[Tig] Why isn't HDR(-like) rendering more widespread?

glenn chan glennchan
Thu Mar 30 00:33:19 BST 2006


--
Thanks to oktobor for supporting the TIG
--
I suppose I was using sloppy terminology in my original post.

What I wanted to get at was expanding video values into the linear-light
domain.  i.e. Take the values to an exponent of 2.2.

It implies the use of 32-bit floating point numbers in rendering.

While these concepts may not be HDR per se, these concepts from HDR can be
used in color correction.

Are we on the same page here?

Glenn Chan

On 3/29/06, Bob Friesenhahn <bfriesen at simple.dallas.tx.us> wrote:
>
> On Wed, 29 Mar 2006, glenn chan wrote:
> > I'm curious if there are any color correction systems on the market that
> do
> > this.  The latest version of After Effects supports this, although I
> > wouldn't consider it a serious color correction system.
> >
> > Why HDR?
> > It allows more natural-looking corrections.  Saturated colors will
> retain
>
> You are mixing many issues together and entitling them "HDR".  All HDR
> means is that the encoding supports subtantial range above white, and
> substantial range below black.  So, yes, if an intermediate algorithm
> would have caused normal-ranged values to "blow out", "saturate", or
> become zeroed, then extended coding can help prevent that.  The Cineon
> log encoding is considered to be an extended range format (let's call
> it "EDR").
>
> There is extensive use of "HDR" encodings in the rendering and
> compositing areas and has been for years.
>
> It would be wrong to assume that traditional image processing uses the
> same storage size for its computations as is used in the file format.
> For example, pixel values can be converted to 32-bit float or double
> for crunching, and then the final values can be appropriately
> translated back into the simple integer form required by the majority
> of file formats.  The end result can be the same as you are
> describing.
>
> The traditional image processing approach does not help prevent the
> user of the software from destroying the image by using an algorithm
> which promotes the values outside of the storage range.  From my point
> of view, HDR with linear-light values is necessary for rendering in
> order for the math to work, but for the colorist, the main advantage
> of HDR is to allow mistakes to be corrected later.
>
> There is a disadvantage of HDR formats in that the files must be much
> larger so file I/O becomes much slower and schedules must therefore be
> longer.  There is also the disadvantage that there are only a few well
> supported HDR formats.
>
> Bob
> ======================================
> Bob Friesenhahn
> bfriesen at simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/
> GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/
>
_______________________________________________
Tig list - http://tig.oktobor.com/mailman/listinfo/tig
TIG wiki: http://tig.colorist.org/wiki




More information about the Tig mailing list