POV-Ray : Newsgroups : povray.competition : Image Submission Reality Check Server Time: 29 Sep 2020 20:06:03 GMT
  Image Submission Reality Check (Message 1 to 9 of 9)  
From: Jim Holsenback
Subject: Image Submission Reality Check
Date: 10 Sep 2004 14:56:25
Message: <4141c099$1@news.povray.org>
Howdy,

Maybe I'm making this more complicated than it needs to be, so I just wanted 
to fling this out there too see how far off I really am on this issue.

I'm thinking just on render at 6400 x 4800 and then just cropping two 
separate 800 x 800 detail views.

I did a quick preview of my scene (no media, no antialiasing) and it appears 
that bigger image takes more time. Is that true or it it just me??

Jim
From: Christoph Hormann
Subject: Re: Image Submission Reality Check
Date: 10 Sep 2004 15:30:02
Message: <chsh9h$s28$1@chho.imagico.de>
Jim Holsenback wrote:
> 
> I did a quick preview of my scene (no media, no antialiasing) and it appears 
> that bigger image takes more time. Is that true or it it just me??

Are you kidding?

It should be obvious that without aa the render time is directly 
proportional to the number of pixels.  Twice the pixel count -> twice 
the render time.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 06 Jul. 2004 _____./\/^>_*_<^\/\.______
From: Ross
Subject: Re: Image Submission Reality Check
Date: 10 Sep 2004 15:46:05
Message: <4141cc3d$1@news.povray.org>
"Jim Holsenback" <jho### [at] hotmailcom> wrote in message
news:4141c099$1@news.povray.org...
> Howdy,
>
> Maybe I'm making this more complicated than it needs to be, so I just
wanted
> to fling this out there too see how far off I really am on this issue.
>
> I'm thinking just on render at 6400 x 4800 and then just cropping two
> separate 800 x 800 detail views.
>
> I did a quick preview of my scene (no media, no antialiasing) and it
appears
> that bigger image takes more time. Is that true or it it just me??
>
> Jim
>
>

do you mean to say that the time does not scale linearly with the size? for
example instead of a 100 pixel image taking 1 second, a 200 pixel image
taking 2 seconds, a 300 pixel taking 3 seconds it's more like 100:1,
200:2.5, and 300:6.25 or something?

i don't know what i'm talking about. just trying to figure out what you're
talking about ;)
From: Gilles Tran
Subject: Re: Image Submission Reality Check
Date: 10 Sep 2004 15:57:17
Message: <4141cedd@news.povray.org>
"Jim Holsenback" <jho### [at] hotmailcom> a écrit dans le message de
news:4141c099$1@news.povray.org...

> I did a quick preview of my scene (no media, no antialiasing) and it
appears
> that bigger image takes more time. Is that true or it it just me??

Without aa, the tracing time should be strictly proportional to the rendered
surface, i.e. tracing a 6400*4800 image takes 8x8=64 times longer than the
800*600 version.

However, it may happen that the bigger image requires more memory than the
available RAM, in which case the system starts using the swap file, slowing
down the render. Another possibility is that, since the overall render time
is longer, another program or scheduled service such as an antivirus may
start during the render...

G.

-- 

**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters
From: Jim Holsenback
Subject: Re: Image Submission Reality Check
Date: 10 Sep 2004 15:58:41
Message: <4141cf31@news.povray.org>
"Ross" <rli### [at] everestkcnet> wrote in message 
news:4141cc3d$1@news.povray.org...
> "Jim Holsenback" <jho### [at] hotmailcom> wrote in message
> news:4141c099$1@news.povray.org...
>> Howdy,
>>
>> Maybe I'm making this more complicated than it needs to be, so I just
> wanted
>> to fling this out there too see how far off I really am on this issue.
>>
>> I'm thinking just on render at 6400 x 4800 and then just cropping two
>> separate 800 x 800 detail views.
>>
>> I did a quick preview of my scene (no media, no antialiasing) and it
> appears
>> that bigger image takes more time. Is that true or it it just me??
>>
>> Jim
>>
>>
>
> do you mean to say that the time does not scale linearly with the size? 
> for
> example instead of a 100 pixel image taking 1 second, a 200 pixel image
> taking 2 seconds, a 300 pixel taking 3 seconds it's more like 100:1,
> 200:2.5, and 300:6.25 or something?
>
> i don't know what i'm talking about. just trying to figure out what you're
> talking about ;)

I did a full render (all the switches I wanted) at 1280 x 1024 and it took 
6h 16m 57s to complete
if I do a the same render at 6400 x 4800 and just pick out two detail areas 
and crop to get my detail views. The orginal 6400 x 4800 image would be used 
for the full view version. Would the 6400 x 4800 image take approx 5 times 
as long to render? Again is it me or is this more complicated that it needs 
to be? I'm having some difficulty the BEST way to go about meeting 
submission requirements.

