I did not say that I was oftenly going below 0.05 either, Steve. It
happened on several occasions, but this is very uncommon. 0.05 is the
value that I use in the vast majority of cases. Sorry if my statements
sounded absolute, I made a concious effort not to. But oftenly the
treshold is left at the default 0.2, and sampling is increased to
higher values to eliminate artefacting. Lowering the treshold can be a
solution to this problem without increasing sampling. This is
generally the first thing I do to eliminate artefacts, then depending
on the success of this approach, the type of image being rendered and
the render time budget, I work different combinations of sampling and
treshold. No matter what, though, eliminating the artefacts if the #1
priority, as the client will most likely refuse the image and the shot
will return to the farm anyway.
On 6/1/06, Steven Caron <scaron(at)omation.com> wrote:
sure... and the docs say "...in most cases...". Not always in every
single case but rather most of the time it doesn't increase the quality.
but i was commenting to bernard about never going above .05... and since
the docs say in most cases that the quality gain is negligible, just
wanted him to be aware of what the documents say about this parameter.
nothing more.
guillaume laforge wrote:
> >file:///C:/Softimage/XSI_5.0/Doc/mental_ray/manual/node76.html#INDEX184
>
> >"Values less than /0.05/ do not further increase quality in most
> cases..."
>
> If threshold is set to 0.05 then all the contrast below continue their
> flickering game :-). Very often we need smaller threshold to avoide
> this game so I think
> the "RTFM paradigm" is not always true.
>
> --
> Guillaume Laforge
> freelance TD | cg Artist
>
>
>
> On 6/1/06, *Steven Caron* < scaron(at)omation.com
> <mailto:scaron(at)omation.com>> wrote:
>
> well i see it daily when working with other artists shots. i can
> usually
> tweak the aliasing tab back into same quality zone but less time to
> render. for about 1/3 of my shots i use imf_diff to make sure i
> haven't
> changed the output too much. my usual percent diff is below 1%. to be
> honest though contrast(threshold) changes are not the only thing i
> do. i
> mess with filters and aa settings also...
>
> i will also admit that in a many cases i tend to believe mental
> ray docs
> without question it hasn't steered me wrong yet.
>
> send me 2 images with contrast(threshold) settings < 0.05 so i can
> see?
>
> Bernard Lebel wrote:
> > Have you ever tested it, Steve?
> >
> >
> > On 6/1/06, Steven Caron < carons(at)gmail.com
> <mailto:carons(at)gmail.com>> wrote:
> >>
> file:///C:/Softimage/XSI_5.0/Doc/mental_ray/manual/node76.html#INDEX184
> >>
> >> "Values less than 0.05 do not further increase quality in most
> cases..."
> > ---
> > Unsubscribe? Mail Majordomo(at)Softimage.COM
> <mailto:Majordomo(at)Softimage.COM> with the following text in
> > body:
> > unsubscribe xsi
> >
> >
>
> --
> Steven Caron
> doer of things somewhat technical
> scaron(at)omation.com <mailto:scaron(at)omation.com>
> 949-366-4422 x601
>
>
> ---
> Unsubscribe? Mail Majordomo(at)Softimage.COM
> <mailto:Majordomo(at)Softimage.COM> with the following text in body:
> unsubscribe xsi
>
>
>
>
>
--
Steven Caron
doer of things somewhat technical
scaron(at)omation.com
949-366-4422 x601
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
