Talk:Decibel/Archive 3

Doubling
Our article says an increase of 3 dB corresponds to a perceived doubling of sound volume. http://www.phys.unsw.edu.au/jw/dB.html says: "Experimentally it was found that a 10 dB increase in sound level corresponds approximately to a perceived doubling of loudness". Can anyone clear this up? Boris B (talk) 22:43, 19 May 2008 (UTC)


 * This has come up before, but I don't recall it ever being resolved. One problem for me is that I don't understand what is meant by "perceived doubling of sound volume". Do you? Thunderbird2 (talk) 22:51, 19 May 2008 (UTC)


 * http://hyperphysics.phy-astr.gsu.edu/Hbase/sound/loud.html seems quite specific, and even mentions that the dB is popular because it (roughly) corresponds to perceived loudness; "it takes 10 violins to sound twice as loud as one violin". It also notes that sound frequency affects perceived loudness, so the phon scale states that 60phon = 60dB at 1000Hz ... My guess is the "3dB = 2x louder" statement is an error deriving from perceived brightness which uses a log2(luminance) scale. --Redbobblehat (talk) 23:14, 3 February 2009 (UTC)


 * A 3.01dB increase in luminance is perceived as a doubling of brightness and 6.02dB quadruples brightness because 3.01 * 10 * log10(x) ≈ log2(x). Decibels crop up frequently in (video and digital) photography: dynamic range, PSNR levels, etc are often measured and specified in dB. As photographers still work with predominantly log2 exposure scales such as f-numbers and exposure values, converting between dB and f-stops or EVs might be worth a mention in this article? --Redbobblehat (talk) 23:14, 3 February 2009 (UTC)


 * This is confusing me as well. We can't perceive both a 3dB and a 10dB increase as a doubling in loudness.  If anyone fully understands this, please do clear it up. 193.38.155.28 (talk) 17:03, 7 December 2009 (UTC)


 * The 10dB increase value for a "perceived" doubling in loudness is a "subjective" matter, whereas the 3dB increase as a doubling of intensity is an "objective" measure. "Perceived" loudness and actual intensity are "not" identical quantities.  Note that the article doesn't actually say that an increase of 3 dB corresponds to a perceived doubling of sound volume, but it does say that perceived intensity works logarithmically.  Sound pressure gives a reference to http://www.sengpielaudio.com/calculator-levelchange.htm, which gives more detail.  David Biddulph (talk) 17:28, 7 December 2009 (UTC)

I am told that 68db is twice as loud as 65db(3db = 100%). Firstly, is this correct? If so please can you please explain. It relates to engine noise. —Preceding unsigned comment added by 217.33.145.182 (talk) 22:22, 16 June 2008 (UTC)


 * Yes that's correct. A 3 dB increase is very close to double. This is because if a/b = 2 (i.e. a is double b), then log_10(a/b) = log_10(2) =~ 0.30103, and multiplying by 10 you get almost exactly 3 dB. —Preceding unsigned comment added by Fundamentisto (talk • contribs) 16:46, 6 August 2008 (UTC)

Confusion about bel - decibel relationship
" L_\mathrm{dB} = 10 \log_{10} \bigg(\frac{P_1}{P_0}\bigg) \,

Since a bel is equal to ten decibels, the corresponding formulae for measurement in bels (LB) are

L_\mathrm{B} = \log_{10} \bigg(\frac{P_1}{P_0}\bigg) \, " I think the sentence should read 'since a bel is a tenth of a decibel' and not 'a bel is ten decibels'. —Preceding unsigned comment added by 128.220.29.227 (talk) 19:13, 3 September 2009 (UTC)


 * What's there is correct: one bel is ten decibels. The prefix deci- means "one tenth"; you may be confusing it with deca-. Hqb (talk) 20:07, 3 September 2009 (UTC)


 * The formulas given have the opposite relationship though. If the statement is right, the formulas are wrong. 129.78.64.101 (talk) 00:55, 13 April 2010 (UTC)


 * No, the equations are correct too. Oli Filth(talk&#124;contribs) 07:04, 13 April 2010 (UTC)

Non equal impedances
Decibel is a measure of power ratio. Of course voltage ratios can also be expressed as explained in the Amplitude, voltage and current sub section. However, in order to use this relation, one must be sure that the input and output impedances are equal. Otherwise an correction term of such as 10•log (Ri/Ro) should be added. Nedim Ardoğa (talk) 10:00, 17 September 2009 (UTC)

The restriction to power is inherently nonsensical
This notion about decibels being restricted to power is widespread, but it manifestly makes no sense. It is the log of the ratio of two quantities expressed with identical dimensionality, and as such, in its essence it is the ratio of two quantities. It does not make one whit of sense for anyone to insist that a way to express the ratio of two quantities must not be applied to voltage in the straightforward sense, and that when applied to voltage, it must only be done in a way such that 3 dB corresponds to a doubling of the squares of the voltage. When the decibel value is to be interpreted as a power ratio and the given values are voltages, then the "multiply by 20" rule applies, because the effect is to convert a ratio of voltages to a ratio of their squares. But the assertion, that to speak of one voltage that is one-half of another voltage as being down by 3 dB is non-compliant and nonsensical, is simply ludicrous. It belies the mathematical simplicity of the concept. This "you always multiply by 20 when dealing with voltages" is simplistic, sophomoric rubbish, and I am sick and tired of encountering this inane nonsense everywhere I look. If you want the effect of converting a voltage ratio to a power ratio, then yes, this is what you must do, but it assumes that you use the notion of "decibel" only in the context of power, and this belies the inordinate mathematical simplicity of the concept. It amounts to the worst example of "control freak mentality" that I have ever encountered. Additionally, while I am not certain, it seems to be that this also belies common practice of electrical engineering. Again I am not certain, but I seem to recall that the "cutoff frequency" of a simple first-order filter is the frequency at which the voltage is equally divided between the resistor and the reactive component (capacitor or inductor) in series, in which case the voltage at the output (across the inductive component) is one-half of the input voltage. The "cutoff frequency" is also referred to as the "- 3dB" point, which would make no sense if it is not allowed to used decibel in the context of voltage. This business of insisting that "decibel" be restricted to refer to power is simply nonsensical and wrong. The equation where the log of the voltage ratio is multiplied by 20 is an equation that applies when you want to express power ratios in decibels. It is NOT absolutely and universally incorrect to multiply by 10, and if you are interested in expressing voltage ratios in terms of decibels, then you would multiply by 10. This is not the only reason that this article annoys me, and if this is the sort of thing that Wikipedia is going to be known for, I would rather there was no Wikipedia at all. Why does this article use the word "amplitude" as a synonym for "voltage"? "Amplitude" is a more general term that lacks anywhere near the same specificity as "voltage". "Amplitude" is not a synonym for "voltage" except for within a context where voltage is implicit. This article takes a very specific mentality that is applicable in certain professional contexts and tries to superimpose that mentality universally, and this is simply wrong. —Preceding unsigned comment added by Anotherkaisersoze (talk • contribs) 20:06, 2 November 2009 (UTC)

Unfortunately, this is not according to current editorial standards of dB in the peer reviewed literature. A relative ratio (regardless of being a power or not), is defined as a "level". The definition of dB requires it to be a power ratio, which can be voltage or pressure squared, of course. See my comment (with references) below. Shofus (talk) 15:16, 4 May 2010 (UTC)

Popular use
This article is almost useless to the reader who comes here to find out about the popular use of decibel - how loud things are - like myself. Yes, you can eventually find the link to that if you look hard enough, but it should be a hell of a lot clearer. It should be in the lead at the very least. Cheers, Mostlyharmless (talk) 10:54, 26 November 2009 (UTC)
 * Absolutely. Like many WP technical articles it's bloated and jargon packed. It needs a clear section for the layman. 193.91.181.142 (talk) 22:54, 20 May 2010 (UTC) (Nick)
 * Absolutely. Just like the article on metre should tell me how far it is from Scranton PA to East Braintree MB. And the article on gallon should tell me how much gas I have left.  But seriously, you shouldn't expect to be taught everything while standing on one leg. --Wtshymanski (talk) 00:55, 21 May 2010 (UTC)

Format with numbers?
There is only one mention in this discussion of the topic of whether or not a space should appear between the number and dB used as a unit. I think that there should be a space (for basically the same reasons), but the body of the article uses both forms. I hesitate to make the change without more evidence of consensus, though I can say that I've read a couple of books that said the same thing... Only such things as the % sign and the temperature circle should appear immediately after the numbers, and unit labels should have a space (unless it's an adjectival usage that calls for the hyphen). Shanen (talk) 06:06, 2 March 2010 (UTC)
 * There should indeed be a space before the "dB", like for any other unit of measurement. As near as I can tell, the article was missing a few such spaces solely because they were implicitly removed by mode, rather than as a deliberate choice. (The spaces are actually there in the source.) I've made the fixes.
 * Incidentally, the degree symbol for temperatures is normally also written with a preceding space: "20 °C" or "75 °F", rather than "20°C" or (even worse) "75° F". However, when used for angles, or (by extension) geographical latitudes/longitudes, the degree symbol does indeed follow the number immediately: "turn 90° left" or "45.5° N". Hqb (talk) 15:22, 2 March 2010 (UTC)

Warning, Vandalism
Recently User SixteenLeft vandalised this page, specificly inserting an insidiously well written paragraph "Alternate Theory" into Definitons —Preceding unsigned comment added by 81.141.136.200 (talk) 09:04, 30 January 2010 (UTC)

Acoustics - something is wrong
I quote: "The ratio of the sound pressure that causes permanent damage during short exposure to the quietest sound that the ear can hear is above a trillion. Such large measurement ranges are conveniently expressed in logarithmic units: the base-10 logarithm of one trillion (10**12) is 12, which is expressed as an audio level of 120 dB."

