Re: [Openexr-devel] On YUV support

[Richard Annema]

Hi Billy,

Just to jump in on nomenclature, there may be a clash between that 'Z' and
the 'Z' potentially used by 3D graphics applications to indicate the
Z-Buffer.

Hopefully a more meaningful discussion between various developers can be
started at some point on exactly these issues : standardization of channel
names, definitions and formats. Mostly because I think the OpenEXR format
should not end up in a situation in which there will be a "Written-by" type
field indicating the application that wrote the OpenEXR file, with all other
applications having to look up the specifics of how said application writes
things when trying to read a file ( as suggested by somebody who shall
remain nameless ).

We'd gladly be part of such a discussion.

Kindest regards,
Richard Annema
Director of Client Relations
SplutterFish, LLC


----- Original Message ----- 
From: "Billy Biggs" <address@hidden>
To: <address@hidden>
Sent: Friday, November 28, 2003 20:52
Subject: [Openexr-devel] On YUV support


>   The website has some notes on YUV support in OpenEXR and asks for
> suggestions.  I would like to offer some.
>
>   EXR files currently store linear intensity values for RGB.  However,
> the YUV used in video systems is always based on gamma corrected RGB:
> the Y in a YUV system is not really luminance [1].  Because of this, I
> do not see any way to convert from linear intensity RGB to the U, V
> channels of a video system without clamping to [0,1], making the use of
> EXR over a 16-bit per channel integer format like PNG seem useless.
>
>   Because of this, I would recommend dropping the idea of a YUV mode,
> and the channel names U and V, until a better defined use case is found.
>
>   I do have an alternate suggestion for some more channel names in EXR
> files.  Specifying colours as CIEXYZ triples is quite useful, and I
> would recommend "X" and "Z" as channel names.  I would then recommend
> changing this text:
>
>   Y  luminance, used either alone, for gray-scale images, or in
>          combination with U and V for color images.
>
>   to this:
>
>   Y  luminance, used either alone, for gray-scale images, or in
>          combination with X and Z for color images in CIEXYZ.
>
>   Which makes much more sense and seems a lot more useful.
>
>   Thoughts?
>
>   -Billy
>
>   [1] http://www.poynton.com/PDFs/YUV_and_luminance_harmful.pdf
>
>
>
> _______________________________________________
> Openexr-devel mailing list
> address@hidden
> http://mail.nongnu.org/mailman/listinfo/openexr-devel
>