Talk:Decibel

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Dubious[edit]

There is a Dubious claim on the use of dBmV in cable television systems, and the use of 75 ohm cable. The use of 75 ohm cable for CATV systems is so widespread that I don't see anything dubious. In days past, it was not so unusual for TV antenna inputs to be balanced 300 ohm twin lead, that is rare. The impedance of a folded half-wave dipole antenna is close to 300 ohms, but CATV only uses coaxial cable. Gah4 (talk) 17:41, 10 October 2016 (UTC)

There is also Dubious related to nuclear hardness. I presume this is related to hardening of missile silos, and not a property of nuclear physics. I can't say much about the dubiousness in this case. Gah4 (talk) 17:41, 10 October 2016 (UTC)

OK, I removed: {{Dubious|date=March 2016|reason=if it's relative to a reference voltage, the impedance should be irrelevant. If it's a power measurement misnamed as a voltage measurement, this should be clearly specified.}} dB is always a relative power measurement, but voltmeters measure voltage. As TV cable is reliably 75 ohms, this isn't a problem. In other cases, it is much less obvious. It seems that audio uses a 600 ohm reference, as far as I know, even when it isn't actually 600 ohms. At audio frequencies, reflections aren't a big problem like they are at RF. If this is still a question, then somewhere else the article should make obvious the meaning of voltage vs. power measurements. Gah4 (talk) 06:29, 13 October 2016 (UTC)

The reference value is 1 V, not 1 V rms[edit]

In fact there is no such thing as 1 volt RMS. It's just one volt. Period. Dondervogel 2 (talk) 14:35, 26 November 2016 (UTC)

Not true, despite your "period".

Pros and Cons[edit]

The "cons" part is much longer than the "pros" part, and seems contain mostly complaints that people who don't understand decibels don't understand decibels. Logarithms are confusing to people who don't undersand logarithms. Long division is confusing to people that don't understand long division. Is any of this worthy of putting in an encyclopedia article? Sorry, ignorance of math does not score a point against math. — Preceding unsigned comment added by 139.68.134.1 (talk) 21:34, 19 January 2017 (UTC)