This is inconsistent with the definition of dB on sound pressure and elsewhere: L = 20*log(P/P0) dB It should be 240 dB, not 120 dB. Now since we know that ear damage appears at app. 120 dB, we must conclude that the ear can stand a sound pressure that is app. 1 mio. (10**6) times stronger than the 'quietist' pressure. Therefore, I have changed the quote. —Preceding unsigned comment added by 93.178.142.237 (talk) 07:42, 1 April 2010 (UTC)
 * The error was that it's not the ratio of sound pressure but sound intensity (or power) that's 10**12. Sound pressure is convenient to measure, but does not properly represent how much sound energy is being transferred from source to listener per time unit. For example, it is not the case that if some sound source is barely audible, then a million instances of that source would cause hearing damage. Hqb (talk) 07:21, 2 April 2010 (UTC)

I have a concern about the acoustics section & would say something is wrong, but in a different sense. I came in search of some basic information about sound measurement / decibels. (Due to a licensing commission hearing about a neighborhood sound issue.) This article seems exhaustive in scope but lacks introductory information for those of us who are not physics or math-minded. I would encourage some great wiki author to add some introductory simple information on the use of the decibel measurement scale with examples and also providing some simple examples of how much louder for example, something at 30 db is than something at 40 (or any other gap or three.) —Preceding unsigned comment added by Wikikd (talk • contribs) 15:46, 22 June 2010 (UTC)


 * I think that the simple information you want is at the last bullet point in Decibel, and examples are at Sound pressure are linked from Decibel. David Biddulph (talk) 16:07, 22 June 2010 (UTC)

Definition of dB
The current version says: “The definition of the decibel varies depending ...”. The definition does not vary. The dB is always a power ratio, but squared field quantities that are proportional to a power is accepted since the proportionality constants cancel. This is editorial practice in IEEE J.Ocean.Eng., i.e. "Carey, W. M. Editorial Standard Definitions for Sound Levels in the Ocean. IEEE Journal of Oceanic Engineering, 1995, 20, 109-113" and "Carey, W. M. Sound Sources and Levels in the Ocean. IEEE Journal of Oceanic Engineering, 2006, 31, 61-75" These editorial decisions are based on: "Standard CEI-IEC-27-3 Letter Symbols to be used in Electrical Engineering. Part 3 Logarithmic quantities and units, International Standard, 1989" These citations were also removed from the edit. I think the last one should be in.

My edit from 23 of April was correct, but I agree it was a bit difficult to read. The current edit is formally not adequate as a definition.

Shofus (talk) 14:31, 4 May 2010 (UTC)

Axial ratio
Regarding the paragraph starting An example illustrating the subtleties of the 20-log vs. 10-log rules is given by the so-called polarization ellipticity,..., I have a number of quibbles. Would anybody object if I deleted it? or can something be done with it? --catslash (talk) 23:51, 4 May 2010 (UTC)
 * The term polarization ellipticity isn't in common usage (try ["polarization ellipticity" dB] vs ["axial ratio" dB] in Google).
 * It's not clear why a reader unfamiliar with the applicability of each 20- and 10-log rules might suppose it refers to a ratio of amplitudes on the principle axes, rather than to the ratio of the powers carried by the corresponding orthogonal polarizations.
 * It sounds like WP:OR.
 * It was out of place and we beat the 10 and 20 thing to death anyway - I agree with you and I've taken it out. --Wtshymanski (talk) 00:09, 5 May 2010 (UTC)

The decibel clarified: serving intuition and rigor
Can we really let this new section stand as it is? Interesting point of view, but as far as I know it is a view held only by the author of the paper (apparently the same person as the editor here). So currently the section is falling foul of WP:UNDUE and WP:COI as well as needing a copyedit. My opinion is that Boute's paper can be covered in the article but it needs to be drastically reduced and made clear that it is not mainstream. Anyone else agree?  Sp in ni ng  Spark  11:09, 31 July 2010 (UTC)


 * No, emphatically not. The section is one man's concept of a situation that has not been implemented or seriously considered for implementation. It is prescriptive, proposing change, not descriptive of what change has occurred. This is an encyclopedia, not a soapbox. If Raymond Boute's proposal is taken up, or even if standards committees publish their discussions of it, we can print that. Binksternet (talk) 17:12, 31 July 2010 (UTC)


 * It is strange to hear the appropriateness of a coherent set of definitions and derived properties being made dependent on standards committees. The essence of the decibel (representing large numbers) is not changed by the proposed section, nor are the numerical values.  What the proposed section does is providing a definition that is intuitively simpler, and also more satisfactory since it supports writing true equalities instead of vague correspondences. Improving understanding is a service to the Wikipedia reader community.  In view of this goal, it is not relevant how many other people have also provided such a simple formulation.  Only factual arguments (soundness of the development, possible advantages) should decide whether or not such a section is appropriate.  Since the section was removed without contacting me first and this might happen again, I wait for more arguments pro and contra before reinserting it.  Especially technical comments and presentation suggestions are welcome.Boute (talk) 20:15, 31 July 2010 (UTC)


 * As shocking as it may sound to a smart guy such as yourself, Wikipedia does not publish things based on whether they are true or not. What is more important to the encyclopedia is that the facts it presents are notable, reliably sourced and verifiable. Notability is what is at issue here. Your simpler definition for decibel, has it obtained notice by others? Point us to where it was discussed by others in the field, rather than presented at IEEE.
 * Additionally, the work you introduced was far too detailed for an encyclopedia article. You may want to investigate another online venue: Wikisource. Published papers can be uploaded in their entirety at that website, if the author is willing to give them freely to public, as it appears you are. Binksternet (talk) 20:58, 31 July 2010 (UTC)


 * Wikipedia is not a place to promote or publish original research, no matter how ingenious it may be. We report established knowledge, verifiable by *independent*, notable, third-party sources. Primary sources are permissible in technical fields when published by standards committees for the general good, etc., but not by individuals with interest in self-promotion. Aside from that, the additions do not integrate with or support the existing material, in any form. In fact there is very little discussion of the material presented, mostly just mathematical derivation. If you can't express in words the content of your ideas, then this isn't the place to put it. WP isn't the place to publish lengthy original academic mathematical derivations. No general purpose encyclopedia does, AFAIK. It should be pretty obvious to anyone, that this doesn't fit. Kbrose (talk) 22:03, 31 July 2010 (UTC)


 * The material is not original research, as it has already been published (referenced in the proposed section). Peer review by a reputable journal amounts to independent third-party verification and evaluation (soundness, relevance, usefulness etc.).  Clarifying existing explanations serves the general good.  The suggestion of self-promotion is inappropriate, also factually since I am retired, which makes the "publish or perish" incentive irrelevant.  Why an alternative explanation that is simpler and clearer than the traditional ones does not "integrate or support" existing material is not at all obvious.  Evidently, the content can be explained in words, but (not ulike a picture) a formula is worth a thousand words; however, if that is the problem it can be easily mended.  The derivation was step by step for the sake of clarity, but of course it can be made less "lengthy" by skipping steps (the traditional definitions dump everything at once on the reader, mixing concept with one particular application).  In any encyclopedia, clarity is valued; this is precisely what the proposed section offers.  So saying that it "should be obvious" (that the section doesn't fit) is not a convincing argument.  As a demonstration of the fact that the traditional explanations are confused and confusing, I will correct a few mistakes in the existing article; mistakes that would be immediately apparent (hence avoided) with the proposed alternative.  That is the "proof of the pudding" regarding clarity.Boute (talk)


 * Raymond, I downloaded and read your paper. Very interesting, and logical.  But from the point of view of wikipedia policy, it rates at most a brief mention.  If it catches on, or even gets commented on by others, then it might be worth saying more about.


 * In the mean time, you've removed the "dB" off of results such as "30 dB" that were in an equation where your notation would make putting dB there incorrect. Yet these are pretty conventional.  Is there any reason not to leave them alone?  I reverted.  Dicklyon (talk) 04:33, 1 August 2010 (UTC)


 * (a) I respect Wikipedia policy, but the following quotes from the policy statements seem to favor inclusion of the (now deleted) section, possibly in abbreviated form. (i) NPOV: "representing fairly, proportionately, and as far as possible without bias, all significant views that have been published by reliable sources". (ii) Reliable sources: "Academic and peer-reviewed publications are usually the most reliable sources where available". Apart from that, should an encyclopedia not present the clearest explanation possible?


 * (b) Your qualification "pretty conventional" is very appropriate, as it reveals a flaw (one of many) in the conventional formulations. Indeed, in the chained equalities

G_\mathrm{dB} = 10 \log_{10} \bigg(\frac{1000~\mathrm{W}}{1~\mathrm{W}}\bigg) = 30~\mathrm{dB} \, $$
 * the middle expression has undisputably the value 30 (by mathematics). Hence writing dB in the third expression would yield 30 = 30 dB, making dB a trivial identity operator (which is certainly not the intention). Unfortunately, this is just the tip of the iceberg. Boute (talk) 05:40, 1 August 2010 (UTC)
 * I agree with Dicklyon, that inclusion is inappropriate. At most, a brief mention, but even that would seem to be against the consensus here.  WP:NPOV does not support you, the key word here is "proportionately".  Literally tens of thousands of books and papers using the conventional scheme could be cited.  Can you point to even one using yours?  Your paper can hardly be counted a "significant" view: if proportionality is to be taken literally, you get a small fraction of one letter.
 * Your example chain of equalities makes a false argument. First of all because it is common practice to only state the units on the last item in the chain unless the units have been converted midway.  But more than this, I agree that log(1000W/1W) is unitless.  However the coefficient "10" is not unitless - this is carrying the units dB.
 * Finally, I would like to draw your attention to WP:POINT which your most recent edits could be considered to be breaching. Making a point in this way is considered disruptive.  Sp in ni  ng  Spark  07:53, 1 August 2010 (UTC)


 * Talking about "consensus" when the disputed section was removed so fast that only 3 people were able to comment is rather extreme. Also, "proportionately" does not mean word counting, but just staying within reasonable proportions (avoiding disproportionate length).  Similarly, "significant" has nothing to do with numbers of books but with the essence of the content.  Putting "tens of thousands of books" on the scale with your interpretation would prohibit any recently published result from reaching Wikipedia readers for decades.  That cannot be policy.
 * Your interpretation of equality is by no means common practice. Assigning "unit" dB to 10 is very artificial (viewing dB as a unit is even disputable).  Also, what, in your interpretation, are then the units of $$G_{\mathrm{dB}}$$? It is common practice (for the notation at hand) that, if q is a quantity and u is a matching unit, $$q_u$$ is a pure (unitless) number, e.g., for distance, if $$d = 3\,\mathrm{ft}$$, then $$d_{\mathrm{ft}} = 3$$, i.e., the distance in feet is 3, a pure number (not 3 feet).  Hence $$G_{\mathrm{dB}}$$ is unitless.  What would be correct is writing
 * $$G_\mathrm{dB} = 10 \log_{10} \bigg(\frac{1000~\mathrm{W}}{1~\mathrm{W}}\bigg) = 30 \;\mathrm{or, equivalently,}\;G = 30\,\mathrm{dB}$$
 * Finally, the WP:POINT page concerns something totally different, namely attacks on Wikipedia policy by attempting to "discredit the rule or interpretation thereof by, in one's view, enforcing it consistently". The proposed correction of the dB formula is not intended to discredit a policy rule, but is a bona fide, well-founded correction of an error. Boute (talk) 12:01, 1 August 2010 (UTC)

The need for clearer formulations
The purpose of the (now deleted) section "The decibel clarified: serving intuition and rigor" was presenting the clearest possible explanation of the decibel for the Wikipedia readers. This is not a luxury.

