What make up the unit of dB when measuring analog signal?

[MRC01]

Stripped to its bare essentials, a dB is a ratio between 2 measurements or quantities. I used dbV to clarify that I mean 20*log(R) not 10*log(R) for whatever ratio R is being compared. Of course I didn't mean I was comparing voltages.

 

[RayDunzl]

I believe you used the correct # in your dB calc and just fat fingered in typing; it should be around 0.035"

Yes, 0.035 inch, as a typical old-school gap, as in the original thought...

It's been a long time since I gapped a plug, what with the platinum and iridium design for the electrodes.

 

[earlevel]

Stripped to its bare essentials, a dB is a ratio between 2 measurements or quantities. I used dbV to clarify that I mean 20*log(R) not 10*log(R) for whatever ratio R is being compared. Of course I didn't mean I was comparing voltages.

Yeah, maybe we can say... (dBV)/V

 

[xaviescacs]

not to mention distances between stars

Distances between stars are measured in parsecs (pc), and the diameter of the universes is estimated to be about 28.5 Gpc, 2.85*10^10 pc, so not that much

 

[xaviescacs]

0.000000000001W

1·10^(-12) W

 

[MCH]

This makes me think about the electronvolt, the unit of energy. The unit is defined as a reference, so any measure can be seen a proportion with respect to that reference. So one could say that 1TeV = 1*10^(12) eV = 120 dBe with respect to the reference.

But you see? eV is also used to measure energies where the number is still small. More convenient to say tha the ionization energy of Oxygen is 13 eV than 2exp(-18) J, isn't it? Yet another very large scale

 

[antcollinet]

It does. Assuming dBVac. Only reference to voltage and current is ignore. There will be phase shift with capacitor in series or parallel. Meaning with voltage as reference you can measure perfect flat frequency response. Even you measure current, it will be prefect flat response. But the current does not propagate in time with voltage with all the capacitors around. tens of picofards is enough to affect the audible frequency at higher or lower depend on how is it place, being high pass or low pass filter. Of course, only measuring voltage will be flat throughout.

Amazing then that audio reproduction has for nigh on 100 years used voltage as reference and it has worked for all this time. Every broadcast, every recording, every reproduction, every live performance.

I wonder how we have ever heard any audio quality at all.

 

[waynel]

You already induce a phase shift with capacitors around the transistor, opsamp, DAC, that without cables. Those small picofards are enough to cause a shift phase in audible range up 8 degree. They are already used to block DC leakage or wrong connections.

Please take a look at the following link on Bode plots . Small signal transfer functions are characterized by magnitude and phase. Phase is not ignored.

Bode plot - Wikipedia

en.m.wikipedia.org

 

[antcollinet]

Distances between stars are measured in parsecs (pc), and the diameter of the universes is estimated to be about 28.5 Gpc, 2.85*10^10 pc, so not that much

That is only 104.5dBps. Pretty much audible for someone with good ears.

 

[xaviescacs]

But you see? eV is also used to measure energies where the number is still small. More convenient to say tha the ionization energy of Oxygen is 13 eV than 2exp(-18) J, isn't it? Yet another very large scale

Yeah but this is a different thing, is a change of units, but you still have an energy measure in this case. The db thing is dimensionless.

P.S.: You were quick enough to quote this part... I deleted it just because after the second read I thought it was BS.

 

[MCH]

Yeah but this is a different thing, is a change of units, but you still have an energy measure in this case. The db thing is dimensionless.

P.S.: You were quick enough to quote this part... I deleted it just because after the second read I thought it was BS.

Was answering to the original comment of someone else "our scales are too large" or something like that, i don't remember, probably bs from my side too. But now that you say it, a real life example heard today at work, someone determined that the migration of a chemical in a food envelope was 0.16 ppb, so 160 ppt. Very small, adimensional, and no need for db . I go to sleep now

 

[xaviescacs]

That is only 104.5dBps. Pretty much audible for someone with good ears.

You know... with this dB thing and other stuff, I always have the impression that engineers have the ability to make difficult what's easy and easy what's difficult.

 

[antcollinet]

You know... with this dB thing and other stuff, I always have the impression that engineers have the ability to make difficult what's easy and easy what's difficult.

It's the second. Engineering is difficult. Engineers use tricks (like dB) to make it easier to suit our limited intellect.

If you want difficult, go watch a theoretical physicist.

 

[xaviescacs]

Stripped to its bare essentials, a dB is a ratio between 2 measurements or quantities.

Personally, I think that it would be a ratio without the log. To my eyes the log of a ratio if the difference of logs, it's kind of ugly to me that fraction inside the log, so not a ratio, but a "smoothed" difference, or a difference in a logarithmic scale. That's how see it.

 

[MRC01]

Personally, I think that it would be a ratio without the log. To my eyes the log of a ratio if the difference of logs, it's kind of ugly to me that fraction inside the log, so not a ratio, but a "smoothed" difference, or a difference in a logarithmic scale. That's how see it.

Strictly mathematically speaking, the dB is the log of the ratio, not the log of the difference. Suppose we have two values V1 and V2.
To express it in dB, it's 20 * log (V1 / V2). It is not 20 * log (V1 - V2).
The log scale is convenient for expressing ratios that vary over several orders of magnitude.

 

[antcollinet]

Personally, I think that it would be a ratio without the log. To my eyes the log of a ratio if the difference of logs, it's kind of ugly to me that fraction inside the log, so not a ratio, but a "smoothed" difference, or a difference in a logarithmic scale. That's how see it.

Well yes - a logarithmic scale of a ratio. Such that you can add different ratios rather than needing to multiply them. And when you have huge ratios, they can be expressed easily.

They also (conveniently) match how our ears perceive volume (somewhat logarithmically)

 

[xaviescacs]

It's the second. Engineering is difficult. Engineers use tricks (like dB) to make it easier to suit our limited intellect.

Life is difficult.

If you want difficult, go watch a theoretical physicist.

I'm trying to invite my younger brother, who is writing his thesis on quantum mechanics fundamentals, to a family party this Saturday, but he's not answering, too busy I suspect. I left if just after finishing the degree and gave all my books to him. Since I joined ASR I regret I didn't keep the EM ones though.

 

[xaviescacs]

Strictly mathematically speaking, the dB is the log of the ratio, not the log of the difference. Suppose we have two values V1 and V2.
To express it in dB, it's 20 * log (V1 / V2). It is not 20 * log (V1 - V2).
The log scale is convenient for expressing ratios that vary over several orders of magnitude.

Strictly speaking, I said "difference of logs" in the post you are quoting ... se my first post on this thread, second bullet.

 

[MRC01]

Strictly speaking, I said "difference of logs" in the post you are quoting ... se my first post on this thread, second bullet.

If you are saying that log(V1/V2) = log(V1) - log(v2), then I agree

 

[xaviescacs]

If you are saying that log(V1/V2) = log(V1) - log(v2), then I agree

Yes. The log function is a group homomorphism from R* to R⁺, if you prefer a more formal description. There isn't really much room for disagreement here.