Hello all,

In the manual dated 2011-09-15, p. 75, the "amplitude ratio to dB" formula is writen as 10 * lg (R)

Page 76 shows the reciprocal conversion (is that the correct wording?) as using a multiplying factor of 20.

The dB <-> power conversions seem consistant (pp. 76 & 77), and use the 10 factor I am familiar with in EE.

Would anybody know of a field of engineering or science, in which something called an amplitude, would actually have a "dimension" akin to a power (using a dB factor of 10), therefore causing a confusion?

I am a EE, but haven't used these units much.

I do know that electrical power can be in peak values, or RMS average values. I don't know if this comes into play here.

But here is a units website that deals with dB values and there seem to be some agreement with these conversion, but you have to know if you are basing the conversion on peak (full scale), or average.

http://www.unc.edu/~rowlett/units/dictD.html

decibels don't have units, per se, but there are different scales.

This comes into play when talking about gain of, say amplifier, since power ratio is a square of input/output voltage ratio

p1/p2=(v1/v2)^2

G=20log(y/x)[dB]=2x10log(y/x)=10log(y/x)^2[dB]

Something like that (its been decades..)

Bonjour Etienne,

deziBel denote ratios of two measurements of equal kinds. Such ratios are dimensionless (or unit-free). So no problem with a logarithmic scale. Confusing - maybe. But useful e.g. assessing noise, amplifiers, damping, etc.

Edit: The factor on page 75 must be 20, of course. Thanks for pointing to that! Will be corrected in next edition. Merci!

*Edited: 21 Sept 2011, 2:44 a.m. *

Hello,

First let me apologize for not not knowing how to address you, I am not familiar with your mother tongue & culture, so even if I recognized your frst from your last name, I would not be sure which one to use.

Thanks for your input, you are indeed correct.

Hi David,

Thanks for the link. I knew there were units I had never heard about :-)

About electrical power "varieties", they do not come into play (they do "behave" in the same way), but one has to be consistant (peak / peak or RMS / RMS) between the two values being evaluated, in order to keep the scaling factor being meaningful.

Thanks

Etienne

Hello Walter,

It seems I was not entirely clear.

As you noted, my point was that I thought there was an inconsistency in the manual. Not being able to read the code, I could not decide which scale factor was correctly documenting the function you had actually implemented. That is now solved.

BieteschÃ¶n...

Etienne

Thank you for posting the question Etienne. I learnt something today!

Hi all,

FYI, I committed an updated manual at sourceforge some minutes ago. Hope all glitches are treated. But don't nail me for the number of flash pages - it keeps changing all the time.

To those happy few who go: Enjoy the conference!

To the rest of us: Hope we'll get some overlays from Eric nevertheless.

Walter

I've updated the release package with the recent changes.