Indeed, below we show that the traditional formulations of the decibel (unavoidably reflected in the current article) are thoroughly flawed conceptually, at the expense of intuitive understanding and rigorous formulas and calculation rules. The various points raised in the numerous discussions in the "Discussion" section and the changes in wording and formulas reported in the "History" section would have been very quickly resolved (and most might not even have come up) if a clear and sound formulation were available.

Such a sound formulation would provide a reference frame assisting Wikipedia readers in better understanding even the texts and presentations using traditional formulations. Indeed, the (numerous) flaws and abuses of notation would be readily recognized and mentally corrected. If solely traditional formulations (which also are mutually rather disparate) were available as a background, every flaw would have to be handled separately, without fallback position.

Here we (non-exhaustively) give a few examples of the deficiencies of the traditional formulations, in the order in which they appear in the current Wikipedia article (which is not necessarily the order of their importance).


 * Problems with viewing the decibel as a "logarithmic unit". (a) The logarithmic view is counter-intuitive: if an increase from 10 dB to 13 dB amounts to doubling, then that is exponential behavior, not logarithmic. Enforcing the logarithmic view in formulas restricts the freedom needed for clear formulations (e.g., true equalities instead of vague correspondences). (b) A unit is a reference for comparison in measurement.  For instance, measuring length is comparing the given length to 1 meter. This also holds for dimensionless units, such as the radian for angles.  Saying that dB measurement amounts to comparing to 1 dB (the supposed unit) is hardly meaningful.  If there is anything that can be meaningfully considered a unit in the dB context, it is the reference power or reference voltage (etc.)  in the application to ratios; the resulting measure (ratio) is a pure number.  The  decibel itself is just a nonlinear scaling of that number, meant to cover widely varying orders of magnitude.


 * Problems with "a decibel is one tenth of a bel". Writing this literally as a formula, $$1\,\mathrm{dB} = \frac{1}{10}(1\,\mathrm{B})$$ or, equivalently, $$1\,\mathrm{dB} = (1\,\mathrm{B})/10$$ shows that it cannot be correct. Correct are $$1\,\mathrm{B} = 10\,\mathrm{dB}$$ or, equivalently, $$1\,\mathrm{dB} = (1/10)\,\mathrm{B}$$ but this is "one decibel is one tenth bel" (which sounds strange), not "a decibel is one tenth of a bel".


 * Problems with the decibel as a concept. If a layperson asks in the traditional context "how much is 30 dB" or, less prejudicially, "what is 30 dB" or "what does 30 dB mean", he/she will get a convolved answer with power ratios and in the worst case logarithms.  In the context of a clear formulation, the answer could be simply "1000" or "another way for writing 1000", possibly with the simple generalizing explanation that the 3 refers to the number of zeros.


 * Problems with
 * (quote) Thus, the ratio of a power value P1 to another power value P0, is represented by LdB, that ratio expressed in decibels and is calculated using the formula:



L_\mathrm{dB} = 10 \log_{10} \bigg(\frac{P_1}{P_0}\bigg) \, $$


 * (unquote)
 * As noted below, this is by no means the worst formulation in the literature. However, why the vague wording "is represented by LdB"? Why is it nowhere simply said that $$L = P_1/P_0$$ (as an abbreviation)?  This evasiveness is typical in the traditional formulations.  Yet, stating $$L = P_1/P_0$$ allows rewriting the above formula as



L_\mathrm{dB} = 10 \log_{10} L \, $$


 * with a double advantage: it is a single definition without extraneous variables, and it generalizes the definition of LdB to any positive number L (detached from the narrow application to power ratios).
 * Just as another illustration (showing that it can get much worse than the definition in the article): many authors define the decibel as



\mathrm{dB} = 10 \log_{10} \bigg(\frac{P_1}{P_0}\bigg) \, $$


 * Why this is extremely poor needs (by now) no further explanation.


 * Problems with equalities such as
 * $$G_\mathrm{dB} = 10 \log_{10} \bigg(\frac{1000~\mathrm{W}}{1~\mathrm{W}}\bigg) = 30~\mathrm{dB} \,

$$
 * The middle expression has undisputably the value 30 (by mathematics). Hence writing dB in the third expression would yield 30 = 30 dB, making dB a trivial identity operator (which is certainly not the intention).   What can be made correct (with proper definitions) is
 * $$G_\mathrm{dB} = 10 \log_{10} \bigg(\frac{1000~\mathrm{W}}{1~\mathrm{W}}\bigg) = 30 \;\mathrm{or, equivalently,}\;G = 30\,\mathrm{dB}$$

The advantage of a clearer formulation, if properly designed, is that all these isolated flaws are not "patched" one by one, but avoided right from the start.

Hence my proposal is preparing a considerably shortened version of "The decibel clarified: serving intuition and rigor" under a new title "The decibel: a simplified view". As observed at an earlier occasion, this would be serving the Wikipedia guidelines (i) NPOV: "representing fairly, proportionately, and as far as possible without bias, all significant views that have been published by reliable sources". (ii) Reliable sources: "Academic and peer-reviewed publications are usually the most reliable sources where available".

Before investing the time and work, I would like to have a reasonable assurance that the section would not be deleted until a reasonable number of Wikipedia readers have had the occasion to comment on it. Boute (talk) 14:29, 1 August 2010 (UTC)


 * I too would oppose the reinstatement of this material. However logical it is, it is not a sufficiently widely adopted viewpoint, and that makes all other arguments irrelevant. ...That said, I can't resist a few (irrelevant) observations:
 * In it's original incarnation, as a measure of attenuation, the dB is indeed a unit. You can have a length of waveguide that has 1 dB of attenuation. Ten such lengths give 1 Bel. You can ask how many of these 1dB-yard-sticks are necessary to reproduce a given observed attenuation. Unfortunately the article only discusses 'absolute dB' (relative to a standard reference), which is a secondary sort of dB (I ought to fix this).
 * If you apply dB[x] = 10 log10(x) to an arbitrary +ve number x not a ratio, then there's no expectation that the x's should combine by multiplication, so the result is not necessarily additive, and it may not behave like a quantity with a unit (c.f. 10 log10(p1/p2) + 10 log10(p2/p3) ≡ 10 log10(p1/p3))
 * Boute may be interested in Jaynes application of 'db' [sic - for decibel not deciban] to odds as a measure of evidence or belief. This doesn't seem to have been widely adopted though - it's not even mentioned on the logit page.
 * --catslash (talk) 17:42, 1 August 2010 (UTC)

I carefully checked the rules, and none support the phrase "that makes all other arguments irrelevant". Moreover, it seems that by using the term "viewpoint", we are all barking up the wrong tree: the proposed section is not about a viewpoint but about a simpler explanation. More specifically, reading the rules (and the examples given), especially in WP:NPOV indicate the following
 * The spirit of the rules is avoiding undue weight to crackpot opinions when mentioning them.
 * The term "views" relates more to sociological and philosophical issues, where being "widely adopted" is the only criterion for verifiability and significance. Scientific ideas are more directly verifiable, especially the mathematics.  The "turnaround time" is also much shorter, and recent results (properly verified by peer review) clearly have a place in an encyclopedia.

The most important rule seems to be implied by WP:IAR and the essay What "Ignore all rules" means: making a better free encyclopedia. I am convinced that the proposed section contributes precisely to that goal, by helping the reader with the simplest explanation possible.

This is why all discussion should center on the quality of the proposed section and on helping to improve it, not using a narrow interpretation of some rules to exercise censorship.

Hence I decided to put in a drastically shortened version of the clarifying section (NOT "view"!), kindly asking everyone to be civil about it, not removing it before enough people have been able to see it, and using reader feedback to know which people find it helpful. If after, say, one month the feedback is negative, I remove the section myself. Boute (talk) 07:58, 2 August 2010 (UTC)

The meaning of dB
(header inserted to indicate a new track and facilitate editing) The next item refers to the 07:58, 2 August 2010 note. Boute (talk) 13:15, 3 August 2010 (UTC)


 * I removed it before I noticed your note here. It is still easy for interested other editors to review it and comment here.  I think it was still too non-standard of a notation and approach to be acceptable, and still quite long and complicated.  Dicklyon (talk) 15:20, 2 August 2010 (UTC)

Yes, I was just rearranging the material at the same time, resulting in an editing conflict; as a result of resolving this, the section was reinstated. As said before, we are all barking up the wrong tree by talking about a "different view" and (as above) "nonstandard notation". What the clarifying section really does is giving a more precise meaning to the standard notation (not unlike programming language semantics: it does not change the existing language but makes the meaning precise). It also fiils gaps in a fully compatible way. For instance, the article (without the section) tells nowhere simply and explicitly what 30 dB means. The added section makes clear that 30 dB = 1000. Filling the gaps cannot be done by various patches throughout the article, since other gaps may arise and not all crucial questions can be anticipated. Filling the gaps can only be done by taking a step "back to the basics", supplying a small but solid basis from which the reader can answer many questions himself/herself. All suggestions to improve the section are welcome, and I hope you agree with an "evaluation period" (a reasonable request). Boute (talk) —Preceding undated comment added 15:51, 2 August 2010 (UTC).


