Hi Lance,

Perhaps you can clarify something for me: When you say, your LCD is
calibrated with a 2.2 gamma response curve, does that mean your
software/hardware adjusts the gamma automatically? If so, you do not
represent the majority of people who, like myself, uses a system that does
not feature automatic gamma correction, and therefore needs pictures that
have the gamma correction "built in".

I know it isn't possible for my images to look good on all monitors but I
try to accomodate most people.

> I was talking about *just* the speaker -
> there's good luminance range even when
> excluding all areas of the blue backdrop,
> in fact from 0% to nearly 100%
>(around 250/255 - i.e. ~98% luminance).

If you say so. I believe you, but it still doesn't look good here. And most
professional work DO look good on my monitor, so I assume this image has a
way to go. Besides there is a difference between histograms and "the minds
eye". Maybe that is part of the reason. I have distinct ideas about, what
this speaker could look like, and should look like, in a good lighting
enviroment.

> Removing radiosity would lose this subtly
> and make the image not seem as realistic

Yes, true. I've had much fun playing with radiosity, and with scenes that
are solely lit by radiosity, so they behave more like the real world. But
sooner or later we always run into the barrier of extremely slow renders,
and we realise that even those renderings are miles away from complexity of
the real world. Even if the light bounces around correctly, it doesn't
respond well to our faked and always very simplistic textures / materials.

So I encourage people to try and achieve good lighting conditions with
oldfashioned pointlights / spotlights, because it teaches us some important
lessons. If a scene looks great without radiosity, it means we have done a
good job setting up light, and radiosity will add the final touch. On the
other hand, if we have done a poor job setting up light, radiosity will not
help much.

It has come to my knowledge that lots of professional work today does not
use radiosity, or any other kind of advanced algoritm for solving
global-illumination... In movies, G.I. are often faked with fill-lights or
painted into the bitmap textures (eventually baked into them). For example,
there is a new movie out called "Troi". I haven't seen it, but I have read
about the CG sequences of the city Troi and some battles on the fields. It
looks pretty realistic but does not use G.I.! The artists explain that G.I.
would have been impractical. For one thing, it would have been much slower,
but even the large studios and render-farms run into problems with memory
capacity, and even the most expensive render-engines cannot always handle
an animation without G.I. artifacts. (They should've used POV-Ray instead,
hehe... well)

I don't discourage the use of G.I. either! Especially in complex indoor
enviroments with walls, and detailed objects with many holes and creases
where the light creeps in, radiosity is certainly the best solution. And in
still-images it's often not much of a speed issue, provided it's used in
conjunction with pointlights. But radiosity is only an aid; it does not
replace the need for good, manual lighting.

Regards,
Hugo

Post a reply to this message