How to best explain dB to non-audio experts?

[abdo123]

Yes. And reference exponential growth of covid infections to give them something to chew on...

One of the many situations when humanity fucked up by having zero comprehension of exponential growth.

 

[RayDunzl]

In the case of 130 db, that represents a ratio of more than 3 million to one. If full scale is 2 volts then the -130 db value is a bit less than .65 millionths of a volt.

What's a Volt?

 

[Wes]

What's a Volt?

an EV formerly made by GM

 

[Blumlein 88]

What's a Volt?

One of these might be the one definition we are looking for:

• n.
  
  A circular movement executed by a horse in manège.
• n.
  
  A sudden movement made in avoiding a thrust in fencing.
• n.
  
  The SI-derived unit of electric potential and electromotive force, equal to the difference of electric potential between two points on a conducting wire carrying a constant current of one ampere when the power dissipated between the points is one watt.

Or alternatively one volt is 1/12th of how many volts your car battery has assuming you don't have an electric car.

 

[DVDdoug]

Lets say you have a DAC where the jitter is at -130dB or lower. How you would explain to somebody who is not familiar with audio how good that is?

I'd make several sample files with Audacity (or your favorite audio editor) each one reduced by 20dB. (Of course, export as 24-bits if you're going to -130dB).

Another "interesting" experiment/demonstration might be to mix two files together, one at various lower levels to demonstrate masking.

And if it's not too confusing you could talk about SPL levels and the difference between the threshold of human hearing and 130dB SPL. i.e., If you were listening to music at 120dB then any noise or distortion at -130dB is 10dB below the threshold of hearing. And... below the ambient noise, plus you probably have some temporary deafness from listening at 120dB!

 

[spacevector]

The SI-derived unit of electric potential and electromotive force, equal to the difference of electric potential between two points on a conducting wire carrying a constant current of one ampere when the power dissipated between the points is one watt.

That's the one RayDunzl is looking for I think and of course, the next question is what is an Ampere. That would be one Coulomb of charge going through a conductor every 1 second. How about a Coulomb? We need charged particles to help define that. I think one electron is supposed to have 1.6*10^(-19) Coulomb worth of charge.

@Koeitje - are you trying to explain jitter to your friend or are you trying to convey what dB means?

 

[MRC01]

I am 2m tall. My brother is 20m tall. How much taller is my brother than me?

Is your brother a tree?

First, remember that there are actually 2 different "sizes" of dB: for power, and for voltage. Each has a different formula:
dbP (power) = 10 * log(R)
dbV (voltage) = 20 * log(R)
Where R is the ratio you are expressing in dB.

From the math you can see that a dbP is bigger than a dbV.
1 dbP means 10 * log(R) = 1, which means R = 10^(1/10) = 1.26
1 dbV means 20 * log(R) = 1, which means R = 10^(1/20) = 1.12

So we could say that sales tax in the Seattle area (10%) is about -20 dbV.
Your brother is 10x taller so R = 10.
dbP = 10*log(R), so he's 10 dbP taller than you.
dbV = 20*log(R) so he's 20 dbV taller than you.

Put differently, every factor of 10 in the ratio is 10 dbP, or 20 dbV.
In reverse, suppose your speakers measure distortion at -40 dB. Every 20 dB is a factor of 10, so this is 2 factors of 10, which is 10*10 = 100. So the distortion is 100x quieter than the signal.

 

[Wes]

if they take photos much, tell them it is sorta, kinda like f-stops

 

[Cbdb2]

Tell them every 10db is a 10x change. So +20db is 100 times, (-20 is 1/100 change) +30db is 1000, so +130db is a 1 with 13 zeros after it.

 

[digitalfrost]

Is your brother a tree?

You should've watched the video before replying to me but I appreciate the explanation

The point I was trying to make was, to get people to understands dBs, they need relatable size differences. Voltage, Power, nobody cares. It's already too complicated at this point. Size is something everybody can understand. Give them entertaining questions like the one I asked and you get a reaction out of them. Like I got out of you.

Then scale them up orders of magnitute more and they will realize the "little" difference in dBs these make.

 

[faheem]

As an electrical illiterate. These 2 videos helped me alot

They're long but thorough.

The easy way I understood it, dB is a ratio and not a value. Understanding the log part was the most difficult and took me time to grasp.

 

[MRC01]

Tell them every 10db is a 10x change. So +20db is 100 times, (-20 is 1/100 change) +30db is 1000, so +130db is a 1 with 13 zeros after it.

That's true for a dbP, but not for a dbV. Most of the specs that Amir measures here are voltage, where every 10x change is 20 dB.
For example, if someone says SINAD is 100 dB, that's 10^5 or 100,000:1, or 0.001%.
It's not 10^10, or 0.00000001%.

This is another common confusion about dB: when to use dbV vs. dbP.

 

[DonH56]

Jitter intro: https://www.audiosciencereview.com/forum/index.php?threads/digital-audio-jitter-fundamentals.1922/ -- has a graph relating jitter (aperture error) to bits. For SNR with no other error sources converts bits to dB using SNR (dB) = 6.021*Nbits + 1.8 where Nbits is the number of bits. There is more info and more pictures in the second part: https://www.audiosciencereview.com/...igital-audio-jitter-fundamentals-part-2.1926/

For other articles look in the reference sub-forum. Here is a link to a list of the ones I authored: https://www.audiosciencereview.com/...and-inquisitive-minds.8121/page-3#post-808622

HTH - Don

 

[Propheticus]

Just in case all these (correct) explanations are far too technical for your audience:

A really rudimentary way to show exponential behaviour is the task: guess how often you can fold in half a piece of A4 paper. Now go try... Once you reach like 7 times (2^6) it becomes pretty hard and soon impossible.

Examples with bacteria doubling periods also work nicely.

 

[Koeitje]

You shouldn't try, because it would not be true. That DAC is sonically indistinguishable from any other properly functioning one.

Don't promote audiophoolery.

It can be used to explain that two values are both crazy good. Not to explain why one sounds better.

 

[Koeitje]

This is the right approach. First off db is a ratio, and if you can get across the idea of a log scale and that db is a ratio on the log scale you'll have it.

In the case of 130 db, that represents a ratio of more than 3 million to one. If full scale is 2 volts then the -130 db value is about .63 millionths of a volt.

I think this is a good way. If you add that the voltage determines how loud something is it becomes even more clear what kind of differences we are talking about.

 

[Wes]

just be sure to emphasize that natural logs are much better than the other kind

 

[Blumlein 88]

just be sure to emphasize that natural logs are much better than the other kind

Is that because natural logs are more analog like?

 

[danadam]

Lets say you have a DAC where the jitter is at -130dB or lower. How you would explain to somebody who is not familiar with audio how good that is? I feel this is something that could be communicated in a better way.

Play them this: https://www.audiocheck.net/audiotests_dynamiccheck.php . It goes "only" to -90 dBFS, but should be illustrative enough.
Or this:

 

[danadam]

It goes "only" to -90 dBFS, but should be illustrative enough.

Ok, so human hearing is most sensitive around 3 kHz. In attachment there are 4 files, each contains G7 note (3136 Hz) beeps, 2 beeps per level, the levels are decremented by 6 dB. The files are -60 to -78, -78 to -96, -96 to -114 and finally -114 to -132 dBFS. At what point do you (or them) stop hearing the beeps when listening at your usual volume?