 * Sorry, I didn't state my objection very clearly; not a sufficiently widely adopted viewpoint, was not a reference to WP:NPOV, but was meant to be polite for WP:OR (WP:SELFCITING) (which could be taken as meaning something you just made up). If the material does indeed constitute original research, then Wikipedia is not the right place to publish it, and this is the case irrespective of the merit of the material. --catslash (talk) 20:30, 2 August 2010 (UTC)

I do not believe this has a place in WP. Apparently this viewpoint, and that's really all this is, has only been published recently and the attempts to introduce this here appear as an attempt to popularize the idea. The fact that is was published doesn't make it more noteworthy, particularly when the initiative came from its original author. Typically WP bars people from promoting their companies or products here. I see this as no different than creating an article on my own product, or adding information of my product to existing information. It is considered WP:COI and it will get purged. If others pick up on the concept and text books and other secondary source start reporting the idea, only then does WP allow inclusion as a tertiary publication. Many topics of science have found differing mathematical or phenomenological descriptions throughout their history. The point here seems to be that this is easier or clearer for some reason, for which there is no evidence provided nor are there any independent references to support the notion. The presentation appears to redefine the dB as not a unit but a mathematical function, which is in conflict with the definitions of the dB in the standards bodies and the deci- prefix in the International System of Units. Therefore it may be argued that this confuses a general audience even more, if it ever did previously. Frankly, the presentation previously was not confusing, only perhaps by its extensive provision of examples for different measurements. Furthermore, I don't recall ever anyone suggesting they had trouble understanding the traditional definitions of the unit. Please note in particular also the guidelines on WP:SELFCITING. Kbrose (talk) 23:11, 2 August 2010 (UTC)

The proposed section is not a (new, divergent) "viewpoint" in the meaning inferred from the WP guidelines, it is a clarification of existing dB semantics and pragmatics in a systematic way, in the sense that programming language semanics does, without changing the "language". A semantic analysis could at most could be termed a "view". Hence it is not prescriptive, something for others to "pick up" (programming language semantics does not say how to program): it is to enhance understanding, so WP:COI is not relevant. Although I agree that "easier and clearer" is to some extent subjective and therefore always debatable (an easy target, survival depending on fairness), the proposed section contains ample evidence that this is indeed the case: just consider the "what is 30 dB?" question, which in the proposed section is answered right from the start, without obfuscation by power ratios (which is just one very specific use). Just tell me what, according to the original article, 30 dB means: either you will tell me that the article does not provide the answer (revealing a gap), or you will infer from the article exactly the same explanation (x dB = 10^{x/10}) as given in the proposed section. Indeed, this explanation is inescapable: in any amplifier spec sheet, saying that Gain = 30 dB is synonymous with Gain = 1000; hence 30 dB = 1000 (barring obfuscating arguments). This is just one example. Suggesting that the resulting formulations would confuse a "general audience" seems most contrived. Also, defining the dB as a function does not conflict with units or "standards", to the contrary: functions are a well-known way to describe and reason about units, e.g., in dimensional analysis. The same holds for the prefix deci. Finally, as regards WP:SELFCITING, which I read before placing the section, quote: "If an editor has published the results of his or her research in a reliable publication, the editor may cite that source while writing in the third person and complying with our neutrality policy.". By the way, a uniform and consistent way of formulating existing semantics and pragmatics is by definition neutral. Prior publication removes the WP:OR consideration. A final observation: centering all the argumentation for removal of the article on a very narrow interpretation of the guidelines seems in violation of WP:NOT, in particular "Wikipedia is not censored" and "Wikipedia is not a bureaucracy"; quote: "Do not follow an overly strict interpretation of the letter of policy without consideration for the principles of policies.". If there are real arguments why the proposed section does not serve the readers, those should be made explicit, so the section can be improved. Censorship by "purging" does not seem to serve the spirit of WP. Boute (talk) 04:19, 3 August 2010 (UTC)


 * It is not surprising that as a new user you are unfamiliar with how policies, guidelines, and consensus work on WP. This is not censorship.  But also note that WP:IAR is a last resort when the consensus is that ignoring rules will improve the encyclopedia.  That's not the case here.  When you find some support among other editors, or in publications about your idea, there will be something to talk about.  Until then, there's not really much to be done.  I wish you luck promoting your idea, but wikipedia is not the place to try it. Dicklyon (talk) 05:18, 3 August 2010 (UTC)

Thank you for your wishes, but there is no "idea to promote", just providing the simplest explanation of an existing formalism in the interest of the reader. The false impression that it was something else was present from the start and poisoned the entire discussion, despite my clear explanation about semantics.

I am indeed a new user, but I carefully read the guidelines. The principle of consensus is also evident. However, just three people, one of them is so fast in removing the text that no other readers can evaluate it, is far from representative. It gives the impression that some editors think they "own" a specific WP article. Removal ("purge") seems a defensive reaction based on considering any improvement as an implicit criticism of the work invested in an article. To summarize, there are two kinds of vandalism: (a) placing a text whose content is garbage, (b) "purging" a well-founded contribution. Agains (a), "purging" is an effective defense, against (b) there is no defense: with a few keystrokes eliminating the text, anyone can enforce his/her viewpoint and/or their very narrow interpretation of the guidelines. This asymmetry is clearly a flaw in the Wikipedia setup.

Anyway, let us forget this bureaucratic "rule-wielding", and look at the content, as befits scientists. I am interested in your comments on the following questions.

(1) In the current article,


 * (quote) Thus, the ratio of a power value P1 to another power value P0, is represented by LdB, that ratio expressed in decibels and is calculated using the formula:



L_\mathrm{dB} = 10 \log_{10} \bigg(\frac{P_1}{P_0}\bigg) \, $$


 * (unquote)

Why the vague wording "is represented by LdB"? Why is it nowhere simply said that $$L = P_1/P_0$$ (as an abbreviation)? This evasiveness, in fact: obfuscation, is typical in the traditional formulations. Yet, simply stating $$L = P_1/P_0$$ allows rewriting the above formula as

L_\mathrm{dB} = 10 \log_{10} L \, $$

with a double advantage: it is a single definition without extraneous variables, and it generalizes the definition of LdB to any positive number L (detached from the narrow application to power ratios).

The result of this simple step (eliminating obfuscating elements) is precisely one of the equations in the proposed section. What is wrong with presenting the simplest explanation to the readers? Do people like obfuscation? (Note: this question reveals the motivation behind my insistence on inserting "The meaning of dB".)

(2) What, according to the article, is 30 dB? If it is something else than as explained in the proposed section, this would be very interesting. Indeed, the common practice (e.g. in RF spec sheets) of considering Gain = 30 dB synonymous with Gain = 1000 (which provides the most, if not only, consistent semantics) would be invalidated!

I am looking forward to your comments. Boute (talk) 10:10, 3 August 2010 (UTC)


 * If I understand correctly, you:
 * reinterpret "dB" in "x dB" as a function of x rather than as x times a unit
 * invent the inverse function [dB]
 * invent the dashed functions dB' and [dB'] to accomomdate the 20-factor when the argument is not a power-ratio
 * all of which is non-standard, i.e. original research. Wikipedia is therefore not the right place to publish this. --catslash (talk) 12:10, 3 August 2010 (UTC)

According to the definition in WP:OR, prior publication (which has happened) makes it non-OR. As regards your questions, that is itself also a matter of interpretation.
 * Functions are the convenient way to manipulate units in formulas; in fact, the dBmW = d;B;m;W explanation in the (purged) section treats W and dB on equal footing. Whether or not you call something a "unit" does no damage mathematically, nor to the proposed explanation: anyone is free. However, the problem with calling dB a "unit" is the intuition: it feels nor behaves in any way as a unit.
 * An inverse function exists automatically for every injective function (which is the case here). The current WP article also "de facto" uses (without realizing) an inverse function when writing $$L_{\mathrm{dB}}$$ for "L expressed in dB".
 * I presented the dashed dB' and [dB'] as a mnemonic aid, which is useful for anyone to do exploit. As argued below (part postponed by edit conflict) engineers often get the right result by not adhering to flawed definitions. Boute (talk) 12:57, 3 August 2010 (UTC)


 * If 30dB = 1000, then 30dB + 20dB = 30.414dB. This is no the usual way to calculate a link budget. --catslash (talk) 12:20, 3 August 2010 (UTC)

A link budget is calculated by adding numbers in a table. Written in formula form, your example amounts to (30 + 20) dB = 50 dB. (see the note on "patching" below). In fact, the formula


 * Received Power (dBm) = Transmitted Power (dBm) + Gains (dB) − Losses (dB)

from link budget can be considered formally correct if the round parentheses are seen as just another notation for the inverse functions. Boute (talk) 12:57, 3 August 2010 (UTC)


 * You've noted that "This evasiveness, in fact: obfuscation, is typical in the traditional formulations." I agree.  And that's why we follow it in wikipedia:  because it's typical in the traditional formulations.  That's really the only reason.  When some other approach becomes represented in the literature, in more than your one primary article, we can start to move that way.  Dicklyon (talk) 05:00, 4 August 2010 (UTC)

I understand, but the variations in the traditional formulations are large, and some differences are FUNDAMENTAL, as outlined in the topic discussion below. The definition in the article is not representative for all, nor can it be: the variation is too large. However, for this very same reason there is plenty of room for a pure and clear explanation which semantically explains ACTUAL PRACTICE. I carefully checked the relevant Wikipedia guidelines, especially WP:VERIFY, and nothing indicates that for any topic wider representation is needed than a reliable source (peer-review). Especially since we are talking about a DESCRIPTIVE EXPLANATION, which is NOT a prescriptive proposal for anything, the objections are not relevant. See also the quotations from the guidelines in the topic below.

Following the WP:BRD directions, and based on the comments of various editors in this discussion, I will install a thoroughly revised version of "The meaning of dB". It avoids the reference to function composition and inverses which are typical in the literature about semantic formulations but not in the literature using dB. It uses ONLY notations widely represented in this literature. Boute (talk) 10:03, 4 August 2010 (UTC)

Why most definitions of dB are flawed
As an addendum to the 10:10, 3 August 2010 (UTC) note, here are some crucial observations on the definition of the decibel. The definition in the WP decibel article is not in any standard, nor could I find it in any of the books and articles I have on the subject. Hence, by the strict interpretation of the guidelines used to eliminate the section on "The meaning of dB", it has no place in WP, until it can be shown that many texts and articles use it! Although correct, this observation is only meant ironically, and the essential point is the following.

The definition of the decibel is unique in that all its traditional variants, on one hand, clearly seem to be copied from each other (while also reflecting to some extent a common source), whereas, on the other hand, all are formulated somewhat differently, as if the various authors were aware of something being askew and were tinkering with it to get it "right" somehow. The problem is that something is askew indeed. The analysis in shows that there are two main "styles". One is represented in the WP article, which can therefore be quoted literally:



L_\mathrm{dB} = 10 \log_{10} \bigg(\frac{P_1}{P_0}\bigg) \, $$

This is "salvageable" as shown earlier by recognizing that $$L = P_1/P_0$$, leading to $$L_\mathrm{dB} = 10 \log_{10} L$$.

The other style, unfortunately even more prevalent, is



\mathrm{dB} = 10 \log_{10} \bigg(\frac{P_1}{P_0}\bigg) \, $$

Here dB itself appears as a variable, which is useless. For instance, given $$\frac{P_1}{P_0} = 1000$$ it yields $$\mathrm{dB} = 30$$. This is not salvageable.

A further analysis of various sources discussed in shows that, in actual calculations, nobody uses these formulas "literally". What happens instead is the following (depending on the author):
 * Either the flaws in the formulas are tacitly patched up in the course of the calculations to get the right result.
 * Or a correct interpretation of some definition in words is de facto applied, such as "The number of decibels is ten times the logarithm of the given power ratio".

Hence the traditional definitions in formula form serve at best a "decorative" purpose. No wonder no one copies them literally; unfortunately very few get them "right".

A similar problem affects the calculations in formula form. Yet, whereas we could classify the definition styles in two main groups, the calculations in formula form are so diverse (due to the different way of "patching up" as the calculations proceed) that we found no systematic classification. No two authors seem to do their calculations in essentially the same way. Incidentally, none of the calculations suggest that the dB is treated as a "unit". This is also just a word copied by successive generations.

