[Hervé]

mmmm, ok, I explained it wrong... (damn english.. hehe)

normally, to obtain a good HDRI you blend (mix) all your LDR's together with let say HDRShop.. and you obtain one single hdr... but the right thing would be to cut (select) in each LDR the part that is right exposed... and do a sort of montage... maybe... very tough job..

... but I do get what you say Hyltom...

[giacob]

i agree with hylton..... would be preferable if we have images that match what our eyes perceive in reality rather tha having image that imitate what a camera is capable of reproduce since human eyes give to our mind a different result than a camera ... surely more near to what we really perceive...
.... this question remind me of another old question: mathematich perspective reproduce our perception of 3d object ?.. also in these case oure eyes make some correction that make what we perceive different from the results of appllication of scientific perspective.....
in this case though if u look at a picture from the very point of view ( scaled in relation to the scale of the picture) from wich the picture is taken the coincidence is perfect... perspective aberration disappears...

[Hervé]

The brain edits the scene...

he he....

[Fernando Tella]

If we get so picky about what we see with our eyes we should consider too that we only see crisp details at the point we are focusing to and all around it should become blurry, cause our eyes only get the details at the macula (don't know if that's the name in english), the most important and sensitive point of the retina.
If we move the eye we are taking another picture (making another render).

Just making a racket.

[Boris Ulzibat]

The brain edits the scene...

he he....

No laugh! It really does!
Remember that almost right in the center of each eye there is a blind spot - NOTHING can be seen with it. Now close one eye and look at something - do you get a black empty circle anywhere? That's because you brain edits this spot out, filling it with parts of the scene close to it (think of it as clone stamp tool in PS). you can see the effect by placing 2 dots horizontally approx 10 cm away from each other on a piece of checkered paper close 1 eye, look on one of the spots, then slowly move the paper from and to yuorself. You'll notice the second dot disappears at some point but the checkers dont!
Pity i forget

[iandavis]

we don't actually see the full dynamic range at one time. We do the same thing as any digital camera, as we look out a bright window our eyes 'stop down' and our brain re-balances for the exterior light values. Granted, your dynamic range for vision is substantially higher then a digital camera, but not wide enough to see everything all at once as some people have suggested.

Our brain is the magic tool here. It fools us into thinking we see a contiguous world when in fact our eyes are just as limited (more so actually) then most lenses, and our retinas are better then film in most respects, but by no means capable of dealing with the full dynamic range of light in nature simultaniously.

it's a tough call. do we render something with blown out, overexposed windows as it actually is... or as the fake world our minds build for us. We build the picture in our heads by scanning repeatedly and creating a persistant 3D representation of the room/outdoors. If that ability were lost, it would seem like we were viewing the world through a paper towel core.

Personally I usually try to photograph and render scenes that lean more toward reality then perception. Which translates to perfectly exposed interiors with overexposed exteriors... or underexpose the interior a few stops and expose the exterior properly. By perfectly exposing both areas one is saying that we are seeing two light values - thousands of times brighter/dimmer - as the same, which raises red flags in most viewers.

Most digital sensors are capable of capturing 8 or 9 stops at once... our eyes are a few stops better... but to capture bright sunlight AND interiors you would need even more. So assuming the exterior is perfectly exposed, the interior should be 3-5 stops below. Putting the highlights of the interior in the middle grey area. The trick is to compress the interior exposure so that the highlight and shadow all fall within 5 or so stops. In this way it looks like a feasable exposure, even though to actually capture it in real life would require two exposures and compositing.
cheers

Fernando - it's called the 'fovea' The macula is the broader area which includes the fovea, which is where the light is theoretically focused. Assuming your not near or far sighted. ;P

[iandavis]

boris,

yeah, the nerve bundle blindspot... not very big, but exactly what i'm talking about. Vision is a fabrication of the brain.

We also have very limited color vision in the dark. The fovea is packed with cones and in the dark they are almost non-functional. Our highest resolving power is in the fovea so in the dark we basically are using our greater retinal surface to see. Though almost nobody feels that at night they see in monochrome. Our minds 'fill in' color where possible. Try looking at well known objects under a sodium vapour light. You will SEE colours which are physically impossible to see under those circumstances.

needless to say... it's quite illuminating and disconcerting all at the same time.

[Hervé]

The brain edits the scene...

he he....

No laugh! It really does!
Remember that almost right in the center of each eye there is a blind spot - NOTHING can be seen with it. Now close one eye and look at something - do you get a black empty circle anywhere? That's because you brain edits this spot out, filling it with parts of the scene close to it (think of it as clone stamp tool in PS). you can see the effect by placing 2 dots horizontally approx 10 cm away from each other on a piece of checkered paper close 1 eye, look on one of the spots, then slowly move the paper from and to yuorself. You'll notice the second dot disappears at some point but the checkers dont!
Pity i forget

I trust you.. I was not LOL, but just the way he said that was funny.. anyway.. back to Alaska...

[Boris Ulzibat]

I trust you.. I was not LOL, but just the way he said that was funny.. anyway.. back to Alaska...

I know you were not lol, i meant "no matter how funny it sounds - it is the way things are!"
Sorry if it was not clear enough.

[chrisnvp]

Thanks for the great read!!! I'll keep all these things in mind while I render my up comming projects

Best regards
Chris

[Leonardo]

WOW!! a complete topic about my observations.... and there's arguments and discussion............... COOL!!

[Peder]

My understanding of the tonemapping process is that you can exploit the fact that the eye (brain) is more sensitive to contrasts between surfaces than the luminance of a surface itself. There are a lot of interesting experiments showing this effect -called the Cornsweet illusion i believe:
http://www.cut-the-knot.org/Curriculum/ ... rast.shtml
http://www.purveslab.net/research/expla ... sweet.html

Or this:
http://infohost.nmt.edu/~armiller/illus ... rboard.htm

The tonempapping process increases local contrast in order to create an illusion of a greater dynamic range in order to trick the eye into believing there is more information in the image. Overdoing this creates the typical halo effects often seen in HDR manipulated images like here:
http://www.flickr.com/search/?q=hdr+architecture&s=int

Just my 2 Cents... Interesting topic BTW.
Peder

[Becco_UK]

I like hyltoms' painting example. It matches what my eye sees inside a room and not those dreadful burnt out windows seen all too often in photo's and renders. Give me 'eye realism' anyday!

[Bubbaloo]

Leonardo, how did you post this in the future???
Check out your "posted date" above your post.
It says posted Mon, Nov 20.
So what's it like in the future?

[lebbeus]

actually he posted it in 2006…