Jim
>
>
>
>
From: Gilles Tran
Subject: Re: Image Submission Reality Check
Date: 10 Sep 2004 16:05:40
Message: <4141d0d4$1@news.povray.org>
"Jim Holsenback" <jho### [at] hotmailcom> a écrit dans le message de
news:4141cf31@news.povray.org...

> for the full view version. Would the 6400 x 4800 image take approx 5 times
> as long to render?

It will take 25 times as long to render (at most) as it's proportional to
the rendered area. If the 1280*960 version is already slow, just render the
close-ups with the +sr +er +sc +ec switches (you'll need a paint utility
such as the gimp to cut out the rendered part) or use the zoomin macro. See
http://www.povcomp.com/samples/ for an example.

G.

-- 

**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters
From: Christoph Hormann
Subject: Re: Image Submission Reality Check
Date: 10 Sep 2004 18:00:02
Message: <chsq2h$tje$1@chho.imagico.de>
Gilles Tran wrote:
> 
> However, it may happen that the bigger image requires more memory than the
> available RAM, in which case the system starts using the swap file, slowing
> down the render. Another possibility is that, since the overall render time
> is longer, another program or scheduled service such as an antivirus may
> start during the render...

The memory use does not depend on the render size (except the line 
buffers which are not significant).  So at least without display the 
large render will take the same amount of memory.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 06 Jul. 2004 _____./\/^>_*_<^\/\.______
From: Gilles Tran
Subject: Re: Image Submission Reality Check
Date: 10 Sep 2004 20:18:48
Message: <41420c28@news.povray.org>
"Christoph Hormann" <chr### [at] gmxde> a écrit dans le message de
news:chsq2h$tje$1@chho.imagico.de...

> The memory use does not depend on the render size (except the line
> buffers which are not significant).  So at least without display the
> large render will take the same amount of memory.

Not with radiosity (and perhaps photons, I didn't test).
Here is memory use for the scene below, with and without radiosity:
320 x 240 : 13 Mb (150 kb)
640 x 480 : 30 Mb (190 kb)
1600 x 1200 : 58 Mb (324 kb)
6400 x 4800 : 105 Mb (10 Mb)

Fortunately, the increase isn't proportional, but it's still important, and
this scene is minimal.

global_settings{
    radiosity{error_bound 0.02 count 300 recursion_limit 1}
}
sky_sphere{pigment{rgb x*0.25+y*0.5}}
camera{
  location <0, 4, -10>
  direction 2.2*z
  look_at   0
}
union{
    sphere{y,1}
    plane{-y,0}
    rotate y*45
    texture{pigment{rgb 1}finish{ambient 0 diffuse 1}}
}

G.

-- 
**********************
http://www.oyonale.com
**********************
- Graphic experiments
- POV-Ray and Poser computer images
- Posters
From: Christoph Hormann
Subject: Re: Image Submission Reality Check
Date: 10 Sep 2004 21:40:02
Message: <cht6tt$2j$1@chho.imagico.de>
Gilles Tran wrote:
> 
> Not with radiosity (and perhaps photons, I didn't test).
> Here is memory use for the scene below, with and without radiosity:
> 320 x 240 : 13 Mb (150 kb)
> 640 x 480 : 30 Mb (190 kb)
> 1600 x 1200 : 58 Mb (324 kb)
> 6400 x 4800 : 105 Mb (10 Mb)
> 
> Fortunately, the increase isn't proportional, but it's still important, and
> this scene is minimal.
> 
> global_settings{
>     radiosity{error_bound 0.02 count 300 recursion_limit 1}
> }

Well - with that settings for sure. But that is not a question of render 
size but of radiosity samples taken and stored.  If your settings lead 
to a lot of samples being taken during final trace the memory use will 
raise during the render.

Memory use by photons do not increase during render.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 06 Jul. 2004 _____./\/^>_*_<^\/\.______

Copyright 2003-2008 Persistence of Vision Raytracer Pty. Ltd.