[prev in list] [next in list] [prev in thread] [next in thread] 

List:       freedesktop-openicc
Subject:    Re: [Openicc] ICC meta Tag for Monitor Profiles 1.0 - DRAFT 5
From:       Kai-Uwe Behrmann <ku.b () gmx ! de>
Date:       2012-01-16 9:32:31
Message-ID: alpine.LNX.2.00.1201161020220.8294 () siras ! rasena
[Download RAW message or body]

Am 15.01.12, 19:39 -0000 schrieb Richard Hughes:
> On 15 January 2012 15:02, Kai-Uwe Behrmann <ku.b@gmx.de> wrote:
>>>> * EDID_red_x:
>>>> * EDID_red_y:
>>> What floating point precision? For instance, if GCM writes "0.8789123"
>>> and Oyranos writes "0.87891229" would that match?
>> Applications should write to the EDID delivered precission. A link to the
>> EDID spec is in the DRAFT. "2.2" and "2.2000000" are numerically the same.
>
> Okay, but there is no numerical precision specified for the primaries.
> A client has to do bit manipulation, followed by repeat multiplication
> of a power of two to another value. It's perfectly conceivable that
> the same code on i386 and ppc64 is going to produce slightly different
> results in this case. Comparing floating points for equality like this
> is not a good idea at all. The spec should really have some kind of
> guideline for a good enough match, i.e. a delta of =< 0.0001

Thats a implementation detail. Oyranos uses 0.0005 and matches 
successful.

I can put that value in the discussion section if that helps you.

regards
Kai-Uwe Behrmann
-- 
developing for colour management 
www.behrmann.name + www.oyranos.org
_______________________________________________
openicc mailing list
openicc@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/openicc
[prev in list] [next in list] [prev in thread] [next in thread]