Am 01.12.2010 03:30, schrieb Stephen Klebs:

>> As a result, a truly linear gradient doesn't /appear/ linear to us: The
>> "distance" between 10% and 20%, for instance, is percieved as roughly
>> equal to that between 50% and 100%. Go figure.
>>
> What you are missing here is that this ability for the brain to retain a
> constant relation between widely varying changes in size, shape, colors or
> perspective is not something you need to correct in the picture. The brain does
> it. It works in a photograph as well, which is just a neutral recording device.

You'd be surprised how non-neutral photographic paper can be.

Fact is: A grey tone that looks like halfway between black and white 
does /not/ reflect half as much light as a pure white, but a lot less, 
and this fact causes a lot of problems wherever you try to compute the 
addition of light intensities (which is a pretty common task in 
raytracing), as you'll find out that "50%" perceptual brightness plus 
"50%" perceptual brightness does not add up to "100%" perceptual 
brightness in reality. So if you try to do your math with perceptual 
brightness rather than physical, you'll end up getting nonsense in the 
output image that your brain cannot fully compensate; sometimes you'll 
not even be able to pinpoint them because your brain does its best, but 
you'll still be able to "feel" that the image is CG).

Post a reply to this message

Am 01.12.2010 01:48, schrieb Stephen Klebs:

> With regard the new gamma model that's being applied to POV, two points: POV is
> first a description language not like Photoshop were you just slide sliders back
> and forth until things look right. There is nothing wrong and even wonderful
> about this approach as long as you have immediate visual feedback. In POV you do
> not. It's like taking photographs in the pre-digital age. You use an educated

I fully agree here: POV-Ray should not be a tool where you need to tweak 
and try until everything looks right.

But that's exactly the point: As long as you don't do proper gamma 
handling, you /will/ need to tweak and try - not for the individual 
colours, but for the scene lighting - to get it look /somewhat/ 
convincing (and it's guaranteed that you /cannot/ get it completely 
right except for trivial scenes).

With proper gamma handling, specifying colors gets a little bit more 
complicated if you're used to color values from image processing 
software, but a simple rule of thumb to always put "gamma 2.2" or "gamma 
srgb" after any color literal should get you going. And you'll have a 
much easier time lighting your scene in a convincing way.

> curve but is a curve. So to get to the point of the gradient example. It tells
> POV that we want a linear succession values over a certain distance what it does
> is give an exponential curve. That's fine but not what I intended and, if you
> can't depend on 1 plus 1 plus 1 etc coming out a 3 or 4 or what ever, how do you
> use a language that's going to be retranslated to mean something totally
> different than what you expected it to say.

No, the point here is that what you expected wasn't what you told 
POV-Ray to do, because despite all your knowledge about human visual 
perception you forgot that a linear succession of brightness values - 
which is what you told POV-Ray to do, and what POV-Ray did - isn't 
/percieved/ as linear.

So essentially you told POV-Ray to add 1^g plus 1^g plus 1^g etc, and 
expected it to come out as 3^g or 4^g or whatever.

> In 3.6 the evenly ramped gradient looked right. The values reported in Photoshop
> or whatever said yes indeed 1 plus 1 is two. I've used POV in every possible

Photoshop is a liar when it comes to color maths.

Ever tried to blur a pattern of alternating black and white horizontal 
stripes, and wondered why the result would be darker than expected?

That's because Photosop thinks 1^g + 1^g = 2^g. Unfortunately that's 
only true for g=1, while normally it will be about 2.2.

> way, on Macs and PCs, to make images that looked fine in IE and FireFox or
> Chrome or in print, and have in fact never encountered problems as serious as
> this major revision seems to solve. I would like to see a concrete scene we
> could render in both versions where it is demonstrated the need for the change.

See p.b.images.

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> Am 01.12.2010 02:32, schrieb Stephen Klebs:
> >>
> >> No, it's your eye playing tricks on you.
> >>
> >> Our eyes are highly calibratible optical measurement tools - they work
> >> as well at a brightness of 10% as at 1,000% (and then some), and get
> >> object colors right regardless of lighting conditions.
> >
> > This is not correct. You only understand the physics of the optical system, but
> > not the physiology nor the psychology. This is not how we see. You are talking
> > as if you could see yourself seeing and you can not. What you are seeing is a
> > model of how you think things appear. This is a very subtle and complex mystery.
>
> I'm well aware that I'm simplifying here, but I'm not holding a lecture
> in physiology nor psychology here - I'm merely making a point that
> visual perception of brightness does not match physical brightness when
> it comes to gradients.