I agree. And the sources being cited don't really support the implication that they are "complaining" about how decibels work; they're just explaining. That's not a con. Dicklyon (talk) 23:11, 19 January 2017 (UTC)
The sources cited refer to confusion caused by use of the decibel. Explaining the confusion can be a benefit, but the confusion itself is not. Dondervogel 2 (talk) 00:15, 21 January 2017 (UTC)
I agree that confusion is not a benefit, but I mostly don't agree that the sources are referring to confusion caused by use of the decibel. The 1954 "bewildering" paper is proposing an alternative that didn't catch on; obviously COI there in criticizing what he's trying to replace. Some of the others assert confusion, but these primary sources don't represent any significant viewpoint. Dicklyon (talk) 05:11, 21 January 2017 (UTC)
Those primary sources are all we have. I witness the confusion almost every day of my working life, so I would not agree that those primary sources are somehow unrepresentative. Dondervogel 2 (talk) 08:34, 21 January 2017 (UTC)
"Those primary sources are all we have." And they are weak. If they're all you have, then maybe the material should be deleted. Multiplication is like addition; but repeated. That's so confusing! We should just use addition, because someone who couldn't finish sixth-grade math is confused by stuff that they can't understand. It's all a pretty pathetic excuse for a "criticism" section. Is Wikipedia supposed to be supporting lackwits and losers and their ignorance? — Preceding unsigned comment added by 108.49.176.208 (talk) 05:10, 25 January 2017 (UTC)
Greetings. Following the discussion with interest. Just wondering if Dondervogel 2 would elaborate on the confusion that he witnesses every day. Constant314 (talk) 05:20, 25 January 2017 (UTC)
I'd like to hear about that, too. Seems odd to me, as I've worked with people using decibels for decades, with no problems. And without secondary sources, I don't see how we can say that decibels cause confusion. Dicklyon (talk) 05:36, 25 January 2017 (UTC)
I'm not entirely sure of the relevance of my own experience. What matters are the sources. But seeing as you ask, my work involves interaction with people who use the decibel without an understanding of its meaning. These include regulators, administrators, and similar, and in general are not well versed in mathematics or engineering. For example, they read a sound power level of 120 dB and a sound pressure level of 100 dB and infer from that that the power level represents more noise (and therefore requires more of their attention) than the sound pressure level. Alternatively they might see an rms sound pressure level of 80 dB and a peak sound pressure level of 83 dB and imagine using similar reasoning that the peak sound pressure level is a bigger concern. I have also seen journalists make ridiculous claims that s source level of 190 dB is equivalent to this many nuclear explosions or that many billion Boeing 747s. The confusion is rampant. I can dig out the crappy journalism for you if that helps at all, but even thought the comparisons made are clearly nonsense, their interpretation as such would be OR in my view. In what sense are the multiple letters not secondary sources? Dondervogel 2 (talk) 07:58, 26 January 2017 (UTC)
Thanks for sharing. I see what you mean. Like Dicklyon, I have worked with telephone and CTV technicians for 40 years who use dBs daily. It was hard to see any confusion. Constant314 (talk) 08:13, 26 January 2017 (UTC)
DV, do you think people would be less confused to see sound pressure in pascals? Human hearing works well from about 0.000020 to 2.0 pascals, and rock concerts might go to 20 or more, and that 190 dB explosion/jet engine would be about 100000 pascals. Would that help anyone? Would they think that 2 pascals is twice as loud, or twice as much sound power, as 1 pascal (hint, it's neither). Sound power flux in watts per square meter might be more meaningful, but then the sizes of the numbers would be even more ridiculous. Of course, you could use scientific notation, with powers of ten, but then the power is essentially decibels, so it just makes an even more confusing mixed notation. Dicklyon (talk) 16:32, 26 January 2017 (UTC)
What I find causes most confusion is that the difference between power and intensity is at best hidden in the reference value and too often lost altogether because the reference value is not even stated. If I were to tell a journalist that the mass of my body (say 100 kg, using round numbers here) and that its height is 2 m, he or she would not imagine that my mass was 50 times greater than my height. Why? Because they learn at school that fings wot hav diffrant units are diffrant in naytshah. For the same reason, if I were to state (taking my first example) that the rms sound pressure were 0.1 Pa and the sound power was 1 W, no one would conclude that the power was 10 times the sound pressure. By expressing quantities in decibels we take unsuspecting journalists, sociologists, regulators and the like, including some scientists who should know better, too far from their comfort zone, and the result is confusion. Dondervogel 2 (talk) 16:36, 27 January 2017 (UTC)
OK, that could happen, but it's not usual, and is best avoided by using dB correctly with reference specified, such as dB SPL. And in cases where it's really just dB characterizing a ratio, they really can be added or compared. Like this cable has 10 dB loss and that one has 5 dB loss, which is half as much, so if it's the same cable type it must be half as long. Or if I use two 10 dB amplifiers I can get 20 dB of gain. Very useful and intuitive. Dicklyon (talk) 17:13, 27 January 2017 (UTC)
Yes, for some operations (especially multiplicative one like 10*1.5 W = 15 W) a logarithmic unit can help, for others (especially additive ones like 15 W = 10 W + 5 W) they do not. The bottom line is that neither your experience nor mine are relevant to this discussion. We just follow the sources. Dondervogel 2 (talk) 17:34, 27 January 2017 (UTC)

Criticism[edit]

I recommend to convert the

  • "Advantages and disadvantages / Supporting arguments" subsession into an "Advantages" session
  • "Advantages and disadvantages / Criticism" subsession into an "Frequently asked questions" session
  • and remove the remaining empty "Advantages and disadvantages" session

The reason is, that better if no editor blames him- or herself by supporting the silly "Criticism" subsession, what anyways blames the quality of the Wikipedia itself as well.

To understand the level of same I have to state, that I give frequently references to Wikipedia pages for my university students, but this case I disregaded giving rerence to this page, due to this silly "Criticism" subsession, I simply do not want to teach how some people do not understand the basic arithmetics of logarithmic scales.

prohlep (talk) 09:31, 22 March 2017 (UTC)

There's nothing silly about the criticism. If you are teaching only the advantages of using the decibel you are misleading your students. Dondervogel 2 (talk) 11:10, 24 March 2017 (UTC)
I think the Criticisms section needs work (especially the last paragraph) and I am not opposed to exploring options to rename sections or reorganize. I don't like Prohlep's "Frequently asked questions" suggestion. It may be possible to rework "Criticisms" into a "Misapplications" section. There's nothing wrong with a logarithmic scale, you just have to know when and when not to use it. ~Kvng (talk) 13:54, 9 June 2017 (UTC)

