If that's wrong, then I haven't understood anything.
Christian Rittener
> -----Original Message-----
> From: owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] On
Behalf
> Of Luc-Eric Rousseau
> Sent: July 11, 2007 14:45
> To: XSI(at)Softimage.COM
> Subject: RE: Slider values and bit depth
>
> Parameters like the color in the render tree are not anything
absolute,
> they are simply weights used in a formula, which is already operating
in
> linear space and how you use it doesn't change depending on the bit
> depth. They're mathematical factors. Color is not absolute but
relative
> to what's already there and how it will be mapped to display.
>
> For Photoshop where you're not only transforming pixels relatively but
> setting new values, the user is picking color based on what it will
look
> like for printing and display, and therefore there is no change with
the
> bit depth storage.
>
> For an app where you would paint the 'raw' color value in a floating
> point buffer, no one can tell you what is the 'possible range', it
> depends how you will use the data to a device later. Linear floating
> point space only means that for a given color value, a value twice its
> amount is twice as bright. There is no maximum or minimum, it only
> defines how values are to be interpreted relatively to each other.
>
> OpenEXR anchors by default the color of a photographed 18% gray card
to
> the floating-point value 0.18, and you'd be painting values relative
to
> that, hopefully with a look up table so you can tell what it'll look
> like, unless you're only doing relative changes. There is no white
> point at this stage because white points are defined as the maximum
> limit of the display medium, it's arbitrary.
>
> HLS/HSV ranges never changes. It defines a convenient cone of color
for
> humans on 8-bit interface display, in monitor gamma, with 0.0 being
> black and 1.0 being the white point of the monitor. It predates any
> concept of color management. It's possible in some apps to poke HSV
> values outside of 0.0 and 1.0 range, but these values have no absolute
> meaning without the rest of the color management pipeline, HLS/HSV is
> not defined when out-of-range.
>
> The raw linear floating point data needs to be exposed and mapped to a
> display before it has meaning, like taking a picture. When you expose
> it, the raw floating point data is in fact used as a mathematical
factor
> in a formula, again.
>
> If you don't use any look-up-table to view floating point data, then
you
> are not in fact working in linear space and all the color controls
work
> the same was a they do in 8-bit, with 1.0 meaning the white point of
the
> monitor.
>
>
>
> > From: Christian Rittener
> >
> > Damn. Nobody cares. Or nobody knows.
> >
> >
> > > -----Original Message-----
> > > From: owner-xsi(at)Softimage.COM [mailto:owner-xsi(at)Softimage.COM] On
> > Behalf
> > > Of Christian Rittener
> > > Sent: July 9, 2007 11:08
> > > To: XSI list
> > > Subject: Slider values and bit depth
> > >
> > > The other thread about bit depth gives me the opportunity to
finally
> ask
> > > this: when working with 32-bit linear float images, what's the
> > > possible range of normalized-for-8bits values for RGB sliders in
> > > your standard graphics app (Photoshop, Fusion, XSI, ...) ? what
> about HLS values?
> > > Does
> > > this range apply to lights in XSI as well?
> > >
> > > Christian Rittener
>
> ---
> 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
