POV-Ray : Newsgroups : povray.bugreports : one source, two images Server Time
23 Feb 2024 12:52:40 EST (-0500)
  one source, two images (Message 1 to 9 of 9)  
From: jr
Subject: one source, two images
Date: 18 Nov 2023 11:55:00
Message: <web.6558eb62bdeec745f11225116cde94f1@news.povray.org>
hi,

when rendering WFP's "op art" scene[*] from 2014, I get different images,
depending whether I use a "stock" POV-Ray (3.8, both alpha and beta.2), or WFP's
own (6e4ed6c2), see attached.

[*] <https://news.povray.org/53b35d39%40news.povray.org>

regards, jr.


Post a reply to this message


Attachments:
Download '140701_same_source.png' (2780 KB)

Preview of image '140701_same_source.png'
140701_same_source.png


 

From: jr
Subject: Re: one source, two images
Date: 19 Nov 2023 02:00:00
Message: <web.6559b2551cb2cd1df11225116cde94f1@news.povray.org>
"jr" <cre### [at] gmailcom> wrote:
> when rendering WFP's "op art" scene[*] from 2014, I get different images,
> depending whether I use a "stock" POV-Ray (3.8, both alpha and beta.2), or WFP's
> own (6e4ed6c2), see attached.

version 3.7.0.8 works as intended, from 3.7.1-rc.1 the results/images differ.


Post a reply to this message

From: William F Pokorny
Subject: Re: one source, two images
Date: 19 Nov 2023 02:15:46
Message: <6559b622$1@news.povray.org>
On 11/18/23 11:50, jr wrote:
> hi,
> 
> when rendering WFP's "op art" scene[*] from 2014, I get different images,
> depending whether I use a "stock" POV-Ray (3.8, both alpha and beta.2), or WFP's
> own (6e4ed6c2), see attached.
> 
> [*] <https://news.povray.org/53b35d39%40news.povray.org>
> 
> regards, jr.

Hi.

v3.7 renders the bottom version too.

My vague recollection is that some time during v3.71->v3.8 development a 
typo was introduced somewhere in the class set up for the spiral 1 & 2 
patterns. We ended up with two spiral1(s) or two spiral2(s) instead of 
one of each - or something similar to that. Likely in pattern.h or 
pattern.cpp or both.

I thought I'd mentioned it somewhere out here in the news groups, but 
looks like I did not add the fix to any of my commit messages - so maybe 
not.  I was just something I saw while working on other issues.

I'm fighting a northern hemisphere cold/flu/bug at the moment so I don't 
feel like doing any more rooting around.

Bill P.


Post a reply to this message

From: jr
Subject: Re: one source, two images
Date: 19 Nov 2023 06:45:00
Message: <web.6559f4c81cb2cd1df11225116cde94f1@news.povray.org>
hi,

William F Pokorny <ano### [at] anonymousorg> wrote:
> ...
> I'm fighting a northern hemisphere cold/flu/bug at the moment so I don't
> feel like doing any more rooting around.

perhaps I ought to have tried to render a "get well soon" card :-), best wishes.


regards, jr.


Post a reply to this message

From: Cousin Ricky
Subject: Re: one source, two images
Date: 19 Nov 2023 19:59:24
Message: <655aaf6c@news.povray.org>
On 2023-11-19 03:15 (-4), William F Pokorny wrote:
> 
> My vague recollection is that some time during v3.71->v3.8 development a
> typo was introduced somewhere in the class set up for the spiral 1 & 2
> patterns. We ended up with two spiral1(s) or two spiral2(s) instead of
> one of each - or something similar to that. Likely in pattern.h or
> pattern.cpp or both.

This is a montage of renders of the scene with the POV-Ray versions that
I have, plus 3 versions of UberPOV along the bottom row.  From these, it
appears that the bug was introduced prior to the release candidate phase
of 3.7.1, but after whichever version of 3.7.1 that the latest UberPOV
was based on.


Post a reply to this message


Attachments:
Download 'wfp-op_art.montage.jpg' (1465 KB)

Preview of image 'wfp-op_art.montage.jpg'
wfp-op_art.montage.jpg


 

From: Kenneth
Subject: Re: one source, two images
Date: 21 Nov 2023 14:10:00
Message: <web.655cff811cb2cd1d9b4924336e066e29@news.povray.org>
Cousin Ricky <ric### [at] yahoocom> wrote:
>
> This is a montage of renders of the scene with the POV-Ray versions that
> I have, plus 3 versions of UberPOV along the bottom row.  From these, it
> appears that the bug was introduced prior to the release candidate phase
> of 3.7.1, but after whichever version of 3.7.1 that the latest UberPOV
> was based on.

