POV-Ray : Newsgroups : povray.competition : Detail Views Server Time
29 Mar 2024 06:11:47 EDT (-0400)
  Detail Views (Message 4 to 13 of 13)  
<<< Previous 3 Messages Goto Initial 10 Messages
From: Renderdog
Subject: Re: Detail Views
Date: 26 Aug 2004 14:15:00
Message: <web.412e277f2c3552b96bcddb6d0@news.povray.org>
"Jeremy M. Praay"  wrote:
> Have I interpreted that (in)correctly?

It looks like a correct interpretation to me. That's a lot of
rendering for a contest not much longer than an IRTC round,
especially since many of us will want to push the envelope on
features, etc.

Not to highjack this new thread, but I also have a question
relating to the image sizes...

I notice the guidelines state an image larger than 1600x1200
will be scaled down to screen size for judging. I've decided on
an image that will be around 1.4* taller than it is wide. To meet
the minimum pixel count of 1,228,800 my submission will be
more than 1200 pixels tall, so it will be scaled down to 1200
pixels tall for judging.

The resulting image will be about 857 pixels wide (1200/1.4),
which means tall images will be judged at a lower pixel count
than wide images (unless they're really wide). Am I interpreting
this rule correctly?

The full size image will still be used as additional information,
just like the separate detail views, but I wonder if this puts tall
images at a disadvantage in the judging?

-Mark Slone
From: Gilles Tran
Subject: Re: Detail Views
Date: 26 Aug 2004 15:35:48
Message: <412e3b94@news.povray.org>

news:412dee31$1@news.povray.org...

> "The minimum size for each of the detail views is the same as for the main
> render ; the area of each detail view must not be larger than 1/100 of the
> main image area (meaning no more than 1/10 the width and 1/10 the
height)."
>

Damn, you're right. I'll see what I can do about this.

G.

-- 
**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters
From: Peter J  Holzer
Subject: Re: Detail Views
Date: 26 Aug 2004 17:00:15
Message: <slrncisjeu.lp.hjp-usenet2@teal.hjp.at>
On 2004-08-26 17:23, Slime <fak### [at] emailaddress> wrote:
> But aside from that, my monitor is 1280x1024, and I won't be able to see a
> single image all at once. I'll have to scroll to see every single one of
> them. I might be able to alleviate this problem by viewing in IE's full
> screen mode

Or with Mozilla, which can scale images down to fit into the browser
window.

	hp


-- 



__/   | http://www.hjp.at/ |   -- Hannes Petersen in desd
From: Gilles Tran
Subject: Re: Detail Views
Date: 26 Aug 2004 18:00:19
Message: <412e5d73$1@news.povray.org>

news:web.412e277f2c3552b96bcddb6d0@news.povray.org...

> The full size image will still be used as additional information,
> just like the separate detail views, but I wonder if this puts tall
> images at a disadvantage in the judging?

Tall images are always at a theoretical disadvantage since computer screens
are horizontal. Now, frankly, I doubt that it will make a difference, since
the judges will have the full one anyway. All my IRTC entries - like most of
my POV-Ray work in the past years - had odd ratios (vertical or panoramas)
that didn't do them justice on screen, and it didn't really hurt them...

G.

-- 
**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters
From: Slime
Subject: Re: Detail Views
Date: 26 Aug 2004 21:11:49
Message: <412e8a55$1@news.povray.org>
> Or with Mozilla, which can scale images down to fit into the browser
> window.


IE can do that too (and actually does by default), but does Mozilla do a
nice job of it? (That is, how does it resample: nearest neighbor or a smooth
averaging of the pixels / linear interpolation?)

 - Slime
 [ http://www.slimeland.com/ ]
From: Jim Charter
Subject: Re: Detail Views
Date: 27 Aug 2004 00:33:41
Message: <412eb9a5$1@news.povray.org>
Slime wrote:
>>Or with Mozilla, which can scale images down to fit into the browser
>>window.
> 
> 
> 
> IE can do that too (and actually does by default), but does Mozilla do a
> nice job of it? (That is, how does it resample: nearest neighbor or a smooth
> averaging of the pixels / linear interpolation?)
> 
>  - Slime
>  [ http://www.slimeland.com/ ]
> 
> 
No, in FireFox at least, it stinks.  Not sure if I would say IE is *as* 
bad but personally I would not use a browzer at all for this purpose.
From: Jeremy M  Praay
Subject: Re: Detail Views
Date: 27 Aug 2004 08:15:10
Message: <412f25ce@news.povray.org>
"Gilles Tran" <gitran_nospam_@wanadoo.fr> wrote in message
news:412e3b94@news.povray.org...

> news:412dee31$1@news.povray.org...
>
> > "The minimum size for each of the detail views is the same as for the
main
> > render ; the area of each detail view must not be larger than 1/100 of
the
> > main image area (meaning no more than 1/10 the width and 1/10 the
> height)."
> >
>
> Damn, you're right.


It was bound to happen sooner or later, I suppose...

> I'll see what I can do about this.

My current test render is running at 5pps, and I still have a ton of work to
do.  Now is the time when I start making sacrifices, I suppose...

-- 
Jeremy
www.beantoad.com
From: selsek
Subject: Re: Detail Views
Date: 27 Aug 2004 17:45:00
Message: <web.412faadc2c3552b9f320a5370@news.povray.org>
As Gilles said: detail images are here to show that size reduction hasn't
affected the image quality, or details loss.

So you don't need to have an exact "extract" of the image. The perspective
may be changed, etc.

Can you go further? For example: i made a detailed model for my scene, but
you don't see all of it, or there's a shadow masking some details, etc.
Can i make a "model show"/demo instead of having an image extract (you know,
several point of views for one lone object)?

Could that be interpreded as "show us how far you've gone into details by
sending us two detailed views in 1024*748 minimum"? This would be less
restrictive, and easier to understand (i still haven't understand what
those detailed views should be... not all povers have a high english
level!).
From: Ross
Subject: Re: Detail Views
Date: 27 Aug 2004 17:49:56
Message: <412fac84$1@news.povray.org>
"selsek" <sel### [at] despammedcom> wrote in message
news:web.412faadc2c3552b9f320a5370@news.povray.org...
> As Gilles said: detail images are here to show that size reduction hasn't
> affected the image quality, or details loss.
>
> So you don't need to have an exact "extract" of the image. The perspective
> may be changed, etc.
>

i guess i was under a different assumption, which is that the detailed
versions were to show that the winning image would be suitably detailed to
look good when printed as a poster. and if that were the idea behind it,
then you would in fact need to provide a detailed view in the same
perspective to illustrate your attention to detail.
From: Gilles Tran
Subject: Re: Detail Views
Date: 28 Aug 2004 10:49:39
Message: <41309b83@news.povray.org>

news:412fac84$1@news.povray.org...

> i guess i was under a different assumption, which is that the detailed
> versions were to show that the winning image would be suitably detailed to
> look good when printed as a poster. and if that were the idea behind it,
> then you would in fact need to provide a detailed view in the same
> perspective to illustrate your attention to detail.

That's it. Typical example: let's imagine that the winning image contains a
mesh textured with an image map (correctly interpolated). It looks perfect
(and it is) at screen size. Now this image is used for the cover of a CG
magazine, and the requirement from the publishers is for a 20x30 cm, 300 dpi
image, roughly 2400 x 3600 pixels. Aw, suddenly the mesh looks coarse and
angular and the texture all blurry. Or (a problem I had in the past when
porting screen-sized image to the print world) objects that looked properly
positioned appear now floating over each other. Also: primitives that look
really primitive, normals that look really flat, this sort of thing.
The main idea is that the image needs to look perfect on something else than
a computer screen, so that modeling, texturing and detailing should be tuned
accordingly.

G.


-- 
**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters
<<< Previous 3 Messages Goto Initial 10 Messages

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.