Fix sun positioning in sky shader example #21575
Conversation
|
Nice catch! I guess it applies to the ocean demo too? |
sirxemic
force-pushed
the
sky-example-fix
branch
from
2876492 to
1cb3e79
Compare
|
Indeed. Fixed it for that example as well just now. |
| @@ -75,7 +75,7 @@ | |||
| gui.add( effectController, "rayleigh", 0.0, 4, 0.001 ).onChange( guiChanged ); | |||
| gui.add( effectController, "mieCoefficient", 0.0, 0.1, 0.001 ).onChange( guiChanged ); | |||
| gui.add( effectController, "mieDirectionalG", 0.0, 1, 0.001 ).onChange( guiChanged ); | |||
| gui.add( effectController, "inclination", 0, 1, 0.0001 ).onChange( guiChanged ); | |||
| gui.add( effectController, "inclination", -1, 1, 0.0001 ).onChange( guiChanged ); | |||
There was a problem hiding this comment.
Does a negative inclination make sense? I think the range should be like in webgl_shaders_ocean.
There was a problem hiding this comment.
I think the proper term is solar elevation. (The term solar altitude is also used, but that can be confusing for obvious reasons.)
Solar elevation is in degrees, and can be negative (at twilight, for example). A proper model should handle that case.
Solar azimuth is also in degrees.
/ping @zz85
There was a problem hiding this comment.
Does a negative inclination make sense?
I'm asking this because changing to a negative value has no visual effect in the demo.
There was a problem hiding this comment.
Any value smaller than -0.0257 has no visual effect, to be more precise. Values lower than that may make no difference visually, but it does allow the user to see the effect of when the sun is at the opposite side of the world.
There was a problem hiding this comment.
There was a problem hiding this comment.
The fact that scene goes dark when the sun hits the horizon is either an error in the model, or an error in the implementation of it. That can be saved for another time if someone is interested in pursuing it.
There was a problem hiding this comment.
@WestLangley What range of angles would you suggest for azimuth and elevation?
Looking at the Wikipedia article it seems the angles are always positive. Assuming we would have [-90, 90] for elevation. What part of the hemisphere would be expressed in negative angles?
There was a problem hiding this comment.
From https://www.esrl.noaa.gov/gmd/grad/solcalc/azel.html
Azimuth is measured in degrees clockwise from north. Elevation is measured in degrees up from the horizon.
I think it is OK for azimuth to be in [ -180, 180 ] for the demo, or [ 0, 360 ] -- whatever you prefer. We do not have a concept of 'north'.
Since the model goes black below the horizon, [ 0, 90 ] for elevation in the demo seems appropriate to me. (Negative values, if supported by the model, would correspond to pre-dawn or post-sunset, when the sun is just below the horizon.
JMHO. :-)
There was a problem hiding this comment.
thanks for spotting this, and for interesting discussion. agreed with the points @WestLangley brought up - azimuth and elevation make sense, and so is having some notion of sun below the horizon for twilight. the model used in the shader have some limited support for negative elevation, which have already been pointed up here :)
There was a problem hiding this comment.
the model used in the shader have some limited support for negative elevation
Hi @zz85! Very limited -- and not to this extent, I think.
|
Thanks! |
I noticed while experimenting with the example that changing the azimuth would influence the inclination as well. Turns out the sun position was calculated incorrectly from these two parameters.
I also changed the UI values for the inclination such that 0 means no inclination, which I thought makes more sense.