On Fri, 24 Aug 2012 09:22:17 +0200, Thomas de Groot <tho### [at] degrootorg>  
wrote:

> On 23-8-2012 21:27, Nekar Xenos wrote:
>> I was trying to figure out how big an area light should be to simulate
>> the sun when the light is a lot closer than the sun actually is. Looking
>> at a wall about 3m high I noticed the shadow on the ground blurred about
>> 3cm. I came to the conclusion that the diameter of my light should be
>> about 1% of the distance from the object. To test my theory I looked up
>> the diameter of the sun and the distance from the sun to the earth.
>> Close enough! The diameter of the sun is 1 391 000 km and the distance
>> from the earth to the sun is 149,669,180 km. That's 0.929% :)
>>
>>
>
> I use the following, originally proposed by Cousin Ricky iirc, which  
> seems fairly correct:
>
> #declare SunDis = vlength(SunPosition)*2/215;
>
>    light_source{
>      SunPosition
>      color SunColor
>      area_light
>      SunDis*x, SunDis*z,
>      5, 5
>      adaptive 2
>      jitter
>      circular
>      orient
>      parallel
>      point_at <0, 0, 0>
>    }
>
> Thomas
>
>

Yes, that works out to 0.93% closer than my 1% guesstimate :)

-- 
-Nekar Xenos-

Post a reply to this message