Loudness Question

[JohnA]

How many dB is "twice as loud"? Where did I get the idea it was 10 dB?

[DrWho]

I think it depends on which frequency range and what SPL you're starting the comparison at...as can be seen from the fletcher-munson curves.

I've always heard something from 6-10dB is perceived as twice as loud.

[PrestonTom]

John, you are quite correct.

The quick lesson in psychoacoustics is:

1. the smallest increment (or decrement) that folks can detect (not all the time, but at above guessing levels) is a bit less than 1 dB (and is mostly true over a range of frequencies and overall levels). This will only happen if the any degradation due to memory, alertness, attention etc are minimized. Causal listening at home would require closer to at least 3dB

2. If a person is asked to adjust a sound or assign a number regarding the ("perceived") loudness, then a 10 dB increment is required to "perceptually double" the apparent loudness.

As Mike alluded to this second effect needs to be qualified when you look at extremes (eg, low intensities especially at low frequencies). Then the required increase can be smaller. This deviation is seen in the Fletcher-Munson curves. Incidentally, those curves are spaced at a 10 dB difference (wrt to 1000Hz).

Good Luck,

-Tom

[WMcD]

I agree with Tom. Of course.

This does appear in Olson and/or Barenek. The unit of loudness is a phon. That corresponds roughly to the dB level above 40 dB. IIRC.

It does mean you need 10 dB work of power (10 times) to get a perceived doubling of loudness. That seems like a lot but consider that it kind of helps when we're going in the other direction. Losing 10 dB of power only halves the perceived loudness.

There is something odd which happens in conjunction with FM equal loudness curves. If you look at them at the very low end, you go from almost inaudible to loud (or the same as a 1000 Hz signal) in about 10 or 20 dB. Tomlinson Hulceman (sp???) said something about this in his book "5.1 Sound, Up and Running."

If you go to something like 30 Hz, you need a lot of power to get it perceptible. But once you get there, the equal loudness curves are crowded together. Accordingly, you don't need a wide dynamic range of bass. This is why the subwoofer range does not need as great a number of bits in a multi channel recording. I still have not figured it out to my satisfaction.

I recall that old timers, speaking bass perception, that if you had the correct filters, you could set the low bass to a constant level, and then only vary program for the rest of the range above. Part of this is the FM curve. But it might be the phon issue too.

Maybe Tom can comment.

Gil

Gil

[garyrc]

.... and, in psychoacoustic experiments, subjects (oops!, people) vary around a center point in their perception/judgements of subjective amounts such as "twice as loud," so I'll bet that the 10 dB figure is an average.

I believe 10 dB was also called "twice as loud" by Martin Mayer in his 1958 book "High Fidelity." Ahhhhm High School memories!

[djk]

I generally figure twice the voltage (6dB of power increase) is twice the loudness.

Depends on the levels and frequencies, of course.

The contours of equal loudness (FM) are only 3dB apart in the 30hz~40hz region at about 90dB. At really high volume it can take more than 10dB to sound twice as loud, your eardrum gets 'tight'.

[jacksonbart]

Its because I first coined that term back in 1923 while working for Henry Ford.

[PrestonTom]

I generally figure twice the voltage (6dB of power increase) is twice the loudness.

Depends on the levels and frequencies, of course.

The contours of equal loudness (FM) are only 3dB apart in the 30hz~40hz region at about 90dB. At really high volume it can take more than 10dB to sound twice as loud, your eardrum gets 'tight'.

You are correct about those low frequencies pinching together, especially at the lower levels. In many ways the sensation at a very low frequencies "is different" compared to other spectral regions. However the 10 dB increment (or decrement) to approximately "double" (or halve) the loudness is a good number to remember. It captures most of the relation over a broad range of frequencies and overall levels.

For historical reasons, these loudness contours are ultimately anchored to loudness relations relative to 1000Hz at 40 dB SPL (phon scale)

-Tom

[pauln]

Is this anything other than the very operational definition of "twice as loud"?

Let us have two speakers... one comes on at x dB, it's loud.

Second one comes on at x dB.

With both speakers on you have x+3 dB. So, now it is twice as loud, 3 dB louder.

Any different subjective notion of what is twice is loud; +6 dB, +10 dB or whatever is simply not the right answer. It may be true that the average person makes this error consistently, and it may be true that people in general make close to the same degree of error. For psychoacoustical research that is interesting, but to answer the question of what is really twice as loud, looks like it is 3 db.

[djk]

"Any different subjective notion of what is twice is loud; +6 dB, +10 dB or whatever is simply not the right answer. It may be true that the average person makes this error consistently, and it may be true that people in general make close to the same degree of error. For psychoacoustical research that is interesting, but to answer the question of what is really twice as loud, looks like it is 3 db. "

Hey, I'm sorry you didn't follow much of the previous talk on the subject.

3dB is twice as much POWER, and other than the exception I cited it DOES NOT PLAY TWICE AS LOUD.

[rjrbass]

According to the attached link, "twice as loud" requires a 10 db increase, which requires 10 times the power.

http://www.analogrules.com/dbwatts.html

[dgoreck]

Good article.

I guess it is right in 10db increase is 10X power. +9db would be say 8W, and +12db would be 16W, when you think of it in +2db increaments.

[pauln]

Feeling twice as dirty as earlier in the day, I run a bath, but find that it feels twice as hot as I expected. This makes me mad, and in a few minutes I'm twice as mad as I was before. Is it twice as inconvenient to wait for it to cool or draw the bath again? Now I am resolved to apply twice as much attention when preparing my next bath. Will I smell twice as good when I'm done?

[DrWho]

Any different subjective notion of what is twice is loud; +6 dB, +10 dB or whatever is simply not the right answer. It may be true that the average person makes this error consistently, and it may be true that people in general make close to the same degree of error. For psychoacoustical research that is interesting, but to answer the question of what is really twice as loud, looks like it is 3 db.

Since when are watts a measure of loudness? Why not choose sound pressure, or sound velocity, or temperature change, etc...?

There's lots of different behaviors that can be measured that can all correlate to a specific acoustic intensity - doubling one may not necessarily double all the other parameters though.

To me, the only way it makes sense to reference watts as a measure of loudness is to say that our ears measure power - sadly, power is not one of those things that can be directly measured.

[JohnA]

I'm too old.

The question came about because my Ducati came with an oversized racing exhaust that produced 105 dB at fast idle. I bought some extra baffles that cut that racket by 11 dB. I've been calling that just over half. At least it doesn't sound like the common Hardley.