Talk:Weighting filter

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Professional sound production (Rated C-class, Mid-importance)
WikiProject icon This article is within the scope of WikiProject Professional sound production, a collaborative effort to improve the coverage of sound recording and reproduction on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.

First section[edit]

I propose this entire article be moved to Frequency weighting or something, which will address all the different types. - Omegatron 01:39, Mar 5, 2005 (UTC)

Moved. - Omegatron July 1, 2005 22:13 (UTC)

Moved to talk until someone points out the relevance:

Note: The use of F1A-line or HA1-receiver weighting must be indicated in parentheses as required. A sound level of 0.63 pascal 1000 Hz tone will read +90 dBA, but the same sound pressure as white noise, randomly distributed over a 3 kHz band (nominally 300 to 3300 Hz), will read +82 dBA, due to the frequency weighting.

Probably just cruft from Federal Standard 1037C. - Omegatron July 1, 2005 22:13 (UTC)

Note that this is Omegatron's first ever use of the word cruft. - Omegatron July 1, 2005 22:13 (UTC)

Which curve at which loudness?[edit]

From the article: "The curves approximate the human ear at different loudnesses". So each curve of those 4 curves stands for the ear response at different loudness, right? Shouldn't we add the respective absolute loudnesses in dBs for each of the curves, where each of them are applicable? Or have I understood something wrongly here? Peter S. 23:31, 26 September 2005 (UTC)

Yep, that would be good info. In practice everyone just uses A-weighting for most things. I don't know what the values are. — Omegatron 01:58, 27 September 2005 (UTC)
The value of 40 Phon (40dB SPL at 1kHz) is stated in the article. This is correct (according to ISO), though many sources wrongly give other figures like 30 or 60. I don't think it is meaningful to quote a level for the other curves - they are just intended for higher levels but not necessarily derived exactly from any Fletcher Munson curve. Even if they were, the recent revisions in the ISO standard (see Equal-loudness contours because of supposed experimental errors, render precise association invalid. --Lindosland 14:11, 24 June 2006 (UTC)

generalized weightings[edit]

i think this article should be about audio filters only... — Omegatron 02:52, 24 October 2005 (UTC)


I have made some major revisions to this page. I do not agree that this page should be about audio filters only, because that would not be true to the general heading of Weighting filter. I do think a separate page should be made for A-weighting (currently linked to this) and the information on this page split so that all the detailed stuff on A-weighting is separated out. I have just put up a page on CCIR (ITU) 468 noise weighting to cover my comments regarding this more thoroughly. -- Lindosland 24 Oct

It's fine to move the bulk of this content to audio weighting filter and the rest to just weighting, but I think the audio field has plenty of content for its own article. I disagree with splitting out A-weighting from the others. — Omegatron 13:33, 24 October 2005 (UTC)

Correction appreciated. I think we can probably agree on this, though I need to look more carefully at what is there already. I think Noise weighting filter should perhaps cover A,B,C,D, ITU-R468, Psophometric (telephony)reasonably fully, but then more detailed individual pages are probably warranted for each of these (I can add detailed graphs,specs for time constants etc in due course and would also like to see comparison of old and new A-weightings in the detailed A-weighting page). Weighting filter/curve should stand separately to cover all types of weighting. Lindosland 00:02, 25 October 2005 (UTC)

Good ideas. — Omegatron 00:09, 29 October 2005 (UTC)

Gamma correction[edit]

Is gamma correction used in TVs and VDUs (monitors) considered a weigthing. If so it should be included.--Light current 18:49, 29 October 2005 (UTC)

Spell checker split up micropascals--Light current 02:50, 30 October 2005 (UTC)

I considered this, and decided that gama correction is an example of 'weighting' but not a 'weighting filter' which tends to be a physical realisation, either electronic or optical. Gamma correction uses s set of three numbers, applied to the three colour channels - there is no spectrum involved. Lindosland 16:02, 1 November 2005 (UTC)

Fair comment!--Light current 16:13, 1 November 2005 (UTC)

Actually I didn't quite get that right, and I realised later that my mind was on the three weightings applied to RGB signals to derive chrominance and luminance components. My conclusion holds though. Gamma correction is the application of a non-linear function to a single variable (brightness). While you might think of a function used to compress or expand in this way as a form of 'weighting' (over the range of values) it's really too basic to call weighting, and weighting normally refers to signals with a spectrum or a number of separate parts. Lindosland 22:38, 1 November 2005 (UTC)

Objection to merging[edit]

I do not agree that this page should be merged with A-weighting for the simple reason that A-weighting is not a weighting filter, it is just one of many weightings (a weighting is a mathematical concept while a filter is usually a physical realisation) that might be used in the creation of a weighting filter. I went to some trouble to emphasise this, and separate the concepts, in the past. A-weighting deserves its own page as it is an important subject. It is also commonly misunderstood and so keeping it on its own page avoids encouraging further misunderstanding. --Lindosland 23:38, 13 May 2007 (UTC)

It seems that little of the information in this article is addressing the filter aspect, but mainly targets (merely) weighting. Most of the relevant information is only a repetition of what is already stated correctly in A-Weighting and other articles. And unfortunately the article is flawed by incorrect information (e.g. dBrn adjusted is NOT a synonym to dB(A) - a typical mistake done by someone, not knowing the field, spotting in a list of designations that dBa - which is NOT the same as dB(A) - also is denoted dBrn), doubtful personal interpretations of facts (e.g. 'and so noise measurement on appliances should state "at 1 m in an open field" or "at 1 m in anechoic chamber' - in open fields and anechoic environments such a measurement value could be totally useless or misleading without an accompanying radiation pattern), omission of relevant facts (e.g. a long informal section on noise measurements of electrical appliances without even mentioning that there are relevant standards, as IEC 704-1), personal opinions (e.g. "when, for example, hum is present at 50 or 100 Hz at a level above the quoted (weighted) noise floor this is of no importance because our ears are very insensitive to low frequencies at low levels, so it will not be heard", which cannot be generally stated) or obscure information without references (e.g. "the best 468-weighted results are in the region of −68dB relative to Alignment Level (commonly defined as 18 dB below FS)"). Therefore, I suggest a total revision of the article or a removal and merging of applicable parts into other articles on weighting, filters and so on (it seems most correct and relevant information is already available in related articles, though). — Preceding unsigned comment added by (talk) 15:44, 21 July 2012 (UTC)

Please remove dBrn[edit]

This article has several flaws, which are pointed out above, but please at least remove the strange statement that dBrn adjusted is a synonym to dB(A). It is simply wrong and such false statements undermines the credibility of Wikipedia, particularly since it has been pointed out before here but still stands uncorrected. dBrn adjusted is a synonym to dBa which is not the same as dB(A) and "a" has nothing to do with A-weighting here ("a" denotes "adjusted"...). Check out (as an example I paste the dBa definition):

dBa Abbreviation for decibels adjusted. Weighted absolute noise power, calculated in dB referenced to 3.16 picowatts (-85 dBm), which is 0 dBa. (188) Note: The use of F1A-line or HA1-receiver weighting must be indicated in parentheses as required. A one-milliwatt, 1000-Hz tone will read +85 dBa, but the same power as white noise, randomly distributed over a 3-kHz band (nominally 300 to 3300 Hz), will read +82 dBa, due to the frequency weighting. Synonym dBrn adjusted.

The good thing is that someone with any knowledge in this area can tell that this was written by someone who did not have a clue himself, which makes the reader more cautious (which obviously and unfortunately is needed here...) — Preceding unsigned comment added by (talk) 08:41, 20 September 2013 (UTC)