|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Thorsten Froehlich wrote:
> chaps wrote:
>> Hello,
>> I like your testure, even if it must be quite slow to render.
>>
>> I guess the problem comes from max_trace level. In your function:
>>
>>> function {(f_sphere(x/4.2+2, y/2.8-2, z/1.6-12, 3.9)*f_bozo(x*6.2,
>>> y*6.4-32, z*.18+16))+
>>> f_granite(x/15-13, y/25-13, z/700-74)}//+f_granite(x/30, y/30, z/30) }
>>
>> bozo and granite add real bumps on the underlying sphere, so the ray that
>> are shot near the edge of the cloud may have to go through more than 5
>> surfaces before hiting the sky sphere and you get a black dot. if you add
>> the statement:
>>
>> global_settings {
>> max_trace_level 15
>> }
>>
>> it should disapear. Note that 15 is an example, and it may adjust to yor
>> scene and image resolution.
>
> This advice is plain wrong. Isosurfaces themselves are not controlled by
> max_trace_level, but by the trace_level/all_intersections statements.
> Increasing max_trace_level has no effect whatsoever here.
>
Did you read *all* that he wrote? And in particular the first
paragraph? The max_trace_level *can* have an effect here! Not on the
evaluation of the isosurface itself, sure, but in the number of
times a ray that's nearly tangent to the surface can keep crossing
it before giving a black point. See the following ASCII art:
__ _
/ \ /\ / \
ray:========x====x==x==x=========X---x------>
_ / \/ \ /\ / \_
\_/ \_/ \/
With the default max_trace_level 5, the ray would go only as far as
the '=' signs before giving a black point (well, if you want to
nitpick, it's a point of the background color whose default is
black). Increasing it can allow the ray to go all the way through
the bumps.
Jerome
- --
+------------------------- Jerome M. BERGER ---------------------+
| mailto:jeb### [at] freefr | ICQ: 238062172 |
| http://jeberger.free.fr/ | Jabber: jeb### [at] jabberfr |
+---------------------------------+------------------------------+
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
iD8DBQFF4EZOd0kWM4JG3k8RAtCTAJ4ntAWnZp6lw/FnMoZYYzN1h9bvJgCdFd3r
LP6BsKUklNJXamHiU/4qBGk=
=0HMl
-----END PGP SIGNATURE-----
Post a reply to this message
|
|