[Public WebGL] gamma/color profile and sRGB

Mark Callow [email protected]
Sun Nov 23 04:57:41 PST 2014

> On Nov 23, 2014, at 4:34 PM, Florian Bösch <[email protected]> wrote:
> BTW, I hope current browsers are converting the current linear back-buffers to their compositing color space. I wonder because such conversion has considerable overhead.
> I hope that is exactly not the case, because that would be quite the wrong thing to do, without the application programmer having indicated his willingness for this to happen.

> So if you where to enforce linear space conversion upon output, but do not enforce gamma space conversion upon input, you'd force anybody wishing to output colors unmodified, to modify the colors upon reading them in.

The display in most cases is sRGB. It is not possible to display part of the screen (or composited web page) with a different transfer function from the rest. If the display is sRGB, or whatever, a linear buffer/canvas must be encoded (a.k.a gamma corrected) to the display’s transfer function, otherwise the result is completely wrong. You would be displaying the linear data as if it were encoded in sRGB.

Even in native land, access for apps to change the transfer function is limited to full screen apps. How do you propose that a browser pass through just some parts of the displayed web page? If we limit pass through to full screen canvases, how do you cope with blending other page content with or displaying it over the canvas?

I am not saying pass through is a bad idea. I don’t see how to implement it.



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://khronos.org/pipermail/public_webgl_khronos.org/attachments/20141123/8a3a3f86/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 495 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://khronos.org/pipermail/public_webgl_khronos.org/attachments/20141123/8a3a3f86/attachment.asc>

More information about the public_webgl mailing list