Wow! this is really a tough one. Like a political standoff between
picture-making as art or as science. There seems some kind of blinders on our
perspectives. Like "you just don't get it!" "No you don't get it!". But I think
we get a feeling for each others approach. So yes, "visual perception of
brightness does not match physical brightness". You seem to somehow to be
wanting to "correct" the visual representation of gradients to more realistic
reflect the real visual world. And I'm saying this can't be done and doesn't
need to be done. You seem to be saying the gradient that looks to me perfectly
smooth and convincing is really a deception. Whereas, the one that looks skewed
and unbalanced is really how the physics of things really is. Like this picture
may look flat and washed out but it's fine because that's how the optics of the
situation really is. Like we both think the other is deceiving themselves. One
viewers reality, the others illusion

Post a reply to this message

clipka wrote:
> as you'll find out that "50%" perceptual brightness plus 
> "50%" perceptual brightness does not add up to "100%" perceptual 
> brightness in reality. 

Of course. The speed of light is constant, so 50% plus 50% doesn't add up to 
100% either.

-- Darren, not helping matters. ;-)

Post a reply to this message

Christian Froeschlin <chr### [at] chrfrde> wrote:
> I have to admit I don't understand this either. Until now I
> thought that since 3.7 now uses a linear color space, 0.5 would
> represent a true midgray and all the nifty gamma handling ensures
> that everyone perceives that midgray on their display when viewing
> the image file. So I'd also have expected the end result of your
> test scene to visually yield evenly spaced brightness steps.

  Well, there's a traditional test to see if rgb 0.5 (as produced by
POV-Ray 3.7) is equivalent to half-brightness of your display: Make
an image which is 0.5 gray all over, and another which has horizontal
lines of pixels alternating between black and white. Put these two
images side-by-side and then watch them from far enough that you don't
distinguish the individual lines in the second image. Compare the
brightness of the two images. Make the same test with a rgb 0.5 image
produced by POV-Ray 3.6.

  When I made this test, POV-Ray 3.7 produced approximately the proper
brightness (the two images looked very similar from far away), while
POV-Ray 3.6 produced an image which was *clearly* darker than the
alternating-lines test image.

  Now, if you do make a gradient from black to white with POV-Ray 3.7,
the transition doesn't look linear for some reason, even though the
midpoint has exactly the correct brightness to be 50% gray. I don't know
what causes this.

-- 
                                                          - Warp

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:

> Photoshop is a liar when it comes to color maths.
>
You've obviously never worked as a graphics artist. Eventually, everything goes
through Photoshop at one point or another. If Photoshop is a "liar", then the
lie is the truth. Us kids have to play with a lot of multi-colored graphics
tools - raster and vector and tracers and converters and raytracers and scanline
and cad, etc. - but we can't play nice together without some supervision and,
like it or not, Momma Photoshop holds the ruler. Otherwise, we can just play
only in our own backyard, with only are own toys. Why we accept it -- or any
other well thought out graphics program is that, whatever it's bias, it is
equally tolerant of many approaches. If you don't like things one way, you play
with the levels or curves or color profile or styles until it looks like you
want it, something that says "hey, that looks cool", which is not necessarily
the "proper" -- whatever that is -- behavior. All those shaders and sliders and
tweakers that you hope to avoid are really very necessary because the relation
between the painting and the painter, the seeing and the seen, is relative and
readjusting, with only the eye the final judge. We say "no that's not quite
right, a little more this, a little less that". There will always be in any
picture making inevitable adjustment and readjustment. It's in the nature of the
visual process. Nothing ever comes out just as we want the first time. POV, on
the other hand, can not make use of such real-time adjustments -- as yet. It's a
language. A language needs dictionaries like Photoshop to tell us what thing's
mean.

Post a reply to this message

> Bluntly put. You're making this much too complicated. It's certainly less
> complicated than dealing with color management for print. You make a 
> picture
> that looks good on your computer. It doesn't matter how you do it.

