http://www.gigapxl.org/gallery.htm

http://www.gigapxl.org/gallery-GGSunset.htm


This site, displaying photographs taken with a
eight gigapixel camera seems to show evidence that
the information contained within a povcomp style
image is better reprensented at detailed levels
than the same scale of data is represented through
optical photography.

I.E., the loss of edtail and resolution evident in the
sectional displays of the overall photograph would
not occur with a Pov-Ray generated image.

Anyone agree, or see else of significance?

Pan

> http://www.gigapxl.org/gallery.htm
> 
> http://www.gigapxl.org/gallery-GGSunset.htm
> 
> 
> This site, displaying photographs taken with a
> eight gigapixel camera seems to show evidence that
> the information contained within a povcomp style
> image is better reprensented at detailed levels
> than the same scale of data is represented through
> optical photography.
> 
> I.E., the loss of edtail and resolution evident in the
> sectional displays of the overall photograph would
> not occur with a Pov-Ray generated image.
> 
> Anyone agree, or see else of significance?

Classical Jpeg vs Lossless...
All digital camera do compress (with loss).

Moreover, even 8 Gigapixel camera do not have a 24 bits definitions
(and therefore, if you ever consider using povray with a 48bit PNG...)

And the glass-lenses used for the camera is either very (very very) 
expensive or just not as good as it used to be for the old chemical 
camera (the expensive one, also and at least).


-- 
This is an unauthorised cybernetic announcement.

When someone says "I want a programming language in which I need only
say what I wish done," give him a lollipop.

Le Forgeron wrote:

> Classical Jpeg vs Lossless...
> All digital camera do compress (with loss).

I think the OP meant that POV images can be "zoomed in" without any loss 
of detail. This is obviously not the case in real photographs.

That is hardly a surprise as POV images are made from mathematical 
models that don't necessarely have any precision limits. There is also 
no contrast reducinbg MTF involved so one can just create any kind of 
close up. The only limit is probably the accuracy of floating point numbers.

pan wrote:
> http://www.gigapxl.org/gallery.htm
>
> http://www.gigapxl.org/gallery-GGSunset.htm
>
>
> This site, displaying photographs taken with a
> eight gigapixel camera seems to show evidence that
> the information contained within a povcomp style
> image is better reprensented at detailed levels
> than the same scale of data is represented through
> optical photography.
>
> I.E., the loss of edtail and resolution evident in the
> sectional displays of the overall photograph would
> not occur with a Pov-Ray generated image.
>
> Anyone agree, or see else of significance?

It's a bit of a meaningless comparison.  I can render a 1 mm sphere on top
of a 6400 km sphere, then zoom in on the little sphere and it will all be
perfectly sharp.  No optical system can match this.

The realy challenge is, can *you* model enough detail in POV to make it look
real at those zoom levels?  I doubt it.

news:4207ae97@news.povray.org...

> Anyone agree, or see else of significance?

If you zoom in on a primitive, or anything algorithmic, you don't really see
additional, meaningful information: a procedural rust texture stops looking
like rust at some point. In fact, it's always a little bit frustrating to
work on large CG pictures because the resources (time, RAM, GHz) necessary
to add detail at large sizes are always limited. CG pictures, after all, are
based on models of the world, and are not the real thing.

In a photograph, the details may be lost for a given range of lenses, but,
up to molecular level, there's always the possibility to see more meaningful
constructs with another type of lens (or technology).

G.

-- 

**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters

"Gilles Tran" <tra### [at] inapginrafr> wrote in message
news:4208c2bb@news.povray.org...

> news:4207ae97@news.povray.org...
>
> > Anyone agree, or see else of significance?
>
> If you zoom in on a primitive, or anything algorithmic, you don't really see
> additional, meaningful information: a procedural rust texture stops looking
> like rust at some point. In fact, it's always a little bit frustrating to
> work on large CG pictures because the resources (time, RAM, GHz) necessary
> to add detail at large sizes are always limited. CG pictures, after all, are
> based on models of the world, and are not the real thing.
>
> In a photograph, the details may be lost for a given range of lenses, but,
> up to molecular level, there's always the possibility to see more meaningful
> constructs with another type of lens (or technology).

And here's a demonstration of this idea:
http://www.tpd.tno.nl/Pics/DII/gigazoom/Delft2.htm

This is made by stitching together around 600 photos taken with a Nikon D1X and
one of Nikon's Nikkor AF VR 80-400mm f4.5-5.6 lenses.  And that's *only* a 400mm
lens... put something like a Nikkor 1000mm lens on and...

I guess the moral of this story is, in the real world when you zoom in, there's
more detail (and something that might not appear to have much detail from far
away has more detail the closer you get).  With a virtual 3D construct when you
zoom in, there's less detail (and something that might appear to have a lot of
detail from far away has less detail the closer you get).

Real world for win!

Lance.

thezone - thezone.firewave.com.au
thehandle - www.thehandle.com

in news:Xns95F772E1020F3jgrimbertmeandmyself@203.29.75.35 Le Forgeron 
wrote:

> All digital camera do compress (with loss).
> 

No, SLR's and the top viewfinder cameras support a 'raw' format that is 
uncompressed. It is kind of a direct output of the sensor. With the right 
software these data can be turned into (at least) 12 bit colour images.

Ingo

"pan" <pan### [at] syixcom> wrote:
> I.E., the loss of edtail and resolution evident in the
> sectional displays of the overall photograph would
> not occur with a Pov-Ray generated image.
>
> Anyone agree, or see else of significance?

Hmmm, ask me that again after you produce a render at 20ftx10ft with the
kind of detail in that Balboa Park photo...

I think the real issue isn't how much detail is lost, but how much detail
exists in the original. And you probably need (at least) a supercomputer to
model & render something like that.

Be loads of fun though!

RG

watch and read what I found at http://www.gigapxl.org/gallery-Parasail.htm, 
it's hilarious

Eli wrote:
> watch and read what I found at http://www.gigapxl.org/gallery-Parasail.htm, 
> it's hilarious 

Haha.. Wonder if they're getting a good enough view of the scenery below :)

-- 
~Mike

Things! Billions of them!