I like the black one best (hee hee). That one must be from POV-ray
version -001...

William, I hope you feel better soon! Having the flu is no fun.


Post a reply to this message

From: William F Pokorny
Subject: Re: one source, two images
Date: 1 Feb 2024 07:28:14
Message: <65bb8e5e$1@news.povray.org>
On 11/19/23 02:15, William F Pokorny wrote:
> My vague recollection is that some time during v3.71->v3.8 development a 
> typo was introduced somewhere in the class set up for the spiral 1 & 2 
> patterns. We ended up with two spiral1(s) or two spiral2(s) instead of 
> one of each - or something similar to that. Likely in pattern.h or 
> pattern.cpp or both.

In looking at a question elsewhere I realized there is more going on 
with these spiral 1 & 2 changes that I realized when I picked up an 
unexpected difference in spiral1 spin direction in yuqk. I'll need to 
dig for a while.

Not got much coffee in me yet, but I thought my remembered change/fix 
would be obvious. Of course, in a quick look just now I don't see it at 
all!

The v3.7.0.10 result on the left, my current yuqk working code on the 
right of the attached image.

Bill P.


Post a reply to this message


Attachments:
Download 'spiral1puzzle.jpg' (185 KB)

Preview of image 'spiral1puzzle.jpg'
spiral1puzzle.jpg


 

From: William F Pokorny
Subject: Re: one source, two images
Date: 1 Feb 2024 17:57:19
Message: <65bc21cf$1@news.povray.org>
On 2/1/24 07:28, William F Pokorny wrote:
>> My vague recollection is that some time during v3.71->v3.8 development 
>> a typo was introduced somewhere in the class set up for the spiral 1 & 
>> 2 patterns. We ended up with two spiral1(s) or two spiral2(s) instead 
>> of one of each - or something similar to that. Likely in pattern.h or 
>> pattern.cpp or both.
> 
> In looking at a question elsewhere I realized there is more going on 
> with these spiral 1 & 2 changes that I realized when I picked up an 
> unexpected difference in spiral1 spin direction in yuqk. I'll need to 
> dig for a while.

OK. I did not remember correctly - not that much of a surprise these days.

The typo in the v3.8 code is in the parser not in the class definitions 
for the spirals. Further, the typo is in a section which pertains to the 
use of the spiral1 pattern in a function.

In the file parser_materials.cpp there are two blocks of code looking like:


...
         CASE (SPIRAL1_TOKEN)
             New->Type = GENERIC_PATTERN;
             New->pattern = ParseSpiralPattern<Spiral1Pattern>();
         END_CASE
...



...
         CASE (SPIRAL1_TOKEN)
             New->Type = GENERIC_PATTERN;
             New->pattern = ParseSpiralPattern<Spiral2Pattern>();
         END_CASE
...

That second one is in a section which handles function{ pattern {} } 
wrap differences (though the spiral has none) and it should of course read:

...
         CASE (SPIRAL1_TOKEN)
             New->Type = GENERIC_PATTERN;
             New->pattern = ParseSpiralPattern<Spiral1Pattern>();
         END_CASE
...


On the rotational difference between v3.8 and yuqk
--------------------------------------------------

It is intentional!

I forgot that the spiral patterns were two of those touched by the fix 
to POV-Ray's internal Turbulence() function - for the distribution drift 
on the omega values.

While I was doing the work to adjust to that turbulence change and to 
move to the internal turbulence keywords (it_* (a)), I noticed the 
official POV-Ray rotation for the spiral pattern was right handed about 
z as the distance from the origin increased rather than the more normal 
- and likely expected - left handed rotation(b).

(a) - The spiral's turbulence in official POV-Ray has long been 
colliding with the pattern modifier's 'frequency' keyword usage.

(b) - With yuqk, I have, in a few other places too, moved to change 
right handed implementations of stuff to left handed ones.

Remembering all this, I recall there is somewhere a detailed post on 
everything going on that was fixed / changed in yuqk with the spiral 
patterns.

And yep! That I cannot remember my own work is a little disturbing.

Bill P.


Post a reply to this message

From: Cousin Ricky
Subject: Re: one source, two images
Date: 1 Feb 2024 18:46:51
Message: <65bc2d6b$1@news.povray.org>
On 2024-02-01 18:57 (-4), William F Pokorny wrote:
> 
> And yep! That I cannot remember my own work is a little disturbing.

Based on my own experience, and hearing from other coders, you are
perfectly normal.

IIRC, this observation is noted in the book _The Elements of Programming
Style_ from back in the 1970s.  They say to write your code for someone
else, because in the future, you *will* be someone else.


Post a reply to this message

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