On 14/11/2017 13:57, clipka wrote:
> Am 14.11.2017 um 10:57 schrieb Stephen:
>> On 14/11/2017 08:03, Thomas de Groot wrote:
>>> On 13-11-2017 21:27, Stephen wrote:
>>>> The tunic and wings are emissive df3s, made from Poser models.
>>>>
>>>> I think I need a dragon and a rocky scene to complete this. :)
>>>>
>>>>
>>>
>>> Yes indeed although I like this ballet school setting with the mirrors.
>>
>> It is old enough to have been made in Moray. :-)
>>
>>> A couple of things, if I may: (1) the overhead lights should be dimmed
>>> - probably by at least 50 % (if not 25%) also for compensating the
>>> light reflections from the mirrors.
>>
>> The lighting is problematic. The light in the hall of mirrors brightens
>> when I switch on radiosity. I did not mean this to turn into an
>> experiment in radiosity but it looks looks like it is going to be. I
>> have a normal area light in there to even things out.
> 
> Problems with different "quality levels" of illumination are to be
> expected here, due to the huge mirrors as well as the glowing wings.
> 
> 
> In reality, objects in the room would be lit by...:
> 
> (1) Light coming directly from the light sources.
> 
> (2a) Light bouncing off diffuse objects.
> (2b) Light emitted by glowing stuff.
> 
> (3a) Light bouncing off diffuse objects and then the mirrors.
> (3b) Light emitted by glowing stuff and bouncing off the mirrors.
> 
> (4) Light coming from the light sources and bouncing off the mirrors.
> 
> (5) Light bouncing off the mirrors and then diffuse objects.
> 
> 
> To simulate (1), "classic" raytracing will do.
> 
> To simulate (2) and (3), you need to enable radiosity.
> 
> To simulate (4), you need to enable photons.
> 
> To simulate (5), you need to enable both radiosity and photons.
> 
> If any of these are not simulated, the corresponding portion of
> illumination will simply be missing, causing the image to appear darker.
> 
> The lack of illumination from (2), (3) and (5) can be compensated for to
> some degree, by making use of the "ambient" mechanism.
> 
> 
> As far as overall brightness goes, you should ideally use renders with
> radiosity and photons as a yardstick, and set up "ambient" accordingly
> for fast test renders.
> 
> (As a faster alternative to photons for this particular scene, you might
> try using mirrored copies of your primary light sources,
> `projected_through` the corresponding mirrors. Make sure to properly
> account for any dimming effect if your mirrors aren't 100% reflective.)
> 

Thanks, they aren't, they are about 95% reflective. But I've ditched the 
ordinary light source and am using emissive cubes as light sources.
I was thinking of adding a scattering media container first. Before 
trying photons.


-- 

Regards
     Stephen

Post a reply to this message

Attachments:
Download 'studio_vic2d_y_901a1_0000.png' (764 KB)

Preview of image 'studio_vic2d_y_901a1_0000.png'