Re: [Openexr-devel] Y RY BY images

[Florian Kainz]

If you are using class RgbaInputFile to read the file, then the
RY and BY channel names are special; RgbaInputFile assumes that
they are sub-sampled with x and y sampling rates of 2.  When you
use class RgbaOutputFile to write a Y/RY/BY image, then the RY
and BY channels will automatically be sub-sampled.

A Y/RY/BY image with without sub-sampling can be read via the
general (arbitrary set of channels) interface, but RgbaInputFile
will reject it.  Maybe this should be fixed; the changes in class
RgbaInputFile would be manageable.  On the other hand, why would
you want to write Y/RY/BY images without sub-sampling?  You might
as well write RGB files.

Florian

Jim Hourihan wrote:
> Hi all,
>
> Is there a trick to writing Y RY BY images? I've tried a number of  
> things and I cannot for the life of me understand what's going on.  I've 
> been able to reduce the problem to the following:
>
> I can successfully write out a scanline piz 3 channel R G B image  with 
> no sub-sampling. By declaring the channels as Y RY BY instead of  R G B 
> (all other things being the same), the Imf library will create  a file 
> that exrheader indicates should be kosher, but exrdisplay will  not 
> read. exrdisplay produces this error:
>
> "X and/or y subsampling factors of "BY" channel of input file  "out.exr" 
> are not compatible with the frame buffer's subsampling  factors."
>
> exheader says this:
>
> file format version: 2, flags 0x0
> channels (type chlist):
>     BY, 16-bit floating-point, sampling 1 1
>     RY, 16-bit floating-point, sampling 1 1
>     Y, 16-bit floating-point, sampling 1 1
> compression (type compression): piz
> dataWindow (type box2i): (0 0) - (609 405)
> displayWindow (type box2i): (0 0) - (609 405)
> lineOrder (type lineOrder): increasing y
> pixelAspectRatio (type float): 1
> screenWindowCenter (type v2f): (0 0)
> screenWindowWidth (type float): 1
>
> This is the case with no sub-sampling indicated (sub-sampling 1) for  
> both the Channel struct and the Slice. Somehow the Imf library is  
> treating the image differently because of the channel names. I've  been 
> using the example image Rec709.exr to generate the input. (I  would like 
> to match my output to Rec709_YC.exr eventually).
>
> I tried to update my CVS tree, but it looks like savannah is down. My  
> last checkout was a couple of weeks ago. (using 1.5)
>
>     -Jim
>
>
> _______________________________________________
> Openexr-devel mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/openexr-devel