All this shows that there is something askew indeed. Moreover, as a consequence of this discussion in the past few days, and at the explicit suggestion of one of the editors (see above), I contacted a top official in a standardization organization (not further specified here since I am not sure I have permission to do that), and in the reply (received just this morning) it was mentioned that "However I have had an interest in the subject of your paper myself, for indeed I agree with you that the decibel is not well defined at present". Hence there is ample reason for wanting better formulations.

The fact that engineers have been able to overcome the flaws in the manner described above is a tribute to their professionalism, but no excuse for being complacent about those flaws. In fact, the formulation in the proposed section "The meaning of dB" is based on a formalization of correct actual practice in engineering.

As a conclusion, Wikipedia editors owe it to their readers to present the best formulation available at the moment. Boute (talk) 13:03, 3 August 2010 (UTC)


 * Arguing the merits of your view in talk pages on WP is not very productive, it is simply the wrong venue. We don't evaluate the merits of new research results here, no matter how accomplished our editors are. Many do have advanced degrees in all kinds of fields. We publish established and verifiable knowledge, at least we try, that already has found acceptance in the SECONDARY literature. Arguing the virtues of your approach here is simply inappropriate, and this should have been obvious from the beginning for someone with your level of technical understanding. This is why your additions were swiftly removed, not to prevent evaluation or to censor. I would suggest you first convince the standards bodies, industry associations, textbook authors etc, and then someone else undoubtedly will pick up the concept and amend WP articles. Kbrose (talk) 17:15, 3 August 2010 (UTC)

I fully understand your arguments. I even agree with them, up to the point where it prohibits recent results from reaching Wikipedia (the process via standards bodies you suggested is not realistic timewise). For a printed encyclopedia, being years behind is unavoidable, for a digital online version it is ridiculous. The guidelines say nothing about SECONDARY literature. However, all this is not relevant, since the issue is not a diverging view, but a clear explanation of the existing practice around the decibel.

The situation is a question of reader interest versus rules. In WP:BRD it says: "Do not accept "Policy", "consensus", or "procedure" as valid reasons for a revert: These sometimes get worn in on consensus-based wikis." Following worn habits is clearly what is happening here. Note also that "consensus" by definition means that ALL parties agree; as long as the editor whose contribution is purged does not agree, there is no consensus. However, this is secondary. In WP:BRD it also says: "You should not accept, "It's policy, live with it." On the other hand, you should completely understand the implications when someone explains to you, "The flurbeling you suggest caused very bad barbelism, that's why we decided to always floop before we fleep instead.". The latter indicates that discussing the content/merits of a contribution IS the (only) thing to be discussed! Boute (talk) 18:25, 3 August 2010 (UTC)

The current article uses its definition of decibel inconsistently

The defining expression



L_\mathrm{dB} = 10 \log_{10} \bigg(\frac{P_1}{P_0}\bigg) \, $$

yields for $$P_1/P_0 = 1000$$ the result $$L_\mathrm{dB} = 30$$. Observe: NOT 30 dB. The pure 30 is correct.

Later on, the same figures are used in the calculation


 * $$G_\mathrm{dB} = 10 \log_{10} \bigg(\frac{1000~\mathrm{W}}{1~\mathrm{W}}\bigg) = 30\,\mathrm{dB}$$

Now we suddenly have 30 dB, which is incorrect: you cannot have 30 = 30 dB without making dB trivially useless. Boute (talk) —Preceding undated comment added 01:32, 4 August 2010 (UTC).

Addendum to the preceding: the ad hoc "convention" suggested earlier that the units are only added at the end of the line should then also be applied to the $$L_\mathrm{dB}$$-expression, which would not only make it wrong but would cause trouble with

P_1 = 10^\frac{L_\mathrm{dB}}{10} P_0 \, $$ (try substituting 30 dB for $$L_\mathrm{dB}$$).

A final remark: the definition in the article is not in any standard and no reference is given. Given its crucial importance, this should be remedied with the highest priority. Boute (talk) —Preceding undated comment added 07:18, 4 August 2010 (UTC).

The correction by Kbrose eliminates the inconsistency, but creates a new problem: what does $$a \equiv b$$ mean? It is clearly not logical equivalence (the expressions are not boolean). Is it an equivalence relation at all? Boute (talk) 11:50, 4 August 2010 (UTC)

Dear catslash. The added section is NOT prescriptive; it tells no one what to do (per definition, no mathematics is prescriptive). The section is explanatory and descriptive, as it explains the semantics of current practice and of the common elements of the great variety of formulations in the literature. Claiming that the section is something else as an excuse to eliminate it is counterproductive. The Wikipedia guidelines are very clear in WP:OR: once something is published by reliable sources (WP:SOURCES), it is not OR. Moreover, the entire formulation has been changed so that ONLY widespread notation is used. There is no more talk about functions, inverses, composition, so I took into account all your remarks (and sacrificed the elegant math). This shows that I at least have shown willingness to compromise (as the Wiki guidelines suggest). In fact, you removed the section so fast that I am sure you did not even take time to read it. According to WP:BRD, the next step is explaining why your objections still have not been satisfactorily met. In WP:BRD it says: "Do not accept "Policy", "consensus", or "procedure" as valid reasons for a revert: These sometimes get worn in on consensus-based wikis.". The rest of the text in WP:BRD (quoted earlier) clearly indicates that discussion (D phase) must concern CONTENT. I propose that you temporarily revert so you can read the formatted text comfortably. Boute (talk) 12:28, 4 August 2010 (UTC)


 * Dear Boute, Your latest contribution included something like "(x+y)dB = (xdB)(ydB)". This is a function-like use and is incompatible with the standard view of dB as a unit (albeit a dimensionless one like radians). Also included was the square-bracketed and dashed notations [dB], dB' and [dB'] which as far as I am aware are your own invention. It's true that I did not study the material for more than a couple of minutes, but long enough to see these features. I accept that you have published these ideas before and that they are internally self-consistent (though I think they cannot explain why you must sometimes use dB and sometimes dB'). The problem is that they are a different thing to dB as commonly understood. Have you considered adopting a different name or notation (e.g. deBoute x = 10^(x/10) (please excuse my use of your name; I can't think of anything better at the moment)), for this function? --catslash (talk) 13:30, 4 August 2010 (UTC)

Dear catslash. The problem (or, perhaps, feature) of the mathematical concept of function is that, strictly speaking, all units have function-like properties. If you write x ft, then ft unavoidably obeys all the axioms of a function (here a function that maps numbers to lengths), so by definition it "is" a function. Conversely, if you want to highlight the unit-like flavor of dB, you can write $$x\,\mathrm{dB} = (1\,\mathrm{dB})^x$$, showing that it is not a "linear" unit like meter or radian, e.g., $$x\,\mathrm{m} = x \cdot (1\,\mathrm{m})$$. Also, writing dB' is just a way to avoid ambiguity; it is a statement of fact that (as written in the section) "some engineers use it to avoid problems" (this is how, according to Wikipedia, stating opinions is avoided). Using dB' is not an imperative, it is just a convenient way to avoid factors 2 in decibel calculations (powers being proportional to squares of voltages); in fact, it turns a historical accident into an opportunity. Finally, observing that, in RF amplifier specs, one often writes Gain = 30 dB for saying that Gain = 1000, this practice indicates that one considers 30 dB = 1000, leading to the equation x dB = 10^(x/10), so this is the meaning of dB as used in engineering practice (the fact that everything else falls so nicely into place shows that this practice is correctly captured by the equation). So this is not a departure from common conventions, it is just a way of bringing the essence of what engineers always have been doing from the usual obscurity around dB to the surface. This is what formal semantics does for programming languages. I appreciate your questions, they help finding better formulations. Boute (talk) —Preceding undated comment added 15:33, 4 August 2010 (UTC).


 * Boute's continued attempts to insert this material into the article and waste other editor's time, when not only WP rules, but even the consensus is clearly opposed, begin to appear disruptive. It certainly does not improve the encyclopedia as this material does not reflect accepted and verifiable definitions of the subject matter. It appears that no matter what is formulated in the article based on the accepted definition the writer will not be satisfied until his viewpoint is fully presented. WP is not an equal opportunity bulletin board, not a Speaker's Corner. WP replicates only accepted and verifiable content. The standard definition is that bel, dB, neper, rad, and radian are simply other names for the quantity "1".  This is stated explicitly, for example, in the IUPAC Green Book, I believe. This seems to be a core disagreement with this writer who insists that it is a function that can be plotted. —Preceding unsigned comment added by Kbrose (talk • contribs) 17:45, 4 August 2010 (UTC)

Dear Kbrose Please do not be so touchy about my question about your patch. Concernig the claim that an "accepted definition" is used in the article, please provide the fitting reference. I will be perfectly satisfied with any formulation in the article as long as it provides a consistent explanation to the reader. Providing less is vandalism. If "the standard definition is that bel, dB, neper, rad, and radian are simply other names for the quantity "1"" (however improbable that sounds, since that would make all those units equal), something as crucial should be presented (with reference) in the article. As regards functions, you may have missed my earlier remark that all units by definition are functions and hence, being non-pathological ones, can be plotted.

Never underestimate the versatility of functions: they are like chameleons that can blend in any context, also presenting the decibel as a unit. Here is an outline, thanks to the constructive questions by Catslash


 * Linear units: the essence (linearity) is given by


 * $$x\,\mathrm{m} = x \cdot (1\,\mathrm{m}) \mbox{ and } x\,\mathrm{kg} = x \cdot (1\,\mathrm{kg}) \mbox{ and } x\,\mathrm{s} = x \cdot (1\,\mathrm{s})$$


 * What 1 m and 1 kg and 1 s are is by convention some physical reference object (originally) or some more precise modern variant.


