When following the manual and doing the few examples with a lathe (with the 5
coloured discs), at some point the linear_spline is replaced with the
quadratic_spline. When rendering this with the orthographic camera, a lot of
noise pixels are showing up (that is some of the curves are not curved smoothly,
but smeared out to left and right with noise streaks...).
When rendering the model, but then with standard camera (adding approprate
ligth-sources) this also gives a small strip with noise pixels over the lathe.
The
height of this strip seems to be dependend on the y-position of the camera.
Did this with version Pov 3.7RC5 and with version Pov 3.5. Same problem.
When using cubic_spline, no noise band appears.
Anyone to explain?
Peter Puzzelaar
From: Stephen
Subject: Re: a lathe with quadratic spline shows horizontal bar(s) of noise...
Date: 1 May 2012 09:24:35
Message: <4f9fe413$1@news.povray.org>
On 01/05/2012 2:06 PM, peter puzzelaar wrote:
> Anyone to explain?
An image would help Peter.
Did you use the strum keyword?
--
Regards
Stephen
Stephen <mcavoys_at@aoldotcom> wrote:
> On 01/05/2012 2:06 PM, peter puzzelaar wrote:> > Anyone to explain?>> An image would help Peter.> Did you use the strum keyword?>> --> Regards> Stephen
Adding keyword "sturm" solved the issue.
So adding an image seems redundant now...
I assume sturmian root solver (?) is needed for quadratic_spline only as this
one results in non-smooth transitions at the intermediate points. (as
cubic_spline did not need the sturmian solver)
Thanks (again... ;) ) Peter
From: Stephen
Subject: Re: a lathe with quadratic spline shows horizontal bar(s) of noise...
Date: 1 May 2012 13:15:01
Message: <4fa01a15$1@news.povray.org>
On 01/05/2012 4:36 PM, peter puzzelaar wrote:
>> Adding keyword "sturm" solved the issue.
Good.
> So adding an image seems redundant now...>
It does indeed :-)
> I assume sturmian root solver (?) is needed for quadratic_spline only as this> one results in non-smooth transitions at the intermediate points. (as> cubic_spline did not need the sturmian solver)>
I don't think so. I use it for all types but maybe I am being over
cautious. To me the noise always seems to be in line with the horizon.
Anyway with today's fast computers I never notice any difference in time
taken.
--
Regards
Stephen
