[Bf-committers] F-Stop

[Ejner Fergo]

Hi again,

So the defocus node reads the necessary values needed to calculate an fstop
from the camera, but then why isn't it physically correct? I am also very
confused about the added "max blur", which is why I call this node a purely
creative node, and not a "1-to-1" fstop node. When I set the scene camera
in the defocus node (having turned fstop/aperture radius off on the camera,
keeping only distance), and have the fstop value to 128 in the node (should
be no focal blur), I get a result equalling to a physical f/2.0 in that
scale. Why is that?

Here is my test scene I use. You might need to set some things up (disable
fstop in the camera and set up comp nodes). The scene is build to scale and
camera matches my real one.

http://www.pasteall.org/blend/35139


Best regards
Ejner Fergo



On Sun, May 17, 2015 at 9:57 AM, Jeroen Bakker <j.bakker at atmind.nl> wrote:

> Hi all,
>
> just a small one. The defocus node already has a scene option where
> users can select where to read the camera from.
>
> Jeroen.
>
> On 05/17/2015 09:47 AM, Jeroen Bakker wrote:
> > Hi all,
> >
> > Well the defocus node has a setting no_zbuf. this option is used to
> > determine if it needs to preprocess the z input. When no_zbuf is set,
> > the fstop and active camera settings are not used at all. When it is
> > turned off the fstop is read from the node and the active camera
> > distance, lens and sensor are used to preprocess the input zbuffer.
> >
> > Currently when we need to set the fstop we also need the camera,
> > otherwise the missing values will be set to some defaults. (some even
> > uninitialized with a possible division by zero in the
> > ConvertDepthToRadiusOperation).
> >
> > We could add an option to select the actual scene where to get the
> > camera from or when no scene is selected the user could set the options
> > inside the node (distance, fstop, lens, sensor). When the scene is set,
> > these are read from the active camera of that specific scene. The
> > interactive part is still guaranteed. The UI could also be cleaned up a
> > bit by only showing the options that are used by the algorithm.
> >
> > Jeroen
> >
> > On 05/12/2015 12:18 PM, Ejner Fergo wrote:
> >> As I see it, the f-stop value in the defocus node has nothing to do with
> >> what f-stop is in Cycles (and in reality). This value in the defocus
> node
> >> is a purely creative tool, and as such should stay, but the name should
> >> change to "focal blur" so it is not confused with the concept and
> mechanism
> >> of f-stop, IMHO.
> >>
> >> On Tue, May 12, 2015 at 7:10 AM, Knapp <magick.crow at gmail.com> wrote:
> >>
> >>> Is it not so that you set it in cycles and then live with the results
> but
> >>> in the compositor you can change it real time and see the results? I
> would
> >>> not want to loose the interactive aspect.
> >>>
> >>> On Mon, May 11, 2015 at 11:55 PM, Sergey Sharybin <
> sergey.vfx at gmail.com>
> >>> wrote:
> >>>
> >>>> I'm not sure how you can drop F-Stop from the defocus node. it's not
> >>>> guaranteed that render scene is used for compositing, meaning
> different
> >>>> value might be needed.
> >>>>
> >>>> On Tue, May 12, 2015 at 2:27 AM, Antony Riakiotakis <kalast at gmail.com
> >
> >>>> wrote:
> >>>>
> >>>>
> >>> --
> >>> Douglas E Knapp
> >>>
> >>> Massage in Gelsenkirchen-Buer:
> >>> http://douglas.bespin.org/tcm/ztab1.htm
> >>> Please link to me and trade links with me!
> >>> _______________________________________________
> >>> Bf-committers mailing list
> >>> Bf-committers at blender.org
> >>> http://lists.blender.org/mailman/listinfo/bf-committers
> >>>
> >> _______________________________________________
> >> Bf-committers mailing list
> >> Bf-committers at blender.org
> >> http://lists.blender.org/mailman/listinfo/bf-committers
> >>
> > _______________________________________________
> > Bf-committers mailing list
> > Bf-committers at blender.org
> > http://lists.blender.org/mailman/listinfo/bf-committers
> >
>
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers
>