 * Logarithmic units (what's in a name?): the essence is given by


 * $$x\,\mathrm{Np} = (1\,\mathrm{Np})^x \mbox{ and } x\,\mathrm{B} = (1\,\mathrm{B})^x \mbox{ and } x\,\mathrm{dB} = (1\,\mathrm{dB})^x$$


 * What 1 Np (neper) and 1 B and 1 dB are is by convention


 * $$1\,\mathrm{Np} = e \mbox{ and } 1\,\mathrm{B} = 10 \mbox{ and } 1\,\mathrm{dB} = (1/10)\,\mathrm{B}$$

So, if one is willing to discuss content, we can make real progress and not waste anyone's time. Boute (talk) 19:05, 4 August 2010 (UTC)

Dear Kbrose I inadvertently undid your revert since my edit window was still open. It was not my purpose to skip the D in the BRD cycle. Note: any direct logical consequence of the standard falls within the standard. From WP:SELFCITE "Self-published material may in some circumstances be acceptable when produced by an established expert on the topic of the article whose work in the relevant field has previously been published by reliable third-party publications. Anyway, would removal of the citation eliminate your objections?Boute (talk) 04:51, 5 August 2010 (UTC)
 * " From WP:NOR: "If an editor has published the results of his or her research in a reliable publication, the editor may cite that source while writing in the third person and complying with our neutrality policy." - remark by 94.226.127.101, 2:37 Aug 5 2010

Dear Dicklyon When working with systems of units, the "moving to the other side" is standard practice in physics and engineering. For instance, the statements
 * The length in meters is 3.
 * The length is 3 meters.

are equivalent, which in formula form is written as
 * $$\ell_\mathrm{m} = 3 \Leftrightarrow \ell = 3\,\mathrm{m}$$

Since the "standard" (in the sense of the article) calls dB a "unit",
 * $$L_\mathrm{dB} = x \Leftrightarrow L = x\,\mathrm{dB}$$

Why call such logical consequences nonstandard? Anyway, science is about logically or experimentally verifiable facts, not standards. Boute (talk) 05:46, 5 August 2010 (UTC)

Afterthougt: if a spec sheet states "Gain = 30 dB", what, according to the "standard interpretation" is the value of Gain in ordinary numbers? Why say that the (only) answer is a nonstandard interpretation? Boute (talk) 06:44, 5 August 2010 (UTC)

In my experience, the decibel is heavily used in electrical engineering, and quite a few electrical engineering textbooks are not mathematically rigorous when it comes to notation and units. A text that is not only more rigorous, but recognized by the U.S. Secretary of Commerce as the legal interpretation of SI for the United States is NIST Special Publication 330. It says on page 35 The bel and the decibel, B and dB, where 1 dB = (1/10) B, are used to express the values of logarithmic ratio quantities whose numerical values are based on the decadic logarithm, lg = log10. The way in which these units are interpreted is described in footnotes (g) and (h) of Table 8. The numerical values of these units are rarely required. The units neper, bel, and decibel have been accepted by the CIPM for use with the International System, but are not considered as SI units.

The footnotes g of of table 8 relates to Nepers. The footnote h on page 36 states (h) The statement LX = m dB = (m/10) B (where m is a number) is interpreted to mean that lg(X/X0) = m/10. Thus when LX = 1 B, X/X0 = 10, and when LX = 1 dB, X/X0 = 101/10. If X denotes a mean square signal or power-like quantity, LX is called a power level referred to X0. Jc3s5h (talk) 16:17, 5 August 2010 (UTC)

Thank you for the reference, which is really interesting. I downloaded it, but found little more than what you quoted. Yet, whereas it may be true that "the numerical values of these units are rarely required", they just might have given them as a clarification. It is not even clear what they mean, perhaps 1 Np = e and 1 dB = 10^{1/10}. Also, the statement "LX = m dB = (m/10) B (where m is a number) is interpreted to mean that lg(X/X0) = m/10" is not the clearest one I have seen, to put it mildly. It involves an undefined L and an undefined (because being defined) dB. From an abstract view, leaving L undefined is workable, but why not simply define LX = X/X0? Boute (talk) 19:23, 5 August 2010 (UTC)
 * I believe the intent of the SP-330 text is that the reference power level is given as a subscript of L. For example, if one wishes to express a power of 1 W using 1 mW as the reference level,
 * LmW = 10 lg (1000 mW/ 1 mW) dB = 30 dB
 * The information about the reference level is contained in the subscript of L, and the method of calculation is conveyed by "dB" which is treated as a unit, even though the interpretation is more complex than other units, such as watt or meter.


 * I notice this is not always observed, and indeed, can be rather awkward outside the context of an equation. For example, the August issue of QST reviews an accessory for an Elecraft tranceiver on page 48, and in a chart, shows "Sensitivity: −144 dBm". Since there is no equation, and no variable, there is no opportunity to indicate the reference level with a subscript, and instead the author, Joel Hallas, used a unit, "dBm", which seems to not be contemplated at all by NIST or BIPM; they don't even on the list of units that are not recommended. I suppose Mr. Hallas could have come closer to the NIST/BIPM recommendation by writing "SensitivitymW: −144 dB".Jc3s5h (talk) 20:12, 5 August 2010 (UTC)


 * The SP-300 text is very interesting, yet since it is rather obscure it needs considerable explanation (see below). Googling ""receiver sensitivity" definition" yields many good explanations; I cannot say yet which is the clearest. I also noticed in the QST article about the Elecraft K144XV that the sensitivity for FM is specified in microvolt. Boute (talk) 08:34, 6 August 2010 (UTC)

Something missing in the acoustics definition
This is missing important specifics; previous commenters have said this but didn't exactly how to say it.

This article makes it sound like decibels in acoustics are just an abstract ratio to some variable reference. In reality there are specific defintions that include the reference and (sometimes) specific weighting curves. I don't have the sources handy to just write it in but such an addition of specifics would be good.

North8000 (talk) 16:47, 5 August 2010 (UTC)
 * But acoustic decibels are an abstract ratio to a(n agreed) arbitratry reference. And the "acoustics" section mentions frequency weighting and points you at the sound pressure level article which is where reference levels and frequency weighting should be explained. This is why we have links and other articles, otherwise we'd only have one article called The sum total of human knowledge to date which would break older browsers when loading. --Wtshymanski (talk) 18:42, 5 August 2010 (UTC)


 * There is a defined reference for 0dB in acoustics. If an article is going to have 2 paragraphs on accoustics dB, I would think job one would be to define the reference.  I found it and put it in, with a cite/reference. Hopefully .0002 micro bars jives with the 20 micro Pascals mentioned in the other WP articles. North8000 (talk) 03:21, 6 August 2010 (UTC)


 * Perhaps the acoustics definition deserves a more than proportional coverage in WP since (as much as we technical people might regret it) this is what the VAST majority of Wikipedia readers will look for. There is work to do. I am not (yet) volunteering. Boute (talk) 08:45, 6 August 2010 (UTC)


 * Agreed North8000 (talk) 09:11, 6 August 2010 (UTC)

On explaining definitions in the simplest way (essay)
Explanation of a standard definition The short footnote in Special Publication 330 (which is all it says about dB) is very revealing, and also the most rigorous among the standards that I have seen. It says
 * LX = m dB = (m/10) B (where m is a number) is interpreted to mean that lg(X/X0) = m/10.

What this rather cryptic definition says is requires some explanation.

The variables are X and m, whereas X0 is a constant; we assume X and X0 have the same sign, let's say, positive. Assume that LX and m dB are meant to be unambiguous which, by definition, allows to use the word "function" for both L and dB (which does not conflict with seeing it as a unit). Omitting "= (m/10) B", and writing L(X) instead of LX just to avoid small print later, we have


 * $$\mathrm{L}(X) = m\,\mathrm{dB} \Leftrightarrow \mathrm{lg}(X/X_0) = m/10$$ call this (1)

Both L and dB are (still) undefined. Without saying or knowing what the expressions L(X) and m dB mean, their values are equal if and only if lg(X/X0) = m/10 (the RHS, "right hand side" if (1)). In other words, given X, the RHS says how to calculate the m which allows writing L(X) = m dB. The same holds for calculating the X for given m. You can see the expression "L(X) = m dB" as a dB calculator with two dials, labeled L and dB, whose inner workings is described by the RHS. Setting X on the L-dial moves the dB dial to the m that satisfies the RHS and vice versa (technically: the two-directional mechanical interpretation is based on injectivity of the functions involved).

Mathematically, this means that for every L, however silly (but, as seen later, injective, i.e., L(X) should not have the same value for two different X), there is a corresponding dB (function, unit) that satisfies (1). To see this, simply define, for given L,


 * $$m\,\mathrm{dB} = \mathrm{L}(X_0 \cdot 10^{m/10})$$ call this (2)

This satisfies (1), with the RHS rewritten as $$X = X_0 \cdot 10^{m/10}$$, provided


 * $$\mathrm{L}(X) = \mathrm{L}(X_0 \cdot 10^{m/10}) \Leftrightarrow X = X_0 \cdot 10^{m/10}$$

This means that injectivity of L (at least for positive arguments) is the only requirement. An exotic L defines an equally exotic dB. For instance, if we let L(X) = 3X/X0 + 27 then, assuming m = 20, we get from (2) that 20 dB = 327. Yet, this causes no errors since L(X) = 327 yields X/X0 = 100, as it should (the dB calculator at work). To be precise, any such dB should actually be written dBL since it depends on L.

The equality L(X) = m dB allows saying only that X/X0 "corresponds to" (rather than "equals") m dB, which explains the usual vague wordings.

Convolved and obfuscating? Indeed, but that is what the standard says!

One might think that the ability to choose different L (resulting in different dBL) yields some flexibility, but this is useless: no matter how exotic L and dBL, they can provide no more information than the RHS of (1).

Yet, this freedom allows using the standard to our advantage by choosing the injective L that is the simplest possible (up to scaling with X0), which is L(X) = X/X0. Then (2) becomes (writing dB for this dBL):


 * $$m\,\mathrm{dB} = 10^{m/10}$$

One could/should have said this right from the start. No nonsense (sorry!), no obfuscation, no L, no X0. General, since it involves no power ratios. Yet still a "standard interpretation"! —Preceding unsigned comment added by Boute (talk • contribs) 09:46, 6 August 2010 (UTC)

Implications for the current article In the notation of the article, the L in the article is not the L above, but it is itself a variable that "corresponds" to $$P_1/P_0$$ via some not explicitly given function similar to the L above. Yet, as above, this gives the freedom to let $$L = P_1/P_0$$. This justifies adding the following explanation


 * (quote)
 * Finally, letting $$\,L = P_1/P_0$$, we also have $$L_\mathrm{dB} = 10 \log_{10} L\,$$ (for any positive L) or, in dual form, $$L = (10 \log_{10} L)\,\mathrm{dB}$$. This yields the direct formula


 * $$x\,\mathrm{dB} = 10^{x/10} \mbox{ and, correspondingly, } \frac{P_1}{P_0} = (10 \cdot \mathrm{log}_{10} \frac{P_1}{P_0})\,\mathrm{dB}$$


 * Examples: 0 dB = 1, 3 dB = 2, 10 dB = 10, 20 dB = 100, 30 dB = 1000. This also matches universal practice in engineering where, on specification sheets, statements like "Gain = 30 dB" are used as synonym for "Gain = 1000", which implies that 30 dB = 1000 as the above formula indicates.
 * (unquote)

I propose this here first for evaluation rather than placing it directly in the article and having it removed before others can evaluate it.

