|
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <3A0AF5A6.8D03B49D@spiritone.com> , Josh English
<eng### [at] spiritonecom> wrote:
> That is very strange. Was this bottom left to upper right or bottom right to
> upper left? I still don't undserstand why the order the rays are traced matter
> so much. Any enlightenment?
Very simple: Procedural texture are based on pseudo-random number
generators. While you get the same sequence for every render, you do not
get a reverse sequence when you render from bottom to top.
You would have the same problem with any other render algorithm that splits
up the image as well. This is also one of the complicated issue when it
comes to making POV-Ray multithreaded or do network rendering of single
frames.
Thorsten
____________________________________________________
Thorsten Froehlich, Duisburg, Germany
e-mail: tho### [at] trfde
Visit POV-Ray on the web: http://mac.povray.org
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <3A0AF5A6.8D03B49D@spiritone.com> , Josh English
<eng### [at] spiritonecom> wrote:
> That is very strange. Was this bottom left to upper right or bottom right to
> upper left?
left to right for both.
____________________________________________________
Thorsten Froehlich, Duisburg, Germany
e-mail: tho### [at] trfde
Visit POV-Ray on the web: http://mac.povray.org
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <3a0b1a3f@news.povray.org>, "Thorsten Froehlich"
<tho### [at] trfde> wrote:
> Very simple: Procedural texture are based on pseudo-random number
> generators. While you get the same sequence for every render, you do not
> get a reverse sequence when you render from bottom to top.
I always thought this only affected FRAND(), and that patterns based on
Noise() would always give the same result for every point...hmm...
--
Christopher James Huff
Personal: chr### [at] maccom, http://homepage.mac.com/chrishuff/
TAG: chr### [at] tagpovrayorg, http://tag.povray.org/
<><
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Thorsten Froehlich wrote:
>
> Very simple: Procedural texture are based on pseudo-random number
> generators. While you get the same sequence for every render, you do not
> get a reverse sequence when you render from bottom to top.
>
So how come you get the same pattern at different resolutions and with/without
AA, or when rendering only a subset of the image?
E.g. if I were to render a 480-line image by doing
(+SR480 +ER480), (+SR479 +ER479), ...
would I get a result different from a top-down render?
--
Margus Ramst
Personal e-mail: mar### [at] peakeduee
TAG (Team Assistance Group) e-mail: mar### [at] tagpovrayorg
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <3A0B34F2.1E7DF5E5@peak.edu.ee> , Margus Ramst
<mar### [at] peakeduee> wrote:
> So how come you get the same pattern at different resolutions and with/without
> AA, or when rendering only a subset of the image?
You are right, my explanation is wrong. I should look into the code
first...
Thorsten
____________________________________________________
Thorsten Froehlich, Duisburg, Germany
e-mail: tho### [at] trfde
Visit POV-Ray on the web: http://mac.povray.org
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Mark Wagner <mar### [at] gtenet> wrote:
: I'm working on random-order rendering.
Does it work with antialiasing?
Does it work with interrupted renders?
If radiosity is used, will the resulting image be identical from render
to render (if nothing is changed)?
What about other features which use caches to use previously calculated
results (I think the crackle pattern uses something like this)?
How much will it take memory?
--
main(i,_){for(_?--i,main(i+2,"FhhQHFIJD|FQTITFN]zRFHhhTBFHhhTBFysdB"[i]
):_;i&&_>1;printf("%s",_-70?_&1?"[]":" ":(_=0,"\n")),_/=2);} /*- Warp -*/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Warp wrote in message <3a0be9f7@news.povray.org>...
>Mark Wagner <mar### [at] gtenet> wrote:
>: I'm working on random-order rendering.
>
> Does it work with antialiasing?
It will do antialiasing as a second rendering pass.
> Does it work with interrupted renders?
Yes.
> If radiosity is used, will the resulting image be identical from render
>to render (if nothing is changed)?
I don't know. I haven't gotten it to a workable form.
> What about other features which use caches to use previously calculated
>results (I think the crackle pattern uses something like this)?
It won't affect the results from the crackle cache, but in many cases will
slow down the rendering.
> How much will it take memory?
Hopefully, not much more than current rendering methods.
--
Mark
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Mark Wagner <mar### [at] gtenet> wrote:
:> Does it work with interrupted renders?
: Yes.
How?
:> How much will it take memory?
: Hopefully, not much more than current rendering methods.
How can this be possible if you have to keep the whole image in memory
(for example for the antialiasing pass)?
--
main(i,_){for(_?--i,main(i+2,"FhhQHFIJD|FQTITFN]zRFHhhTBFHhhTBFysdB"[i]
):_;i&&_>1;printf("%s",_-70?_&1?"[]":" ":(_=0,"\n")),_/=2);} /*- Warp -*/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 15 Nov 2000 05:19:40 -0500, Warp <war### [at] tagpovrayorg> wrote:
>:> How much will it take memory?
>
>: Hopefully, not much more than current rendering methods.
>
> How can this be possible if you have to keep the whole image in memory
>(for example for the antialiasing pass)?
POV-Ray for Windows already does that when display is turned on.
Peter Popov ICQ : 15002700
Personal e-mail : pet### [at] usanet
TAG e-mail : pet### [at] tagpovrayorg
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Warp wrote in message <3a12633c@news.povray.org>...
>Mark Wagner <mar### [at] gtenet> wrote:
>:> Does it work with interrupted renders?
>
>: Yes.
>
> How?
It stores the number of the last pixel computed in the comments section of
the image file, and uses the number of each pixel to figure out which pixel
to compute next. This isn't true random order, it will just look random.
>:> How much will it take memory?
>
>: Hopefully, not much more than current rendering methods.
>
> How can this be possible if you have to keep the whole image in memory
>(for example for the antialiasing pass)?
It won't need to. It will store the image on disk, probably as a temporary
uncompressed TGA file.
Mark
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |