How to resolve detail problems.

[WWX]

Hi All,

I am busy with this interior, using Photon mapping. I have attached images illustrating what it looks like after calculating the photon map and then switching the first bounce to irradiance.

My irradiance setting are set to High.

What I would like to know is why the detail with in the red circled areas are so bad. Surely I should be getting clean lines at these high settings. It is suppose to represent individual timber members and are modelled as such. The onces above the table is running parrallel to the lentth of the table.

Please help.

Thanks


Photon Map

With Irradiance

[kippu]

anti aliasing ??? what are you using.... try catmull rom or mitchell

[losbellos]

actually, you have a lots of window and so, I think maybe photon mapping is not really neccesarry, put vraylights on them.

otherwise I think it is antialiaging problem. lowering the treshold value and/or increasing the sampling rate helps.

for example try 0/3 0.03

[Leon_Stoop]

What resolution are you rendering?, try 3500 pixels wide.

Leon

[WWX]

Image quality

Thanks guys, I will try your suggestions and keep you posted.

kippu I am using catmull

losbello, it is almost always important for me to reflect the correct shadow cast for the the building location for the time and day of the year. Will using vraylights provide me with this?

Leon I was rendering at 640x480. Can I reuse my photon map and only recalculate the irradiance map or do I need to recalculate both?

Thanks

[losbellos]

well, it depends how do u use it... Anyway what I meant use the one which makes the sunlight into the interior and vraylight to the windows as
skylight portal. (by the way, photon mapping would never shows any shadows from the opposite sides of the house (aginst sunlight direction)
but vraylight yes. And that is far more realistic, then current setup, I mean if u are so much concerned about it

[WWX]

VRaylight

Hi losbellos,

All I trying to establish is if I where to use Vraylights if the shadow casting would still be technically correct. Its because of my lack of knowledge about them and I was not shooting the suggestion down Anyway its more a ? of illustrating to the client (homeowner) what the expected sunlight for the space would be for a given time and date.

So could you please clarify the following for me.

I have one max sunlight ligthing the scene and its location with regards the interior is obvious from the shadows casts. Would it make a differance to how the light enters the interior if I where to place Vraylights switch to skylights at the window openings on the opposite side? How would this work? Also how does a skylight in general work? When placing one we notice a arrow point in a direction. Does this mean it focus its engery only parallel to the arrow over the size of the skylight or does it radiate its engery out in the direction of the arrow, but covering a hemisphere in this direction?

Also with is more effective, placing one skylight covering all opening per side or to place a skylight at each opening individually.

Thanks

[WWX]

Antialias

Hi,

so far I have had only time to test the antialiasing suggested by losbellos and it worked wonders for the elements closes to me.



Cool

[losbellos]

put vraylight as a skilight portal and see what happens. Thats only I can say about it. Look for some refernce photo I think thats the best chioce. There is now way just set the time and press the render button. You have to tweak and feel.

U can still increase the samplingf rate of the AA or lower the treshold even more.
U dont need to render the whole thing just hide everything else and then render it even in raytrace. Put some light to test the AA of that area...
So you get the right values quickly.

[Gijs]

THis has nothing to do with the aliasing problems, but I'd like to mention that the photon map solution looks a bit too rough to get good GI details. Your objects seem to be washed out and floating.

Gijs

[WWX]

Floating Objects

Hi Gijs,

I agree with you, but don't know how to rectify it. I have been trying quite a bit of differant options and this was the best time/quality/file size I got so far. (the photonmap is 288mb in size).

Also I am working in mm and scale 1:1. Most other photon studies the users worked in imperial units and the conversion does not seem to be a simple 10 or 100.

Cheers

[natty]

you need to play with the search distance for the photon map and get it nice and tight.

[WWX]

Photon map

Hello Natty,

For what characteristics should one look in order to determine a good photon map compared to a bad one?

Is this a step in the right direction compared to this first one posted?


Thanks

[natty]

sure is .... much tighter ... now try to get rid of the black areas as much as you can .... the tighter the solution, the better detail you will acieve.