Remarks:


 * The step to "dual form" is self-explanatary to most technical people, and laypersons will skip it. Indeed, when working with systems of units, the "moving to the other side is standard practice.  For instance, the statements
 * The length in meters is 3
 * The length is 3 meters
 * are equivalent, which in formula form is written as
 * $$\ell_\mathrm{m} = 3 \Leftrightarrow \ell = 3\,\mathrm{m}$$
 * Since the "standard" (in the sense of the article) calls dB a "unit",
 * $$L_\mathrm{dB} = x \Leftrightarrow L = x\,\mathrm{dB}$$


 * Especially in a crucial matter like this, EVERY important formula should have a reference, and certainly the definition in the article. However, for the proposed explanation the situation has proven sensitive, since it would be a publication by myself.
 * - On one hand, to this publication, the following applies. WP:NOR clearly states: "In general, the most reliable sources are peer-reviewed journals and books published by university presses; university-level textbooks; magazines, journals, and books published by respected publishing houses" and also "If an editor has published the results of his or her research in a reliable publication, the editor may cite that source while writing in the third person and complying with our neutrality policy". Furthermore, the description of primary and secondary sources in WP:NOR clearly concerns "events" with possible "eyewitnesses".  For scientific material, which inherently is more objective (unless there is falsification of results), it is not sensible to insist on secondary sources if a primary "most reliable source" is available.
 * - On the other hand, given the sensitivity of the situation, a reference can be omitted on the grounds that it is a direct logical consequence of various standards.


 * Note that the current definition in the article does not accurately quote any source or standard, so it is also an explanation (like the proposed addition). In fact, it is better than most standards I downloaded, which shows that other editors see ample reason for "deviating" from "the" standard (which one, anyway).

Concluding remarks Is the small addition I am proposing worth all this discussion and fuss? First of all, I did not start the fuss but, apart from that, YES, it is worth it. The proposed clarification What is my own stake in this? I am "just a little bit" allergic to obfuscation, vagueness etc. in technical writings, especially when a pure and simple alternative is known. A good encyclopedia must give the clearest explanation possible to its readers. Boute (talk) 15:57, 6 August 2010 (UTC)
 * provides the layperson (for a change) with a pure and simple explanation of what dB means, without extraneous factors such as power ratios (which are just one particular application, and hence have no place in a proper definition), without other obfuscating elements or vague wordings like "corresponds to" or "is represented by" etc.. Via 30 dB = 1000, the decibel can be explained as specifying just a number of zeros. The math for fractional values can come later if further questions are asked.
 * provides engineers with a rigorous basis for (algebraic, symbolic) calculation with dB in the same way as with "ordinary" mathematical functions (engineers like to do symbolic calculation by chaining equalities, not involving "corresponds to" with hidden semantics).


 * Have you ever wondered why you're writing a textbook to define decibels when the rest of the world is quite happy with one-line definitions? --Wtshymanski (talk) 13:57, 6 August 2010 (UTC)


 * Here is the resulting 1/10 line definition: $$x\,\mathrm{dB} = 10^{x/10}$$. All the rest is just to show it follows from the standard (which is very indirect). Boute (talk) 15:57, 6 August 2010 (UTC)


 * Please stop altering your posts after others have replied to them, the is considered a bad practice here, see WP:TALK. You may strike (not delete) a comment you no longer wish to make.  If you wish to amend a comment it is best to make a new post at the bottom explaining this.  I realise that you want your "essay" to be perfect, but talk pages are really not the place for essays.  If you feel compelled to write one, then do it in your own userspace, not on the article talk pages, they are not designed for this.  Sp in ni  ng  Spark  11:01, 7 August 2010 (UTC)


 * Since WP:TALK mentions "substantial changes" (possibly altering the context of a reply), I thought minor improvements unharmful. However, you are right, such alterations may be confusing and a strict, uncompromising discipline to avoid all forms of changes is preferable (and I will adhere to that).  I was also unhappy about the length of the essay and its placement in the talk page.  Are there guidelines for moving it to my own userspace at this stage? Since there is already one reply before yours, I suspect it suffices asking this person's permission, which is my next step. Do you agree with the move? Boute (talk) 11:45, 7 August 2010 (UTC)


 * In the context of talk pages, any change that is not a simple spelling correction, fixing malformed link or the like should be considered substantial for the purposes of the guideline. Rather than the complication of trying to unpick this talk page, I suggest you merely copy the text you want into your userspace and then we can archive the talk page leaving a nice blank page on which you can place a link to your essay.
 * By the way, another talk page convention you are not following is to indent your posts according to the post you are replying to.  Sp in ni ng  Spark  12:51, 7 August 2010 (UTC)


 * Another misunderstanding resolved about the purpose of indentation. I corrected the indentations above, which falls within your definition of insubstantial change (but alas violates my self-imposed discipline - last time!). Feel free to remove any or all my preceding texts from this talk page as needed for cleaning up, I have everything archived. I will put a shorter synthesis of the essay and other pieces in my own user space a little later. Boute (talk) 14:03, 7 August 2010 (UTC)

Definition of decibel
The decibel is not "one-tenth the logarithm to base 10 of the ratio of two power quantities". The decibel is (to quote Merriam-Webster's), "10 times the common logarithm of this ratio [of amounts of signal power]". So it should read "ten times the logarithm to base 10 of the ratio of two power quantities"; but since someone was quoting an IEEE dictionary (good), I leave to someone with the dictionary to make it right. A bel (a unit I've never heard of) is apparently defined as the base-10 logarithm of the ratio of the power quantities, and I suppose that is what is confusing people. You multiply by 10 to go from bels to decibels, but that does not mean that 10 bels is 1 decibel anymore than the fact that to convert from meters to decimeters one multiplies by 10 means that 10 meters is one decimeter. People are confusing units with the corresponding functions sending measurables to their measured values, which are dual concepts. Stephen A. Meigs (talk) 00:46, 11 August 2010 (UTC)
 * Done. Also it said the decibel is rather than the number of decibels is, which is just inviting the wrath of Boute. --catslash (talk) 01:05, 11 August 2010 (UTC)

No, the definition was correct. 1 bel is the lg of the power ratio, so 1 decibel = 0.1 Bel = 1/10 the log. Meriam-Webster is wrong. You are confusing this with the value in decibels, which is of course 10 times the log. Kbrose (talk) 01:47, 11 August 2010 (UTC)


 * While I agree that 1 decibel is 1/10 of 1 Bel, this definition is highly unconventional. Do you have a source for it?  You too are confusing the unit with the value, since the log of the power ratio is the value.  Dicklyon (talk) 03:26, 11 August 2010 (UTC)


 * Right, and if one says that the decibel is "one-tenth the logarithm to base 10 of the ratio of two power quantities" (most) people will think that one is informally defining the decibel by declaring what the value of the decibel is. (Most) people would "not" think that one is first informally defining the bel unit by declaring its value (namely the logarithm to base 10 of the ratio of two power quantities) and then multiplying that unit by 1/10. Such informal abuses of language would be rather natural where the verb of the sentence is, but not in the middle of what looks like an equation. The situation is akin to that between a vector space and its dual. The unit bel is more like the linear functional that sends a function obtained as the logarithm of the ratio of a power quantity with a reference quantity to its value at 10 times the reference quantity (the quantity that gives one bel); it's more of an evaluation function. To be perfectly clear, say either "the decibel is a unit whose value is ten times the logarithm to base 10 of the ratio of two power quantities" or say "the decibel is a unit defined as 1/10 the unit whose value is the logarithm to base 10 of the ratio of two power quantities". How about the following as something one could agree on? "The decibel is a unit whose value is ten times the logarithm to base 10 of the ratio of two power quantities. The actual base 10 logarithm of the ratio itself is called a bel. Though less simply defined, the smaller decibel unit has proven to be more popular than the bel unit. Since the value of a measurement using decibels is 10 times that of a measurement using bels, the decibel (as a unit) equals 1/10 bel, whence the "deci" prefix."Stephen A. Meigs (talk) 16:42, 11 August 2010 (UTC)


 * When you boil it down, there are two main uses of the term, and they are fundamentally different:
 * A way to express the DIFFERENCE between two different values.
 * A way to express sound level in air.  (which, secondarily,  has a difference-based definition behind it) North8000 (talk) 11:23, 11 August 2010 (UTC)
 * Yes, and the primary use - to quantify gain, attenuation, reflection, signal/-to-noise - isn't discussed. It should be the first thing mentioned. I keep meaning to fix this, but feel free to get there before I do. --catslash (talk) 12:04, 11 August 2010 (UTC)


 * First of all, I sincerely regret my excessive zeal in the past, causing this talk page to become overcrowded, and I appreciate the tidying up using the archive.


 * Probably most of the above questions are implicitly answered by a very clear explanation in NIST Guide for the Use of the International System of Units, page 29. This document clarifies the more terse definitions (actually just footnotes) in NIST International System of Units and in BIPM The International System of Units (SI)}, 8th edition, which accommodate various interpretations with the same "end result".  The  NIST document, page 29 and other standardization documents indicate that, under the conventional interpretation, the decibel is not a unit for expressing the considered power ratio itself, but for expressing a derived quantity, called the level difference which, expressed in dB, is 10 times the base 10 logarithm of that power ratio (and this assumes a linear scale).  The same level difference expressed in Np (neper) is half the natural logarithm of that power ratio.  Incidentally, the preceding BIPM document states on page 127 "some scientists consider that [bel, neper etc.] should not even be called units", perhaps because of the possible confusion.


