[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Public WebGL] The Newly Expanded Color Space Issue
On Sep 7, 2010, at 3:49 AM, stephen white wrote:
> On 07/09/2010, at 7:57 AM, Chris Marrin wrote:
>>> Perhaps we're missing a point here... the colour of the pixels are in sRGB, while the values of the pixels are linear. Two different things.
>> I don't understand this distinction at all, and perhaps this is my problem!
> The distinction, which I'm less sure of now, is that the mathematical calculations are linear regardless of colour space. That suggested to me that there's no problem with having sRGB colours as long as there's an option for passing in values without modification.
> Since I wrote that message, Thatcher Ulrich wrote a much more thorough summary. This is beyond me now, so I'll have to step aside as the colour space experts battle this out. :)
>> So if I consider my drawing buffer to be sRGB, I would need a value greater than 128,128,128 to get a display of 50% grey. That gives you a greater range of dark colors, which I believe is Thatcher's argument. But figuring out what pixel value to use to get 50% grey requires extra computations, which I believe is Steve's argument.
>> Is there more to it than that?
> No, other than the possibility of disregarding the sRGB pixel value if it's just going to be passed through (eg, not used in mathematical calculations but only as a sampler). That would indicate the colour space should be sRGB while allowing linear data input.
If you're saying what I think you're saying, I'd rather not call it "linear data input". I think you mean that we need an option of bringing in texture data unchanged. But the source data might not be "linear". So I'd rather use the term "raw" or perhaps "unconditioned". I agree we need that option. I also believe that our "conditioned" format for incoming images needs to match the format of the canvas, whether it be sRGB or linear.
You are currently subscribed to email@example.com.
To unsubscribe, send an email to firstname.lastname@example.org with
the following command in the body of your email: