Talk:Spectral flux density

New intro

The new intro [1] is really bad. Encyclopedia articles are about things, not about terms (Wikipedia is not a dictionary). A good intro should always begin by saying what the subject of the article is. If a term really "means different things in different conceptual frameworks", then each of those meanings would require its own article. If the meanings are close enough to be covered in a single article, then the intro needs to focus first on what unifies them, not on their "different meanings in different conceptual frameworks". All the parentheses in the first sentence was also an awful idea. It makes the sentence almost illegible.--Srleffler (talk) 03:46, 9 September 2010 (UTC)

The first sentence of your new wording of the introduction excludes the case in which the flux is defined as a full spherical integral, the "vector conception". Your second dot point thus contradicts your first sentence. The "vector conception" definition has long been used by serious experts. There are not so many measuring instruments that collect radiation from a whole sphere and find its total direction and sense; the whole-sphere quantity is usually obtained as an integral.Chjoaygame (talk) 20:00, 9 September 2010 (UTC)

Thanks, I tweaked the wording. It may need further adjustment; there is a lot going on in this article now.--Srleffler (talk) 03:58, 10 September 2010 (UTC)

What is this article supposed to speak of?

The subject in itself would be extremely simple: the title is almost self-explaining. So, the article should only tell: flux of what (and, waves or particles, light or sound is not conceptually different); density with respect to which variables (whether per unit solid angle, and/or per unit emitting/receiving surface area, per unit wavelength or frequency); in which fields it is used; and with which units. Instead, the article is 13,012 bytes long, has references to 11 boooks, and many sections are written as if they should explain some very deep, mysterious concept. I think the article should be heavily simplified. --GianniG46 (talk) 23:32, 3 December 2010 (UTC)

Dear GianniG46, I see that you think the article should be heavily simplified. It is evident in your comment "waves or particles, light or sound is not conceptually different" that you like simplification.

It is evident from your remarks "whether per unit solid angle, and/or per unit emitting/receiving surface area, per unit wavelength or frequency" that flux is not a quantity that you have considered closely, since your remark does not mention the key element of flux, that is has a definite direction and sense, like a vector.

The description of electromagnetic radiative fields in terms of radiometic quantities is used by a variety of students. They have developed over the years a variety of different conceptions for the purpose. When I found the article it was written solely about a conception that did not have a place in any of the variety of well established textbooks that I was using; the article had the merit of simplicity and the demerit of total uselessness. I thought that another person who looked up the Wikipedia from the same viewpoint that I did might want something useful. I found that the literature has a variety of conceptions and that it was not too easy to see which was which in different texts, and that the texts did not warn the reader that their conceptions were respectively only one of a diversity. When people discuss questions of physics, if they unknowingly use different definitions of the same word, they are likely to have their time wasted in futile argument. I did not feel I was the authority to decide the one right way.Chjoaygame (talk) 05:54, 4 December 2010 (UTC)

You are perfectly right, I like simplicity. And you confirm what I mean, by speaking of "conceptions" while describing just a precise way to combine physical quantities to obtain another obvious physical quantity. And, concerning vectors, also velocity is a vector, and can be treated as a vector (e.g. in cynematics), or as a scalar (in your car). But I would not use dozens of lines elucubrating on the "vector conception" as opposed to the "speedometrical conception", to imitate your (invented) "pyrgeometrical" conception. --GianniG46 (talk) 15:22, 7 December 2010 (UTC)

A great word, 'elucubrate'! I have to take my hat off to that one!Chjoaygame (talk) 18:38, 7 December 2010 (UTC)

Vector conception

From what I can understand, according to the article the "pyrgeometrical conception" is an integral of the flux (or of the flux times cosine of incidence angle?) over a hemisphere, and is the way a pyranometer, a pyrgeometer, a solar cell and any plane surface collect light. Much less clear is for me the "vector conception". The article speaks of a full spherical integral. I suppose it is not the integral of a true net vector flux, which, by Gauss's divergence theorem, would be proportional to the net power generated (or absorbed) in the sphere, so it would be zero in most circumstances. Two pyrgeometers back to back are used, i.e two separate hemispherical measurements of radiation going one way (e.g. light from sky and sun) and the other way (e.g. thermal radiation and reflected light from earth). One gets two positive numbers, and sums them (or subtracts, since it is spoken of an algebraic sum?). Why this is called "vector conception", if it can be obtained by summing two non-vector ("pyrgeometrical") measurements? And in which sense one measurement and two measurements represent different "conceptions"? It is evident I have understood nearly nothing of the whole matter. --GianniG46 (talk) 00:59, 21 December 2010 (UTC)
P.S. I see that in the section "Comparison between vector and pyrgeometrical conceptions" it is told that an isotropic flux has zero vector flux density. So, evidently it is a true net vector flux, and what I have said about Gauss's theorem apply, which means that also for non-isotropic fluxes a) the spherical integral is zero where there are no sources/sinks of radiation; b) it is impossible to measure the spherical integral with ordinary sensors, because the sensor absorbs and perturbates the flux. --GianniG46 (talk) 01:23, 21 December 2010 (UTC)

Dear GianniG46, I see that you are keen to improve things.

