Advanced eclipses #395
Advanced eclipses #395
Conversation
| // Computes the refractive index of a gas at a given wavelength (in m), pressure (in Pa), and | ||
| // temperature (in K). | ||
| double getIoR(GasType type, double lambda, double pressure, double temperature) { | ||
| double mu2 = std::pow(lambda * 1e6, -2.0); | ||
| double n0 = 1.0; | ||
|
|
||
| switch (type) { | ||
| // E. R. Peck and B. N. Khanna. Dispersion of nitrogen | ||
| case GasType::eNitrogen: | ||
| n0 = 1.0 + 6.8552e-5 + 3.243157e-2 / (144 - mu2); | ||
| break; | ||
|
|
||
| // J. Zhang, Z. H. Lu, and L. J. Wang. Precision refractive index measurements of air, N2, O2, | ||
| // Ar, and CO2 with a frequency comb. This is given at 20°C. | ||
| case GasType::eOxygen: | ||
| n0 = 1.0 + 1.181494e-4 + 9.708931e-3 / (75.4 - mu2); | ||
| n0 = 1.0 + (n0 - 1) * 293.15 / 273; // Correct for the actual temperature. | ||
| break; | ||
|
|
||
| // E. R. Peck and D. J. Fisher. Dispersion of argon | ||
| case GasType::eArgon: | ||
| n0 = 1.0 + 6.7867e-5 + 3.0182943e-2 / (144 - mu2); | ||
| break; | ||
|
|
||
| // J. G. Old, K. L. Gentili, and E. R. Peck. Dispersion of Carbon Dioxide | ||
| case GasType::eCarbonDioxide: | ||
| n0 = 1.0 + 0.00000154489 / (0.0584738 - mu2) + 0.083091927 / (210.9241 - mu2) + | ||
| 0.0028764190 / (60.122959 - mu2); | ||
| break; | ||
| } | ||
|
|
||
| // The pressure is given in Pa, the temperature in K. The IoR above is measured at STP (273 K, | ||
| // 101325 Pa). We need to correct for the actual pressure and temperature. | ||
| return 1.0 + (n0 - 1.0) * pressure / temperature * 273 / 101325.0; | ||
| } |
There was a problem hiding this comment.
Wouldn't it be more future proof, to read in the parameters for each gas from config files? This would make adding new gases easier.
There was a problem hiding this comment.
Yeah, I guess we could do this. But this code is currently not strictly required anyways, as we do not support wavelength-dependent refractive indices. It's just quite cool for future reference, I think.
| bool RefractionSupported() { | ||
| #if USE_REFRACTION | ||
| return true; | ||
| #else | ||
| return false; | ||
| #endif | ||
| } |
There was a problem hiding this comment.
Why use a function instead of a constant? A function call is more overhead (although the compiler will most likely optimize it away).
There was a problem hiding this comment.
Because the model shader is linked to the main atmosphere shader which needs to know whether the model supports refraction. I think I cannot read a constant from a linked shader, or can I?
|
|
||
| //////////////////////////////////////////////////////////////////////////////////////////////////// | ||
|
|
||
| std::tuple<GLuint, int32_t, int32_t> read2DTexture(std::string const& path) { |
There was a problem hiding this comment.
I think a tuple is not a good return type here. Since it is not clear to the caller, what each element is representing. The documentation in the header file also doesn't explain this. Only the return in the implementation can give a glimpse.
I think a better return type would either be a std::pair<GLuint, ivec2> or a custom type.
|
|
||
| //////////////////////////////////////////////////////////////////////////////////////////////////// | ||
|
|
||
| std::tuple<GLuint, int32_t, int32_t, int32_t> read3DTexture(std::string const& path) { |
| std::tuple<GLuint, int32_t, int32_t> read2DTexture(std::string const& path); | ||
| std::tuple<GLuint, int32_t, int32_t, int32_t> read3DTexture(std::string const& path); |
There was a problem hiding this comment.
See CPP file for comment on return types.
| const float D = 0.20; | ||
| const float E = 0.02; | ||
| const float F = 0.30; | ||
| return ((c * (A * c + C * B) + D * E) / (c * (A * c + B) + D * F)) - E / F; |
There was a problem hiding this comment.
It is hard to read with having an upper and lower case c in the same equation. Maybe use UVWXYZ or write color instead of c.
Co-authored-by: Jonas Gilg <[email protected]>
This PR brings two major contributions: