Am 28.02.2016 um 12:24 schrieb William F Pokorny:
> On 02/28/2016 05:04 AM, clipka wrote:
>>
>> You might want to test the newest version though :D
>>
> 
> Cool! You got the bug - thanks.
> 
> I am seeing some shift in optics.pov result 3.7.0 stable to POV-Ray
> 3.7.1-alpha.8499454. I like the new, less intense, result over the old,
> but is this 3.7.0 to 3.7.1 change expected ?

There are various factors that could contribute.

For one thing, the old encoding/decoding scheme for the photon colour
data was flawed, leading to a systematic error in the photon brightness,
which has been fixed in 3.7.1. However, that doesn't seem to fit the
bill, as this error should have manifested as a slight bias towards
black, so that in 3.7.1 we should see an _increase_ rather than a
decrease in brightness.

Another thing is that the error in the photons code I've just fixed has
been around for quite a while, and _did_ already exist in 3.7.0; it just
didn't surface until the change in the encoding/decoding scheme
translated it into artifacts Unix systems. It _may_ be that the old
encoding/decoding scheme did already translate the error into artifacts,
and that they just were a lot less bright.

My current favourite theory, however, is that the change in brightness
is related to a change in the media computation. There, too, a
long-standing flaw led to another systematic error, which has also been
fixed in 3.7.1.

Note that this can explain the differences in abyss.pov, too.

Post a reply to this message