For the beginner who wants to learn about this, the warning that there are two radically different main conceptions is I think needed, to save him a vast amount of muddle from different texts. The article when I found it was only about the point source case, a third one, which is highly specialized, and which I have not seen in any text apart from the one cited by the original starter of the article; I did not interfere with his work more than was necessary to indicate that it is not the only way to think.

"the "pyrgeometrical conception" is an integral of the flux" The pyrgeometrical conception (as I have dubbed it for lack of a clear identifier in much of the literature) is the hemispherical integral of the specific radiative intensity, not weighted by the cosine; it measures the total one-way radiative passage of energy through a small 'unit' area at the base in the centre of the hemisphere that the operator has chosen. The usual idea, used for example by Planck, and followed by every textbook that I have read, is that the specific radiative intensity is the fundamental radiometric quantity that characterizes the radiative field; it is built on the idea of rays. Flux is a derived quantity, built on the idea of transfer of energy across an area. Your comments seem to indicate that you regard flux as a primary quantity. The pyrgeometrical conception is used for flat stratified uniformly layered fields. All the energy that hits the collecting surface (from one side) is absorbed and so no cosine correction is needed.

"the "vector conception"" is the definition of flux used by venerable authorities such as Chandrasekhar and Mihalas and Mihalas, followed by authorities such as Goody and Yung who specifically state that they try to follow the conventions set by Chandrasekhar. As noted it is zero in an isotropic field. This is because no net energy transfer is carried by an isotropic field.

"would be proportional to the net power generated (or absorbed) in the sphere" It is not the net power generated (or absorbed) by the sphere. Light is not a material substance like water which is governed by a simply local conservation principle. Light travels fast over long distances and penetrates things, and travels in both senses at once of the same direction. Thus the Gauss theorem does not directly apply in the obvious way. The flux in this conception is not about local sources and sinks; it is about actual total or net radiative transfer of energy.

"Why this is called "vector conception"," This a vector because nature decides its direction and sense. The two scalars used for the 'net flux' way of getting the magnitude of the vector flux are the values of scalar-valued functions of direction dictated not by nature by how the operator chooses to orient the hemispheres of integration. In some cases this procedure leads to the same value as the magnitude of the relevant vector.

"Two pyrgeometers back to back are used" In general this is not how the flux vector is defined. It is defined by what would happen if the specific radiative intensity were measured at the point of interest in every direction and sense. Only in special cases is a pyrgeometer specific enough in frequency to provide the answer.

"it is impossible to measure the spherical integral with ordinary sensors" Agreed. A simple measurement with an ordinary sensor is not possible. Thus the spherical integral is used by theoreticians more than by practical men.

The use of the word density is significant. In ordinary language, a flux is a total flow over any chosen area, not necessarily unit. Here it is customary to mean flux per unit area which is a density per unit area, though not per unit volume; there is no ordinary language word for something per unit area, and 'density' has to stretch to do the job. I think a beginner needs to be reminded of this.Chjoaygame (talk) 12:13, 21 December 2010 (UTC)

A pyrgeometer (like most other sensors) has a flat surface, so you have to multiply the light it receives by the cosine of incidence angle, when integrating. And usually, one wants just this integral, not, as it appears from the section "pyrgeometric conception", the sum of all radiation entering into a hemisphere. This gives a factor pi, as correctly said in the section "comparison", instead of 2 pi, which would be obtained from a hemisphere without the cosine factor.

A similar confusion arises from the "spherical integral", which makes me imagine either integrating the "one-way" flux entering a sphere (like that obtained by a sphere entirely covered by sensors), or the usual integral of flux on a closed surface, which would give zero if there are no sorces/sinks inside, by the divergence theorem. Which is fully valid, of course, for a well-behaved vector field with conservative, inverse-square properties such as radiation (see the book by Mihalas & Mihalas, which uses it a lot of times). Instead, this "vector conception", as I understand, represents just the vector which has the average direction and intensity of the flux, i.e. the (normalized) vector sum of all fluxes.

These concepts are much simpler and intuitive than it appears from the article, which would be more understandable without speaking of "conceptions", but simply saying that when one has a vector field, one may be interested in knowing the average flux density impinging on a particular surface (e.g. on solar cell, or on a pyranometer); or in how much net radiation is exchanged and in which direction, i.e. which is the vector "total net flux density" (e.g. in evaluating heat exchanges); or else, one has to retain the whole vector field, e.g. when performing theoretical analyses of radiative transfer. Anyway, I think it would much more clear to speak of integrations over solid angles, rather than over spheres or hemispheres. --GianniG46 (talk) 12:38, 23 December 2010 (UTC)

Dear GianniG46, Thank you for this reply.

"the article, which would be more understandable without speaking of "conceptions", but simply saying that when one has a vector field,when one has a vector field, one may be interested in knowing the average flux density impinging on a particular surface (e.g. on solar cell, or on a pyranometer); or in how much net radiation is exchanged and in which direction, i.e. which is the vector "total net flux density" (e.g. in evaluating heat exchanges); or else, one has to retain the whole vector field, e.g. when performing theoretical analyses of radiative transfer." The word 'conception' was used to make it easier (and dare I say it, simpler?) to refer to the fact that there are many 'or's in the picture, and to bring it to the reader's attention that there are indeed diverse ways of thinking here. Perhaps you would be happier with 'way of thinking' or 'way of seeing things' instead?

"A pyrgeometer (like most other sensors) has a flat surface, so you have to multiply the light it receives by the cosine of incidence angle, when integrating. And usually, one wants just this integral, not, as it appears from the section "pyrgeometric conception", the sum of all radiation entering into a hemisphere." Yes, you are right, I was mistaken to ignore the Lambert cosine factor. Thank you for this. I have made an edit to correct my mistake.

"when one has a vector field". The radiative field is not a vector field. For radiometry, the radiative field is usually considered to be described by the specific radiative intensity.

"the book by Mihalas & Mihalas, which uses it a lot of times). Instead, this "vector conception" ". The present "vector conception" is just that of the flux of Mihalas and Mihalas 1984.

"A similar confusion arises from the "spherical integral", which makes me imagine either integrating the "one-way" flux entering a sphere (like that obtained by a sphere entirely covered by sensors)". We are here defining flux from scratch. We cannot reasonably use the word flux as part of our definition. It seems you are not interested in the idea of specific radiative intensity as the fundamental radiometric descriptor of the radiative field.

"Anyway, I think it would much more clear to speak of integrations over solid angles, rather than over spheres or hemispheres." Thank you. I have made an edit to make it clear that the integral is over all the solid angle of an imaginary sphere.Chjoaygame (talk) 15:47, 23 December 2010 (UTC)

A further edit to put in mathematical symbols.Chjoaygame (talk) 21:02, 23 December 2010 (UTC)

No, I would'nt be happier if you change a couple of words. The article uses thousands of words just to explain that the "vector conception" is the vector sum of all flux densities at a point, and that the "pyrgeometrical conception" is flux per unit area impinging on a prescribed surface. I think these concepts would be more clearly explained by these twenty or thirty words, to which one can add the way to measure and calculate them, speaking of integrals but preferably avoiding speaking of ambiguous spheres and hemispheres; and in which fields they are used, e.g. solar energy, heat exchange.

In the previous talk I complained about the fact that the article was 13,012 bytes long and used 11 book references. Now the bytes are 15,714 and the books 12. If these are the results, the goal of my talk was completely missed. --GianniG46 (talk) 09:05, 26 December 2010 (UTC)

Dear GianniG46, I am sorry you wouldn't be happier if I changed a couple of words. The definitions of flux density are in terms of specific radiative intensity. You write: "the vector sum of all flux densities at a point". It does not seem logical to define flux in terms of flux. Some writers such as Chandrasekhar 1950 and Mihalas and Mihalas 1984 and Goody and Yung 1989 define it as a vector, and others define it as a scalar. There are other variant definitions as well, including the one that was the only one offered when I found this article. I have made some more edits.Chjoaygame (talk) 18:50, 26 December 2010 (UTC)

Refer to Jansky

Currently the definition of spectral flux density given here seems to be different from the definition found at Jansky. Maybe it should be mentioned that spectral flux density can also be measured per frequency instead of wavelength (e.g. in Jansky), like mentioned in Irradiance. Fkbreitl (talk) 11:26, 8 December 2015 (UTC)

I think the definition at Jansky corresponds to the one in this article that is summarized as follows:

• Characterizing remote telescopically unresolved sources such as stars, observed from a specified observation point such as an observatory on earth.Chjoaygame (talk) 14:05, 8 December 2015 (UTC)

Ok, I have added a sentence with units per frequency and Jansky. Then this discussion is solved and can be removed. Fkbreitl (talk) 18:48, 8 December 2015 (UTC)

Thank you. Usually such discussions are not removed until some time after they seem complete, I suppose in order to give others a chance to review them. Perhaps you will now very kindly say here exactly what you saw as the discrepancy between this article and the Jansky article? At present I am not sure of that. Perhaps I am mistaken in my reading.Chjoaygame (talk) 00:47, 9 December 2015 (UTC)

I just noticed that this article did not mention that spectral flux density can also be measured in W/m^2/Hz (the unit of Jansky) and as mentioned in the related articles of irradiance etc. So this should be fixed now. Fkbreitl (talk) 09:07, 9 December 2015 (UTC)

Thank you for this check.Chjoaygame (talk) 09:58, 9 December 2015 (UTC)