In something like PhotoShop, sure, but POV is almost a light simulation 
tool, it needs to follow the correct laws of physics in the first place. 
Sure you can botch it and create something that looks realistic, but it 
won't be exactly physically correct, and usually would require a great deal 
more skill and effort from the artist.  POV works on the fact that if it 
deems something should be 50% brightness, it will make it look 50% 
brightness to you (which is not the same as 50% pixel value).  If it didn't 
work this way it becomes impossible to get scenes with physically correct 
lighting, and then you need to start the whole skillful and time-consuming 
"bodge" process of trying to make it "look right".

Post a reply to this message

"Stephen Klebs" <skl### [at] gmailcom> wrote:
> clipka <ano### [at] anonymousorg> wrote:
> > Am 01.12.2010 02:32, schrieb Stephen Klebs:

> Wow! this is really a tough one. Like a political standoff between
> picture-making as art or as science... [clip].. You seem to be saying the
> gradient that looks to me perfectly
> smooth and convincing is really a deception. Whereas, the one that looks skewed
> and unbalanced is really how the physics of things really is.

Yes, succinctly put. I'm feeling the same way, like my world has suddenly turned
upside down. :-O  There are compelling arguments on both sides--and the
arguments are diametrically opposed. It will be interesting to see how this
Mexican Standoff develops further... :-P

What all of this says to me (and I'm not saying it's wrong) is that, from now
on, it's "POV-Ray against the world!" The 'world' meaning, all the *other* apps
(like Photoshop) that we're so used to working with, with their apparently
flawed way of handling gamma.

I have a devilish little question: Does anyone know how BLENDER deals with the
gamma issue? I've never used it, but it would seem to be a decent comparison
tool vis a vis POV-Ray.

Ken

Post a reply to this message

>  Now, if you do make a gradient from black to white with POV-Ray 3.7,
> the transition doesn't look linear for some reason,

The key point here is "look" linear.  If you did the gradient with 3.7 and 
measured your monitor with a luminance meter you'd find it was a linear 
gradient.  Your eye/brain however doesn't see that as a "linear" gradient. 
(There are various transformations to convert absolute luminance into a 
"perceived" luminance, and they're not linear).

BTW all this (about how your eye/brain works) is completely unrelated to how 
POV should work, POV should attempt to make your monitor display what the 
scene would look like if it existed IRL (just like how a camera works).  In 
this respect if IRL there existed a truly linear gradient from 0% to 100% 
then POV should attempt to make it look the same.  Which it does in 3.7.

Post a reply to this message

On 01.12.2010 14:31, Kenneth wrote:
> What all of this says to me (and I'm not saying it's wrong) is that, from now
> on, it's "POV-Ray against the world!" The 'world' meaning, all the *other* apps
> (like Photoshop) that we're so used to working with, with their apparently
> flawed way of handling gamma.
>
There is and never was something like POV-Ray against the world. 
Photoshop is a great tool (I'm, using it since the time it was not even 
called Photoshop but Photostyler and was developed by Aldus and not 
owned by Adobe - and a few weeks ago did update to CS5) but it is also 
still flawed in multiple ways and has (as every complex piece of 
software) numerous bugs.

but as Stephen Klebs wrote:
 >like it or not, Momma Photoshop holds the ruler.
...and...
 >A language needs dictionaries like Photoshop to tell us what thing's
 >mean.

well, like it or not, but the only one who will be happy about such 
statements is the marketing department of Adobe - I for one do report 
the usual found bugs back to Adobe and do never trust blindly any piece 
of software. And just as a side-note: the long history of Adobe 
Photoshop is full of changes that also caused particular image file 
formats written by some older version to be rendered completely 
different with the new one.

I'm also using frequently Blender as a modeller and always POV-Ray as 
final render engine for it.

In fact and for short: since the applied gamma handling changes in 3.7 
making POV-Ray work together with applications like Photoshop and 
Blender has become much easier, less painful and without time consuming 
workarounds.

And to put it a bit rude: the only thing that *is* apparently flawed is 
the knowledge of most users about the tools they use - and well, I'm not 
talkin' especially about POV-Ray here ;)

-Ive

Post a reply to this message