Skip to main content

just a quick question to clear up some confusion that's really bugging me this morning!...

i've been gonig along for the past few years knowing/assuming that a 6db increase in level is considered "twice as loud" to the human hear. or, a6db decrease being half as soft, etc..

i'm in the middle of reading the Sound Reinforcement Handbook, and on page 172 the author says that a 10db increase is "twice as loud"... !

which one is it?



10db increase is "twice as loud"... which one is it?

Every time you double the power output, you are increasing the level by 3dB. Every time you decrease the output by half, you decreased the output by 3dB.

The +6db increase is only because the two speakers have coherent signals ... exactly the same signals.

otherwise it would be a +3dB increase.


audiowkstation Wed, 12/28/2005 - 07:10

The scientific community after peer review has "deemed" the level that appears "twice as loud" to be 10dB increase..or 10 times the actual power output.

Since decibels work on the log. scale, here is a brief rundown of how dB works.

Ref dB is ANY reference to a known factor. Relationship to a "reference" is know as the "difference to that reference. dB is meaningless unless you have a reference to base the new relationship to.

Every time you double the power output, you are increasing the level by 3dB. Every time you decrease the output by half, you decreased the output by 3dB.

Inverse square dictates (and measurements cannot always prove this due to room acoustics and other variables I will not get into here) that when you move a distance of double from the noise source, the resulting drop in decibels is 6dB. 6 more dB for each doubling.

Since we know this is not an absolute, based on acoustics, one must study acoustic engineering to understand the variables.


Science attempts to work with absolutes, we all know the human condition or subjective condition as being relative, not always absolute. Some may say "9dB" sounds twice as loud, someone else may say 7dB. The established reference is 10dB for twice as loud in perception, based on peer review and subjective comments from a cross section of human subjects in experimental lab conditions.

Certain absolutes are simply references we go by. In many examples, some of these absolutes can be shown to be perceived differently..based on "who is listening" and how the actual test is conducted.

The ears are easily fooled.

Relative perception is NOT a point of argument. No one agrees universally with what people claim or claim not of what they hear. For a textbook quiz, go with the 10dB figure. If you want to conduct a test with your system and your ears, please do so and your conclusions, however they differ from scientific extrapolations are your conclusions and go by those for you.

No one can claim that someone can or can not hear certain things..unless they are that particular person.

Subjectivity and objectivity will ALWAYS create disagreements and arguments amongst those who think they know..and those who use the guidelines of established math and science.

Because of this, their is NO definitive answer.


"An auditorium produces a departure from the inverse-square law because the reflected sound adds to the direct sound at a given location."

I cannot tell you how many times I was told "you are wrong" on the internet when I have pointed this out so many times. Scientists simply have a hard time departing from their professors and textbooks. When people tell me my perception is "wrong" my answer is "I am dealing with an idiot, why subject yourself to this crap"

If anyone wants to dispute this, they can do it with themselves. I can't prove "shit" either way.

anonymous Wed, 12/28/2005 - 07:24

thanks for the info... i just find it funny, and a bit disturbing, that for years i've done with 6db being twice as loud.... well, hasn't gotten me into any trouble yet....

i think, actually, that a 6db increase is doubling a VOLTAGE increase... if i remember from somewhere...

but, all of these dbs, voltages, power, volume, level, loudness.. dbU, dbV.... enough to drive someone crazy...

audiowkstation Wed, 12/28/2005 - 07:32

Power referrs to "watts", be it acoustic or electric. yes, you generate acoustic watts as well.

"Unfortuantly, the average 12" woofer in a sealed enclosure would have to travel back and forth (peak to peak) about 8 inches to produce one acoustic watt at 30hz, an obvious impossibility."

(Reprinted by permission from Pat Snyder, speakerlab, publication 1971.)

audiowkstation Wed, 12/28/2005 - 07:39 remember that the 10dB figure was derived from a subjective extraoplation of peoples stated perception..which by no means gurantees any real accuracy. It is simply a consensus and not an absolute..and it varies with frequency and SPL as well as subject to subject noninterchangablitiy AND acoustic environment PLUS the person administering the tests.

Hundreds of thousands of hours have been wasted on this subject alone. No one wants to admit defeat.

anonymous Wed, 12/28/2005 - 07:40

definitely the human ear can hear a smaller increase than 3db.....

ask a mastering engineer who uses .5 db steps (or smaller) on a nice GML or mastering EQ...

i think 1db is pretty easily discernable.... smaller than that takes a more trained ear.. though i'm sure most people who read these forums could tell a .5 db difference....

audiowkstation Wed, 12/28/2005 - 07:44

Going by the textbooks, 1dB is the average level of change that is perceptable with a cross-section of people tested, from great hearing to horrible, it is an average.

Yes, I use 0.25dB as my personal reference. Anything below that with a volume change of a single sound I cannot discern every time.

In a mixdown however, a change of as little as 0.1dB on a single channel with other channels operating would be the increment (not larger than 0.1db per increment to clarify) I will accept in a digital mixer and prefer 0.05dB myself.

My analog mixer with its P&G faders, I have been able to "breath on it" and change the level as little as 0.05dB as measured. Of course, some heavy modifications were done to several of the channel strips to make this happen. Very fine hardly see it move at all with 0.05dB change.

0.01dB resolution is simply beyond needed.

audiowkstation Wed, 12/28/2005 - 07:55

I thought 3dB was twice as loud...? Or at least the minimum increase/decrease that can be heard.

What "sounds" twice as loud to you?

3dB is doubling/halfing of the power as referenced to a "change of 3dB"

Take a source and turn it up to where it "seems" twice as loud and note the level difference. This would be your reference for you.

