Alain <ele### [at] netscapenet> wrote:

> Have you tyied method 2?

I hadn't, but following your suggestion I tried it on the new image. It
seems rather better at dealing with bright speckle artefacts (not visible
in either image, I hope) with lower threshold settings. I only used
threshold 0.3 for the new image, which is why some of the posts surrounding
the exhibit pits have disappeared.

> To help reduce the artefacts you can try using 1 or 2 more pretrace steps, it can
help somewhat. Try
> pretrace_end 0.000625 or 0.0003125
> Increasing count normaly help, but I'm not sure that it can realy help in this case.
> Using a smaller low_error_factor than the default (0.5) may help in some case. It
only affect the
> last pretrace step.
> The artefacts are mainly related to your yellow pannels, maybe some tweaking in
there finish,
> pigment or dimentions and placing could help. Some suggestions:
> - make the pits deeper and lower the panels.
> - make the pannels a bit smaller than the pits. or enlarge them to make them larger
than the pits.
> - a little lower ambient with a higher saturation for the yellow pannels.
> - make the yellow pannels transparent and put some weak yellow area_light spot_light
under them.

Thanks for this advice! I will try out your suggestions.

Post a reply to this message

I think that looks a lot better. It makes the ship's textures pop a bit
more.

Post a reply to this message

"Ross" <rli### [at] everestkcnet> wrote:
> I think that looks a lot better. It makes the ship's textures pop a bit
> more.

Yes, agreed. In other (less uniformly lit) scenes using those ships, the
material seems to go from plastic to metal just with the increase in gamma.
I wonder if it's better matched to the textures, or something. Anyway:

To get rid of some of the artefacts, I followed part of Alain's excellent
advice. The yellow light panel buried in the floor pits has been lowered
further, and turned down a bit. I also decreased the pretrace_end value.
This did have the side effect of punishing my computer with swapping for
about a day, but the artefacting is less noticeable and I think the
radiosity solution is of higher quality. The fog has been slightly
increased compared to the last attempt. These images use antialiasing
method 2 with threshold 0.1. Attached is a smaller version of the image
with assumed_gamma 2.2. A full-size PNG image has been placed at

http://www.zubenelgenubi.34sp.com/temp/tvo_hg.png

(1280x720, 430K)

To appease Tek, here is the new radiosity solution with assumed_gamma 1.0,
and without the fog or additional sunlight. The main difference is the
colours are a little fainter, but they both look similarly "clean" or
airbrushed to me.

http://www.zubenelgenubi.34sp.com/temp/tvo_lg.png

(1280x720, 420K)

Thanks for all the advice!

Post a reply to this message

Attachments:
Download 'tvo_sml_hg.jpg' (70 KB)

Preview of image 'tvo_sml_hg.jpg'

It is absolutely perfect for a "neutral", "boring",  or "daylight" mood.  I 
mean perfect in that it looks like a photograph.  If you want it to be 
exciting, to have the lighting to show a mood, you might have some work to 
do, but it would be hard not to make it "annoyingly obvious raytracing".

Post a reply to this message