was a metric to estimate "typical output" with "typical music" .

I wonder if not the actual possible bass output SPL of many speakers are lower than people expect ?

The measurment i want to see is something like you can listen to music at this average and peak level with almost no compression in any range ie some kind of "usable frequency range for music at x db" .

What I think , many byers (and therefore manufacturers) chases very low -6dB piont in the bass for no good reason , your small speakrs is going to be very ineficient and the SPL levels at this low tuning is almost impossible to hear .

Anectdote: I have my subwoofer xover at 60Hz, at moderate listening levels it can take me a while to realise that i forgot to turn on the sub :facepalm:Depends on the type of music ofcourse, so the fletcher munson curve seems to apply to me too :)

I do realise that this varies a lot with the type of music you use , i reckon that some acoustic orchestral music can go very low but the SPL levels is not that high . And experimental electronic music can be basically anything.

Are there any curves regarding typical frequency contents vs spl level for music genres ? With this speaker manufacturers could give estimates of what you could expect of thier product ? It may not be practical to build speakerss with 125dB output at 20Hz and 20kHz at the same time ?

My personal compromise is floorstander and sub , not small speaker and sub .
 
Top Bottom