Like I say, we ALL hear differently, throw away the textbooks and see for yourself. It is subjective. It is not a scientific fact..just recognized as one.

sub·jec·tive ( P ) Pronunciation Key (sb-jktv)

Proceeding from or taking place in a person's mind rather than the external world: a subjective decision.
Particular to a given person; personal: subjective experience.
Moodily introspective.
Existing only in the mind; illusory.
Psychology. Existing only within the experiencer's mind.
Medicine. Of, relating to, or designating a symptom or condition perceived by the patient and not by the examiner.
Expressing or bringing into prominence the individuality of the artist or author.
Grammar. Relating to or being the nominative case.
Relating to the real nature of something; essential.

Randyman... Thu, 12/29/2005 - 19:35

And just for reference, Doubling the Power and the # of Speakers will also result in a +6dB increase. I would think that doubling the Power and Speakers would constitute "Twice As Loud", but like mentioned, it is how we "Perceive" it - so some people think 6dB is "Twice As Loud", others think 10dB is "Twice As Loud". Just depends on who you ask, and how they "Hear" it :p

We are exposed to waveforms with no meaning. WE are the ones who quantify and "Make Sounds" in our brains (how we decode these meaningless waveforms into something we consciously recognize as "Sound").


Kev Fri, 12/30/2005 - 00:12

Randyman... wrote: And just for reference, Doubling the Power and the # of Speakers will also result in a +6dB increase.

I was trying to keep out of this one
the +6db increase is only because the two speakers have coherent signals ... exactly the same signals.

otherwise it would be a +3dB increase.

these discussions can be difficult and careful use of the words and terms and references are needed.

The dB scale is a relative one and so the reference needs to be quoted.

I'll shut up now.

anonymous Fri, 12/30/2005 - 08:47

Kev brought up a very good point. If anyone is unfamiliar with coherence, and wants to see/hear what he is talking about try the following.

Generate two different tracks of pink noise at -7.9 dBFS. Call them tracks 1 and 2. The RMS value of each should end up at -20 dBFS.

Create track 3, mixing tracks 1 and 2 with no attenuation. The RMS value should end up at -17 dBFS. Tracks 1 and 2 are uncorrelated (random with respect to each other), so this was non-coherent summing.

Create track 4, mixing track 1 with itself with no attenuation. The RMS value should end up at -14 dBFS. Track 1 is by definition correlated with itself, so this was coherent summing.

Solo track 3 and then solo track 4. The volume difference should be very noticeable.

audiowkstation Fri, 12/30/2005 - 09:29

The "doubling" of the sets of speakers must be spoken of with a caveat.

Example...Series a set of 8Ω speakers, ending up with 16Ω nominal impedance and watch the amplifier power change. Do the Parrallel of a pair of 8Ω speakers, dropping to 4Ω nominal and notice another power change. 4Ω loads draws twice the power of an 8Ω load so in testing the output of speakers in pairs, one must see if they are parallel or series wired if attached to the same channel.

Likewise, loudspeaker sensitivity is measured with one loudspeaker and we use them on separate amplifiers so the impedance problem does not show itself but the increase in sensitivity due to having more drivers pushing air...does.

I often wonder why loudspeaker manufacturers don't publish sensitivity ratings per pair instead of each unit.(?) In my reviews, I note the per pair output at 1M 1W 2.83V @ 8Ω per speaker and measured pair.

Randyman... Fri, 12/30/2005 - 16:09

Well, since we are talking about "Doubling Loudness" in respect to a single signal or "mix", I just assumed ( :roll: ) "Coherent Waveforms" were a given (we are talking about raising the volme on a slider, etc). It is hard to compare a Snare hit to a Guitar Strum (2 non-coherent signals) when speaking about relative volume. If you use a peak reference, the Snare will appear "louder" - if you use an RMS/VU type reference, the Guitar will seem "louder". Combine them, and you certainly WON'T get +6dB louder. I guess I assumed too much ;) - but yes I know all about non-coherent summing...

On the power situation - like I specifically mentioned DOUBLING the power (be it from adding an identical set of Speakers and Amps, or by doubling JUST the speakers on an amp that TRULY doubles its output as the impedance is halved - not very common in power amps) - you still end up with 6dB more amplitude. I specifically said "DOUBLING the POWER" - not "Halfing the power" like wiring speakers in series ;) ... I was just trying to correlate the "Twice As loud" concept to a 6dB increase, which happens to be a doubling of POWER and DRIVERS in most circumstances (should be twice as loud in my eyes). Just a point of reference...

Non-Coherent signals are what we experience when summing multiple signals/tracks while mixing, and I don't think we were heading in that direction - but I always Assume incorrectly...


Davedog Fri, 12/30/2005 - 17:09

Everybodys' up on their tippytoes for this question. ONE of these days the scientific community is gonna have to agree on a measurement system that agrees with everyones needs and becomes a standard by which we can all march to....

Until that time I will continue to mix it so that everything is louder than everything else.

It does beg a question....Which is twice as loud? The neighbors little gitar-god with his first real amp, or the significant other, awakened in the middle of the night by you trying out that inspirational mixing information you got at R.O. ???

Would that be measured in 6 or 3 db increments?

Do scientists have to sleep on the couch after discovering something new?

anonymous Sat, 12/31/2005 - 09:02

If I added some chili to the pot, I apologize. As a point of reference, on the scale of talent and experience on this board, I am an audio moron. I do have a lot of signal processing experience in other fields. Since I am an engineer, I also tend to be a bit pedantic...

I mentioned the coherent versus non-coherent experiment because you do have to be carefull you are comparing apples to apples. I have read similar threads in other places where people are assuming a non-coherent situation, eg people talking in a room.