I'm using the following media for the atmosphere inside the spinner.


   interior
   {
     media
     {
       scattering
       {
         4, <0.2,0.4,1.0>  // crappy approximaion of TerraPOV value
         extinction  1
       }
       density
       {
         cylindrical
         poly_wave  0.25
         density_map
         {
           // should be based on city_units!!!
           [0  rgb 100/5000000]
           [1  rgb 010/5000000]
         }
         scale  city_radius
         rotate  x * 90
       }
     }
   }


The "city_radius" is something less than 1000 units. What kinds of stuff 
can I do to improve rendering times with this media enabled? Right now 
it's the single greatest performance bottleneck in my scene.



-- 
http://isometricland.com

Post a reply to this message

Am 25.06.2010 17:11, schrieb SharkD:
> I'm using the following media for the atmosphere inside the spinner.

At this size, wouldn't it be more realistic to have a roughly constant 
atmospheric density?

Post a reply to this message

On 6/25/2010 11:56 AM, clipka wrote:
> Am 25.06.2010 17:11, schrieb SharkD:
>> I'm using the following media for the atmosphere inside the spinner.
>
> At this size, wouldn't it be more realistic to have a roughly constant
> atmospheric density?

I'm not sure. I forgot to say in my last post that 1 unit = 1 meter. 
Would the fact that "gravity" is zero at the center make a difference?

For that matter, how would a gas behave in a cylinder that is spinning? 
I would think that it would tend to remain completely at rest except 
where it comes into contact with the surface objects. Maybe over time 
sustained contact would cause it to build momentum and turn as a 
whole... I dunno.


-- 
http://isometricland.com

Post a reply to this message

Am 25.06.2010 22:24, schrieb SharkD:

>> At this size, wouldn't it be more realistic to have a roughly constant
>> atmospheric density?
>
> I'm not sure. I forgot to say in my last post that 1 unit = 1 meter.
> Would the fact that "gravity" is zero at the center make a difference?
>
> For that matter, how would a gas behave in a cylinder that is spinning?
> I would think that it would tend to remain completely at rest except
> where it comes into contact with the surface objects. Maybe over time
> sustained contact would cause it to build momentum and turn as a
> whole... I dunno.

I guess the air dynamics will probably be... quite dynamic ;-).

Presuming that the spinning motion will in general create a slight 
pressure gradient, you'll get heat convection with air flows between the 
center and the rim. Those currents will be subject to coriolis forces 
and the like, giving the air at higher "altitudes" even a /higher/ 
angular speed than near the ground. So with all this "up" and "down", 
and "horizontal" speed differences, you're in for a lot of eddies I 
guess. Which in turn will contribute to somewhat stabilizing angular 
speed throughout the whole atmosphere.

Post a reply to this message

On 6/25/2010 9:41 PM, clipka wrote:
> Which in turn will contribute to somewhat stabilizing angular
> speed throughout the whole atmosphere.

And pressure?

OT, I was planning on maybe adding a hang glider flying near the lamp. 
They'd be in for one hell of a ride!

-- 
http://isometricland.com

Post a reply to this message

Le 2010-06-25 23:30, SharkD a écrit :
> On 6/25/2010 9:41 PM, clipka wrote:
>> Which in turn will contribute to somewhat stabilizing angular
>> speed throughout the whole atmosphere.
>
> And pressure?
>
> OT, I was planning on maybe adding a hang glider flying near the lamp.
> They'd be in for one hell of a ride!
>

Not for the presure.
As the air in the center may, under some conditions, rotate slightly 
faster than the station, the presure *will* diminishes from the surface 
toward the center.
BUT, as the radius is /only/ 1 Km and quite probably a rim gravity less 
than Earth's gravity, the presure gradient will be relatively small, 
possibly small enough to be neglected. Maybe in the order of 2% or 3%.
After all, you realy don't need a full Earth gravity, something between 
0.5G and 0.7G should be enough.


Alain

Post a reply to this message

On 6/26/2010 1:59 PM, Alain wrote:
> Not for the presure.
> As the air in the center may, under some conditions, rotate slightly
> faster than the station, the presure *will* diminishes from the surface
> toward the center.
> BUT, as the radius is /only/ 1 Km and quite probably a rim gravity less
> than Earth's gravity, the presure gradient will be relatively small,
> possibly small enough to be neglected. Maybe in the order of 2% or 3%.
> After all, you realy don't need a full Earth gravity, something between
> 0.5G and 0.7G should be enough.
>
>
> Alain

If I remove the density map, will rendering times change? Also, how 
should I specify the media's density when the pattern is absent?


-- 
http://isometricland.com

Post a reply to this message

Le 2010-06-27 11:58, SharkD a écrit :
> On 6/26/2010 1:59 PM, Alain wrote:
>> Not for the presure.
>> As the air in the center may, under some conditions, rotate slightly
>> faster than the station, the presure *will* diminishes from the surface
>> toward the center.
>> BUT, as the radius is /only/ 1 Km and quite probably a rim gravity less
>> than Earth's gravity, the presure gradient will be relatively small,
>> possibly small enough to be neglected. Maybe in the order of 2% or 3%.
>> After all, you realy don't need a full Earth gravity, something between
>> 0.5G and 0.7G should be enough.
>>
>>
>> Alain
>
> If I remove the density map, will rendering times change? Also, how
> should I specify the media's density when the pattern is absent?
>
>

You may not need as many samples, whitch will make things go faster.
NEVER use intervals and only use samples. It's MUCH faster.
Set your media base density. The density from the map get multiplied 
with it.
scattering{1, rgb 0.01,.....}
Change the rgb value.



Alain

Post a reply to this message

On 6/27/2010 1:23 PM, Alain wrote:
> You may not need as many samples, whitch will make things go faster.
> NEVER use intervals and only use samples. It's MUCH faster.
> Set your media base density. The density from the map get multiplied
> with it.
> scattering{1, rgb 0.01,.....}
> Change the rgb value.
>
>
>
> Alain

I'll try that then.

OT - Too bad we don't have a GasSys include like we have a LightSys one! :)


-- 
http://isometricland.com

Post a reply to this message

Am 27.06.2010 17:58, schrieb SharkD:

> If I remove the density map, will rendering times change? Also, how
> should I specify the media's density when the pattern is absent?

It's worth a try at least. As for the density, just specify none. All 
you use to control the constant density is the parameter after the 
"emission", "absorption" or "scattering" keyword.

Post a reply to this message