Hi Renaud,
QImage is limited to a maximum of 8 bits per channel, with a very limited list
of pixelformats.. as such you will probably have to do all your operations
(especially OCIO ops) in full 32 bit floating point, and finally resample the
image into an 8 bit per channel RGB/RGBA format with oiio, which you can then
wrap in a QImage.
Assuming you end up with an 8bpc RGB ImageBuf already, you should be able to
directly construct a QImage from this, using
myImageBuf.pixelAddr(xmin,ymin,zmin) as the pointer to the QImages raw data in
the constructor e.g.
QImage myImage( myImageBuf.pixeladdr( myImageBuf.xmin(), myImageBuf.ymin(),
myImageBuf.zmin() ), myImageBuf.width(), myImageBuf.height(), QImage ::
Format_RGB888 );
or similar. Note some formats (particularly exr where dataWindow does not match
displayWindow) will have their pixel data starting at some other location than
0,0 which is why I have used xmin()/ymin()/zmin() but in most cases
pixeladdr(0,0,0) would work fine.
In the case your 'final' processed ImageBuf is something with > 8 bits per
channel, you can also use ImageBuf.get_pixels() to dump the pixels from a float
ImageBuf into a specified format - this is probably a nicer way to do it, and
in this case you would probably allocate a buffer of the correct size using
malloc, and the width/height/nchans info from the ImageBuf - and call
myImageBuf.get_pixels() to copy the image data into it and then wrap a QImage
around it.
e.g.
unsigned char* myBuffer =
malloc(myImageBuf.width()*myImageBuf.height()*myImageBuf.nchans());
myImageBuf.get_pixels(myImageBuf.roi_full(),TypeDesc::UINT8,myBuffer);
QImage::Format QIFormat = QImage::Format_RGB888 //default case is 3 channels
if (myImageBuf.nchans == 4)
{
QIFormat = QImage::Format_ARGB32;
}
QImage myQImage(myBuffer,myImageBuf.width(),myImageBuf.height(),QIFormat);
Note if your ImageBuf has more than 4 channels of data, you will likely need to
extract the R,G,B,A channels to a new ImageBuf, and wrap that in a QImage using
one of the above techniques, as QImage cannot 'wrap' anything other than 3/4
channel RGB image data (unless you need 1 bit masks or 8-bit indexed colour,
which is unlikely).
All this code is straight off the top of my head, and may not compile as-is -
but these are the approaches I would use.
Hopefully that helps?
-Pete
PETER BLACK l PARK ROAD POST PRODUCTION LTD
TELEPHONE +64 4 909 7800 l WWW.PARKROAD.CO.NZ
----- Original Message -----
From: "Renaud" <[email protected]>
To: [email protected]
Sent: Wednesday, February 3, 2016 7:15:39 PM
Subject: [Oiio-dev] ImgBuff to Qimage
Hi !
I've been trying to load an image using the ImgBuff approach for a while now
but I can't quite figure out how to dump the ImgBuff back into a QImage.
Would anyone be kind enough to provide a simple example on how to do that ?
I was able to load images into a QImage using the following method :
http://permalink.gmane.org/gmane.comp.lib.openimageio.devel/140
but I would like to be able to process the image using OCIO: LUT, colorpsace
changes, ... before dumping the result into a QImage.
Any help would be greatly appreciated.
Thanks.
_______________________________________________
Oiio-dev mailing list
[email protected]
http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
_______________________________________________
Oiio-dev mailing list
[email protected]
http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
