RE: [cgmo-webcgm] Transparent Cell Color Updated

[Robert Orosz <roboro@auto-trol.com>]

Hi Stuart,

The value of the MSP parameter in the Tile element depends on the Colour
Model used in the JPEG data stream.  Typically, this is YCbCr for JPEG
compressed color images.

So, the first member of the SDR would be:

11 1 5 (RGB-related)

The second member of the SDR would then be:

11 1 0 (YCbCr)

I haven't looked at your JPEG data stream in great detail, but it appears to
have APPn markers in it.  The viewers that you are using are probably
looking at those to determine the Colour Model.  Or else, they may just be
assuming that the Colour Model is YCbCr without any checking of the data
stream.

In any event it is not good practice to rely on any APPn markers in the data
to determine the Colour Model because the ISO Register does not require them
to be present.

Regards,

Rob

-----Original Message-----
From: Galt, Stuart A [mailto:stuart.a.galt@boeing.com]
Sent: Tuesday, January 27, 2009 1:03 PM
To: WebCGM
Subject: [cgmo-webcgm] Transparent Cell Color Updated


Hello,  

I have renamed the test to transCellColor to be more consistant with the
our naming convention.  I also introduced a .htm file that will display
both the CGM and the reference PNG.

Jan 27 2009 19:05       100886 transCellColor.cgm
Jan 27 2009 18:54         1088 transCellColor.htm
Jan 27 2009 18:54        49524 transCellColor.png

Changes to the CGM image itself:
- Fixed metafile desc to call out WebCGM instead of ATA profile
- Added a MAXIMUM VDC EXTENTS element in the descriptor section
- Removed the ATA escape 4000 elements.

When I run it through metachek I get an error for the TILE element.

Error 6688: Element Class/ID: 4/29   Offset: 586 octets Element No. 55
TILE:  Baseline JPEG compression requires method specific parameters,
encoded as an SDR, stating the Colour Model.  The Colour Model must be
0-5. When the Colour Model is 5, the Colour Submodel must be 0-5 and
also encoded in the method specific parameters.  4.21

I believe that the SDR in question is: "11 1 1 11 1 0"  
When I look at 6.18 
http://docs.oasis-open.org/webcgm/v2.1/cd01/WebCGM21-Profile.html#webcgm
_4_21 
(JPEG compression -type 7) it requires one or two parameters.  I have 
also tried it with "11 1 1" but get the same error.  I looks like if I
use
color model RGB it ignores the second parameter.  Either way I am not
sure I am making the error that metachek calls out.  It appears to
display
the same in different viewers.

Do I need to update the test table?
Also if you know the error that I am making please let me know.


--
Stuart Galt
SGML Resource Group
stuart.a.galt@boeing.com
(206) 544-3656


---------------------------------------------------------------------
To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  Follow this link to all your TCs in OASIS at:
https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php