 * Anyway, the above shows that it is correct writing 1 B = 10 dB and 1 Np = (20 lg e) dB or approximately 8.6858896381 dB. Also, since the decibel is used for many different quantities (even temperatures and frequencies),  power ratios are just one of many applications, and their place in a definition of the decibel itself may be questioned (at best, it clarifies the historical origin). For those of you who are not put off by some math, I just finished a note "The meaning of the decibel" (available for the asking). Boute (talk) 13:05, 13 August 2010 (UTC)
 * It is correct to say that 1 B = 10 dB. The relationship between nepers and bels (or decibels) is more controversial because the decibel is a measure of power ratio, whearas the neper measures amplitude ratio; safer to avoid. Dondervogel 2 (talk) 15:38, 13 August 2010 (UTC)
 * Can you give some references to where the controversy and the unsafe aspects are discussed? This may provide some insight. Boute (talk) 02:12, 14 August 2010 (UTC)
 * I remember an article from about 5 years ago; I did not understand all the details, but I do recall that the precise relationship depends on whether the neper is defined in terms of a ratio of amplitudes or (square root of) a power ratio; I will look up the reference for you. Dondervogel 2 (talk) 16:05, 16 August 2010 (UTC)
 * I found this on the BIPM web site, which says cautiously (see p128) "The numerical values of the neper, bel, and decibel (and hence the relation of the bel and the decibel to the neper) are rarely required. They depend on the way in which the logarithmic quantities are defined." More tomorrow. Dondervogel 2 (talk) 17:22, 16 August 2010 (UTC)
 * Thank you for the reference. You may wish correcting .pd to .pdf to make the above link work.  It refers to the same BIPM document giving (in my opinion) the "clearest" decibel definition published by standards bodies.  By some further formalization of the meaning of the decibel, I found that, for the conventional interpretation, there is no a priori reason for fixing any relationship between neper and decibel: both can be defined in a fully internally consistent way without a mutual link, just as two sheets of logarithmic paper from a different printer need not have the same distance between the 1-10-100 lines.  From this viewpoint, any relationship between decibel and neper can be established by convention, but some provide more "interesting" properties than others. Boute (talk) 08:11, 17 August 2010 (UTC)
 * I found the article I was looking for: On logarithmic ratio quantities and their units, Metrologia 42 (2005) 246-252, by Ian Mills and Christopher Morfey. In my recollection I was wrong about one detail, which is that it is the bel and not the neper that they provide two definitions for.  In one of the definitions they are related by 1 Np = 1 B; in the other by 1 Np = 20lge B.  When you say "both can be defined in a fully internally consistent way without a mutual link" you hit the nail on the head. Dondervogel 2 (talk) 15:41, 18 August 2010 (UTC)
 * The eighth edition of the the SI brochure, cited above (as 'this') is indeed cautious. The seventh edition defined the bel in terms of the neper. Alma Teao Wilson (talk) 21:29, 3 September 2010 (UTC)

Recent edits and reverts between two version
The two version are off from each other by a factor of 100. One would think that we could figure out which one is right! :-) North8000 (talk) 14:43, 11 August 2010 (UTC)


 * They both say the same thing, just very differently. There's not a question of a factor of 100, but of how to express the relationship.  I think that maybe the one that W has sourced to the IEEE may have been mangled along the way, so I've asked him what the source actually says.  Dicklyon (talk) 15:14, 11 August 2010 (UTC)
 * I screwed up; trying to paraphrase a very short definition which on page 288 of the "Authoritative Dictionary" says exactly "Ten times the logarithm to base 10 of a ratio of two powers". How's this current wording? I'd like to get away from the idea that there's a platinum-iridium standard decibel in a vault in Sevres - it's not a unit like a foot or meter, it's a ratio. --Wtshymanski (talk) 15:27, 11 August 2010 (UTC)
 * I suppose all measurements are ratios...but the dB is a ratio that is very lightly tethered to any particular reference...--Wtshymanski (talk) 16:42, 11 August 2010 (UTC)
 * Except in acoustics. North8000 (talk) 20:02, 13 August 2010 (UTC)
 * Well, yes, in the sense that a bushel of apples vs a bushel of oats vs. a bushel of coal are also tied to their references. Or French, German, British and boiler horsepower. Within one field, it's well-understood what the unit is; but problems rise when the same word has different denotations in different contexts.
 * And what's with calling me "W" all over the place? People who's user names aren't even in Roman letters get called by their user name...and all I get is a letter. --Wtshymanski (talk) 20:16, 13 August 2010 (UTC)


 * What I meant is that in acoustics, it's very common to use a standard reference, so standard that it doesn't even need mention. E.G.  "The sound level is 90 dB"
 * Elsewhere, such is rare. And often it specifies a change without even talking about values.  (e.g. the gain of an amplifier) North8000 (talk) 21:12, 13 August 2010 (UTC)
 * In air? Or in water? Spectrally weighted? Using a decibel is a bit like using "degrees" - you have to give the reference level. --Wtshymanski (talk) 21:29, 13 August 2010 (UTC)
 * Answering your questions, the overwhelmingly common meaning is 1.  In Air  2.  weighting is left ambiguous to start,   But always the same 0db standard.  So, even though the "implied standard" is sloppy, there still is one.  Degrees is a little different because there are are three commonly used definitions which massively differ from each other (C,F,K) vs only one for sound level.  The latter is based on assuming weighting is secondary, and that it's being used where a sound level value is implied, e.g. if you're talking sound, "80 db" implies a sound level value, but add a word e.g. : "80 db attenuation" and then it no longer does.  North8000 (talk) 21:53, 13 August 2010 (UTC)


 * Most of the acoustics questions may be cleared up by looking at the Wikipedia article Sound Pressure or, in view of the amount of discussion there, a more definitive text such as Allan D. Pierce, Acoustics - An Introduction to Its Physical Principles and Applications, section 2.3 "Levels and the decibel" pages 60-65. Boute (talk) 06:43, 15 August 2010 (UTC)
 * I agree with Wtshymanski. Any use of "dB" without statement of reference level is ambiguous, unless reporting a dimensionless ratio. Dondervogel 2 (talk) 08:53, 11 October 2010 (UTC)

what is a field quantity?
Everybody uses it as if everybody knew what it meant. 91.7.152.234 (talk) 20:30, 7 October 2010 (UTC)


 * In the context of "logarithmic quantities", the term field quantity is traditionally used for a quantity whose square is proportional to power. Examples are: amplitude (for signals), electric field strength, pressure (in acoustics).  However, the fairly recent ISO 80000-1 International Standard on Quantities and Units (2009-11-15), Part 1: General, Annex C, mentions that the term field quantity is more often used to designate a quantity that depends on the position vector, and recommends using the term root-power quantity for the aforementioned concept in the context of "logarithmic quantities". Boute (talk) 07:59, 8 October 2010 (UTC)

Language consistency
Since when is use of a single language consistently through an article considered "unconstructive"? Dondervogel 2 (talk) 06:50, 15 October 2010 (UTC)
 * Agreed, I have restored your edits. WP:RETAIN states When an article has evolved sufficiently for it to be clear which variety it employs, the whole article should continue to conform to that variety, unless there are reasons for changing it based on strong national ties to the topic. When an article has not yet evolved to that point, the variety chosen by the first major contributor should be adopted.  The first edit to clearly disambiguate the version of English was this edit by Omegatron in June 2004 - not the spelling of honour.  This was itself a merge from dB, the disambiguating edit from Heron occuring in July 2002.  Pretty much conclusive that this article has been Brit Eng since ancient history.  Sp in ni  ng  Spark  15:02, 15 October 2010 (UTC)
 * To Wtshymanski: Sure nobody says "fyb-ree" but the word is spelt "fibre". Dondervogel 2 (talk) 15:32, 15 October 2010 (UTC)
 * But the article title is Optical fiber. It seems wilfully perverse to reverse the letters. Surely the world isn't so illiterate that changing "er" and "re" obscures the meaning; anyone who's been using the Web more than 15 minutes knows there's more than one corect spelling. Come to Canada, where we ignore which way "er" goes, are laid back about which words get a surplus "u". Have you read The Mother Tongue (book)? English spelling is so arbitrary anyway, why swap letters when there's an encyclopedia to be written? --Wtshymanski (talk) 18:37, 15 October 2010 (UTC)


 * Yes, and Wtshymanski's edit is completely against the spirit of WP:ENGVAR which requires a consistent spelling system within an article.  Sp in ni ng  Spark  15:47, 15 October 2010 (UTC)

It's not clear why we chose to prefer to move to UK, rather US, when it was about half and half, but I'm OK either way. Dicklyon (talk) 18:07, 15 October 2010 (UTC)
 * How about: Bell Labs is an American organization, Bell was an American, this article should use American spelling. We should worry less about Indian English vs American English and more about making sure spellings appear in at least one reliable dictionary. --Wtshymanski (talk) 18:45, 15 October 2010 (UTC)
 * Dicklyon, I thought I was quite clear above, WP:RETAIN falls on the side of the original spelling system of the article or the earliest disambiguation. That is unarguably Briteng if you follow the history links I provided.  Wtshymanski, that is a tenuous argument, sure Bell is clearly American, but the unit decibel is internationally used, it cannot be considered an American subject.  Sp in ni  ng  Spark  18:55, 15 October 2010 (UTC)


 * The 2002 edit by Heron includes the word "fiber" and the 2004 edit by Omegatron includes "meter". So you've got your facts wrong on the while disambiguation thing.  Both of them put it into a mixed state. Dicklyon (talk) 01:42, 16 October 2010 (UTC)


 * But it's not even an SI "unit", as Wiki scholars are quick to tell us. --Wtshymanski (talk) 19:18, 15 October 2010 (UTC)
 * That is not germane to the matter. Why is whether or not it is recognised by SI relevant?  It does not stop it being internationally used.  Sp in ni  ng  Spark  19:23, 15 October 2010 (UTC)
 * But there's an entire phylum/clan/sept of Wiki editors whose greatest contribution to the project is converting units to SI for the benefit of "international" readers. Surely this is even more important than the order of ER and RE. I propose we approach the developers so that we could use markup like "fib[er|re|ur]" or "met[re|er|ar]" and the reader will be served up the word in his favo[|u]ite misspelling. Clearly the article has reached technical perfection and this is the sole remaining issue preventing from being a Featured Article full of Brilliant Prose. --Wtshymanski (talk) 20:23, 15 October 2010 (UTC)

Sorry, I had left it in an inconsistent state, too. So, I looked through all the recent versions, for any that did not mix fiber and fibre or meter and metre or fiber and metre, etc. The only one I found was my US version. If there's any evidence that there was ever a consistently UK version, I can't find it. Someone can change back to that if they find the evidence. Dicklyon (talk) 07:23, 16 October 2010 (UTC)
 * Just to be clear about my view on this. It does not matter an iota to me whether this article is in UK or US English.  I chose UK because it seemed mostly written that way at the time of my edit.  I can imagine there are good arguments both ways, but the mixture looked amateurish to me, which I assume is not the intention. Dondervogel 2 (talk) 09:27, 19 October 2010 (UTC)
 * Same here. If your version hadn't left it with mixed "fibre" and "fiber", I probably would have gone back to it.  As far as I can tell, mine if the first consistent version; but if someone finds I didn't make it consistent, they can fix it any way they like. Dicklyon (talk) 04:25, 20 October 2010 (UTC)

Power/Intensity v. Pressure
Hi: Can we change the representation under the "Power" section to I, I_0, etc., for intensity? P seems like pressure. It's a bit confusing because if you are calculating SPL using pressure, the coefficient is 20. Maybe we could add the SPL(pressure) function along with the SPL(power) function just for clarity? I am new here so I'll let y'all take it from here. 128.31.35.57 (talk) 06:09, 12 December 2010 (UTC)

I is a bad idea, it is universally used for electric current and so has vastly more scope for confusion.  Sp in ni ng  Spark  17:44, 21 December 2010 (UTC)