OK, let us put in this way: (1) everything has disadvantages if it is misused, and the "disadvantages / Criticism" is nothing else but report on uneducated misuses. (2) The "Advantages and disadvantages" suggest as if there was a debate on the advantages and disadvantages of using decibel. But there is no debate at all, but only uneducated users of decibel. prohlep (talk) 15:56, 12 June 2017 (UTC)

In what sense are the authors of refs 25-31 uneducated? Dondervogel 2 (talk) 16:12, 12 June 2017 (UTC)

3 dB[edit]

Background: we're discussing the alternative phrasings for describing a factor of 2:

  • More precisely, the change is ±3.0103 dB, but this is invariably rounded to "3 dB" in technical writing.
  • More precisely, the change is ±3.0103 dB, but this is often rounded to "3 dB" in technical writing.

@Dondervogel 2: I don't mean to get into an edit war, but "often" is far too vague of a WP:WEASEL word. This is WP:OR, but I've been doing this for a lot of years, and the only places I've ever seen "3.01 dB" is in tutorials like this which are explaining the approximation, or student exercises where a teacher wants to forestall arguments. (Which is why I added the caveat "in technical writing".)

Go compare:

The exceptions where "3.01 dB" is written in technical writing like [1] and [2] are cases where the value occurs in a table with other values expressed to 2 decimal places.

The vast majority of the time, whether an author writing "3 dB" is actually measuring 3.00 or 3.01 dB is irrelevant. Link budgets are rarely computed to more than 0.1 dB of precision, so 0.01 dB is simply negligible. Real-word component tolerances (and temperature coefficients!) overwhelm that degree of resolution. I may be able to simulate a circuit made with ideal components where I can meaningfully distinguish the 3.00 and 3.01 dB corner frequencies, but the distinction vanishes as soon as I leave the simulation.

The issue arises in theoretical discussions, where there is an exact answer, and in those the half power point is essentially always referred to as "3 dB".

I'm happy to discuss alternative phrasing, but I'd like something considerably stronger than "often". I was trying to keep it brief, and I thought "invariably"'s connotation of "for practical purposes, always" was about right. It's not impossible to not round, but it is, to a first approximation, never done. I haven't yet found a word or phrase that's more precise without being awkward. "Almost always?" 71.41.210.146 (talk) 14:44, 25 March 2017 (UTC)

"Often" has the advantage of being correct. It can be stronger if you prefer but "invariably" is plain wrong. I can accept "almost universally". Dondervogel 2 (talk) 16:21, 25 March 2017 (UTC)
I always round to 3.01 dB. Constant314 (talk) 16:28, 25 March 2017 (UTC)
@Dondervogel 2: "almost universally" it is, thanks! As for "invariably" being incorrect, I understand that its literal meaning is "without exception", but as a literary term it is understood to mean almost that, to the point that exceptions are noteworthy, e.g. "Seattle is invariably cloudy" or "Parisians are invariably rude." Since that's what I was trying to convey, the usage seemed apt to me.
@Constant314: Interesting! What area do you work in that this is common? I use dB primarily in filter design, link budgets, noise analysis (dB rel. 1 nV2/Hz), voice coil transducers (lots of dynamic range!), and a little bit of RF, and I've only seen corner frequencies described as "3 dB". E.g. "The 3 dB point of an RC filter is the frequency at which R = 1/ωC = 1/2πfC". 71.41.210.146 (talk) 20:20, 25 March 2017 (UTC)

"Those primary sources are all we have"[edit]

This may be true for pure facts in the historical sense (see Charles McCabe's maxim about facts). For scientific topics, especially mathematical ones, we also have our brains. For the decibel, sticking exclusively to primary sources (which are poor, as someone else politely remarked) will never yield a decent account. WP:OR is all too often invoked as a pretext for reverting exactly those edits that might be most clarifying for the readers. A major problem is that, in the absence of a proper consensus among editors, the only appropriate main source in the narrow WP:OR sense would be the standards. However, edits to that effect (not made by me!) have regularly been reverted or diluted into nonexistence for various rather strange reasons. A good reason is that the accounts in the standards are far from clarifying to readers and overly restrictive to reflect the freedom required (and taken) in actual practice. But nothing better can be put in their place if WP:OR is systematically taken too narrowly. This is the conundrum for the decibel. The only solution is leaving room for a modicum of clear thinking and interpreting WP:OR wisely. Boute (talk) 18:08, 5 June 2017 (UTC)