Indigo v0.7 test 2
Posted: Fri Jan 19, 2007 4:58 am
Not too many changes this time, mostly fixing + improving the SSS.
There are some important improvements to the Phong model tho.
Also there's a new antialiasing option. (super_sample_factor)
http://www.indigorenderer.com/montecarl ... _test2.zip
Please point out any bugs!
Changelog:
v0.7 test 2
* reverse gamma correcting diffuse and phong diffuse RGB colours
* improved phong sampling
* increased max albedo for diffuse and phong substrate to 0.9
* diffuse colour for diffuse and phong materials now not used when an albedo texture is used.
* added henyey_greenstein phase function
* fixed crash with SSS and no area lights.
* improved SSS russian roulette
* added super_sample_factor option to inifile.
* added display_period option to inifile.
EDIT(something i forgot): fixed energy loss issue in SSS media, causing darkening in highly scattering media.
About super_sample_factor:
This can be set in the inifile.txt.
It basically renders the image internally at a higher res, then downsamples it just before display and disk save time.
Should help aliasing around bright edges a bit; the downside is that it uses lots of RAM (image buffer ram usage is proportional to super_sample_factor squared).
Set it to 2 if you can spare the RAM, 3 or 4 is probably overkill, 1 is no change from before.
more info to follow soon.
EDIT:
Here's an example of a medium using the new Henyey-Greenstein phase function:
Note that the g parameter may vary with wavelength (e.g. g_spectrum could be a RGB spectrum), but here it is a constant with value 0.9, which corresponds to strong forwards scattering.
There are some important improvements to the Phong model tho.
Also there's a new antialiasing option. (super_sample_factor)
http://www.indigorenderer.com/montecarl ... _test2.zip
Please point out any bugs!
Changelog:
v0.7 test 2
* reverse gamma correcting diffuse and phong diffuse RGB colours
* improved phong sampling
* increased max albedo for diffuse and phong substrate to 0.9
* diffuse colour for diffuse and phong materials now not used when an albedo texture is used.
* added henyey_greenstein phase function
* fixed crash with SSS and no area lights.
* improved SSS russian roulette
* added super_sample_factor option to inifile.
* added display_period option to inifile.
EDIT(something i forgot): fixed energy loss issue in SSS media, causing darkening in highly scattering media.
About super_sample_factor:
This can be set in the inifile.txt.
It basically renders the image internally at a higher res, then downsamples it just before display and disk save time.
Should help aliasing around bright edges a bit; the downside is that it uses lots of RAM (image buffer ram usage is proportional to super_sample_factor squared).
Set it to 2 if you can spare the RAM, 3 or 4 is probably overkill, 1 is no change from before.
more info to follow soon.
EDIT:
Here's an example of a medium using the new Henyey-Greenstein phase function:
Note that the g parameter may vary with wavelength (e.g. g_spectrum could be a RGB spectrum), but here it is a constant with value 0.9, which corresponds to strong forwards scattering.
Code: Select all
<medium>
<name>medium1</name>
<ior>1.0</ior>
<cauchy_b_coeff>0.0</cauchy_b_coeff>
<precedence>20</precedence>
<absorption_coefficient_spectrum>
<rgb>
<rgb>0 0 0</rgb>
</rgb>
</absorption_coefficient_spectrum>
<subsurface_scattering>
<scattering_coefficient_spectrum>
<uniform>
<value>500</value>
</uniform>
</scattering_coefficient_spectrum>
<phase_function>
<henyey_greenstein>
<g_spectrum>
<uniform>
<value>0.9</value>
</uniform>
</g_spectrum>
</henyey_greenstein>
</phase_function>
</subsurface_scattering>
</medium>
