Reqiest║Metal temperature

[neo0.]

So, I was in the john and I started thinking of some other ways that indigo could be more physically accurate.. I started thinking of metals.. Different metal glow with different colors when heated to different temps, if I remember correctly (also where fireworks come from I think.) Iron usually glows red, magnesium glows green (blue?).. This would also be great for lighting fixtures.. Im real life, many indoor lighting sources actually come from metal being heated... So, I think it would be great if indigo could physically reproduce this phenomenon.

[Meelis]

But it just works on hi temperatures, don't u set it with emitter the way u like.
Magnesium burnes just white or am i wrong.

[neo0.]

Well.. to be honest, Im kind of confused when it comes to this.. is light temp (in kelvins) equate to the same colors as metal temperature? Obviously, it would vary with different metals, but maybe it is similar for a few?

[lycium]

it is in theory possible for indigo to render images using arbitrary spectra for emission such as given by this page: http://members.misty.com/don/spectra.html

however, depending on how sharp those spectral peaks are, there could well be a lot of noise in the resulting images. mlt might help a bit, but what you really want is a specially-crafted importance sampler for the spectral rendering stuff. i doubt this would be implemented, since it is pretty special-purpose and would slow down the common case.

[gagar]

Well.. to be honest, Im kind of confused when it comes to this.. is light temp (in kelvins) equate to the same colors as metal temperature? Obviously, it would vary with different metals, but maybe it is similar for a few?

There is no such thing as "metal temperature" in the way you are using the term.

When you heat a metal, or another material, it emits a black body spectrum ("light temperature"), which depends on the temperature.

Now, if you excite a metal atom, it will emit a photon at a characteristic wavelength upon relaxation. C.f. http://www.meta-synthesis.com/webbook/1 ... photo.html
That's the effect you see in sodium-vapour lamps (street lamps) and in fireworks.

So, as Lyc mentioned, you just need to find the emission spectrum for the metal you want, and use it.

[CTZn]

If I'm correct, one can use a tabulated spectrum to translate such emission data into Indigo (referring to link from lyc).

[lycium]

yup, and the problem i was referring to is that it would look something like:

0, 0, 0, 0, 0, 0, 0, ... 0, 10, 10000, 10, 0, 0, ... 0, 1, 4000, 1, 0, 0, 0, ...

this is troublesome (you'll get a lot of colour noise) unless specially accounted for. then again, indigo might be even smarter than i imagine (i have a good idea, but nick is full of surprises) and account for this kind of thing already

[fused]

yup, and the problem i was referring to is that it would look something like:

0, 0, 0, 0, 0, 0, 0, ... 0, 10, 10000, 10, 0, 0, ... 0, 1, 4000, 1, 0, 0, 0, ...

this is troublesome (you'll get a lot of colour noise) unless specially accounted for. then again, indigo might be even smarter than i imagine (i have a good idea, but nick is full of surprises) and account for this kind of thing already

i think im going to try that later on

[dougal2]

This might be a useful resource: http://ioannis.virtualcomposer2000.com/ ... amici.html

[lycium]

ahhh, there it is! i was trying to find that link earlier, i'd sent it to nick some time ago.

[dougal2]

There is an implementation of that data set on this other website

[fused]

Code: Select all

    <material>
        <name>0_mid0</name>
        <phong>
            <ior>1.010000</ior>
            <diffuse_albedo>
                <constant>
                    <rgb>
                        <rgb>0.800000 0.800000 0.800000</rgb>
                        <gamma>2.200000</gamma>
                    </rgb>
                </constant>
            </diffuse_albedo>
            <exponent>
                <constant>1000.000000</constant>
            </exponent>
            <base_emission>
                <constant>
                    <regular_tabulated>
                        <start_wavelength>0.34E-06</start_wavelength>
                        <end_wavelength>0.75E-06</end_wavelength>
                        <num_values>64</num_values>
                        <values>0.0359 0.0784 0.0134 0.0258 1000000.0359 0.0784 0.0134 0.0258 0.0359 0.0784 0.0134 0.0258 0.0359 0.0784 0.0134 0.0784 0.0134 0.0258 0.0359 0.0258 0.0359 0.0784 0.0134 0.0258 0.0359 0.0784 0.0134 0.0258 0.0359 0.0784 0.0134 0.0258 0.0359 0.0784 0.0134 0.0258 0.0359 0.0784 0.0134 0.0258 0.0359 0.0784 0.0134 0.0258 0.0359 0.0784 0.0134 0.0784 0.0134 0.0258 0.0359 0.0258 0.0359 0.0784 0.0134 0.0258 5000.0359 0.0784 0.0134 0.0258 0.0359 0.0784 0.0134 0.0258 </values>
                    </regular_tabulated>
                </constant>
            </base_emission>
            <layer>0</layer>
        </phong>
    </material>

4 minutes

[lycium]

pretty much what i expected and it's easy enough to account for (just say "spectral MIS" to nick), but not worth it since such extreme cases are kinda rare. well, there's flourescent lighting that's quite common and very peaky (hence undesirable)...

[CTZn]

yup, and the problem i was referring to is that it would look something like:

0, 0, 0, 0, 0, 0, 0, ... 0, 10, 10000, 10, 0, 0, ... 0, 1, 4000, 1, 0, 0, 0, ...

this is troublesome (you'll get a lot of colour noise) unless specially accounted for. then again, indigo might be even smarter than i imagine (i have a good idea, but nick is full of surprises) and account for this kind of thing already

i think im going to try that later on

Any chance to use ISL to translate one 1d scalar texture into tabulated data ? I don't think that one can handle the relevant data types atm...

Could be a secondary tabulated data type (image) ? At least that could ease use and ressources spotting.

Speculating

[fused]

you can have shaders which are evaluated per wavelenght, iirc.

edit: and yeah, u could also do texture lookups in them