So, I managed to muster enough motivation to doggedly pursue implementing the
user_defined camera approach to modeling lens aberrations with Zernike
polynomials.

At present, I'm attempting to do that using --- derivatives of the polar
equations.  (Not at all sure I'm doing it right)

Anyway, the point is, that after fiddling a bit with the location and direction
parametric equations, POV-Ray suddenly decided to start partial-rendering my
scene with some bad 8-bit blocky censorship.
Sometimes it's fairly evenly distributed across the whole image, other times
it's only in the top half....

This leads to speculate that:
1. There's something wrong with my build (3.8 beta windows executable)
2. There's potentially something wrong with my (computer) memory  :P
3. There exists some sort of state-dependent bug in the user_defined camera code

I'm going to try to work everything out in Cartesian coordinates and see if
maybe that might behave better, plus I'll be a lot more confident about the
derivatives and the resulting normal vector I'll be using for the direction {}

If that fails, I'll try re-starting POV-Ray, and perhaps shutting down until I
get back in from this evening's exciting warehouse work adventure.

- BW

Post a reply to this message

Attachments:
Download 'astigmatism.png' (189 KB)

Preview of image 'astigmatism.png'

Well, the rendered image IS called "astigmatism.png". Maybe POV-ray's new
artificial-intelligence engine is at work, digesting your code and comparing it
to millions of human vision tests that it scraped for free from private medical
records around the world. Then it responded accordingly. Check your messages
pane, to see if the following is there:  "Hey, if this image looks blocky to
you, you definitely have astigmatism! Everyone else sees a perfectly normal
image."

Post a reply to this message

On 1/2/24 14:27, Kenneth wrote:
> Well, the rendered image IS called "astigmatism.png". Maybe POV-ray's new
> artificial-intelligence engine is at work, digesting your code and comparing it
> to millions of human vision tests that it scraped for free from private medical
> records around the world. Then it responded accordingly. Check your messages
> pane, to see if the following is there:  "Hey, if this image looks blocky to
> you, you definitely have astigmatism! Everyone else sees a perfectly normal
> image."
> 
> 

:-)

On the actual issue, unsure.

Looks render block related. Does using a single thread (+wt1) make a 
difference?

I vaguely recall setting all three directional components to 0 in a user 
defined camera, renders black - or transparent if using +ua. However, I 
didn't test my recollection.

Only other thought is that some apparent POV-Ray keywords are in fact 
declared as identifiers (IDs) at start up(a). Two such identifiers are 
'image_width' and 'image_height'. Keywords that are in fact IDs can be 
whacked by users in their SDL.

Bill P.

(a) - Most such 'apparent keyword' IDs are animation related.

Post a reply to this message

William F Pokorny <ano### [at] anonymousorg> wrote:

> On the actual issue, unsure.

Well here's the thing.   the scene was rendering just fine for several renders
before I made a radical change to the functions (just adding a scalar), and then
even when i changed everything back to the original functions - the weird
behaviour persisted.

I'm not redefining any apparent keywords, however I am using the image
dimensions to adjust the aspect ratio in the user_defined block. Maybe I'll
define a right vector just to mix things up and see what happens.

In any event, the Cartesian forms are easy enough to implement, and calculating
the partial derivatives is 1000 times easier than in polar coordinates.

- BW

Post a reply to this message

Attachments:
Download 'astigmatism.png' (168 KB)

Preview of image 'astigmatism.png'