2008.02.07 15:27 "[Tiff] 32-bit floating point RGB images (and gamma)", by Edward Lam

2008.02.12 03:28 "Re: [Tiff] 32-bit floating point RGB images (and gamma)", by Edward Lam

Now, Photoshop DOES correct for the display -- your gamma 1.0 image will look correct on the display because Photoshop converts the data to the display gamma (and chromaticities) when displaying the image.

Indeed, this fact had escaped me. I was lead astray when I sought for reasons why the might be interpreted differently and found the TransferFunction tag.



On 2/7/08 7:27 AM, "Edward Lam" <edward@sidefx.com> wrote:

What's the best way to express a TIFF file that has 32-bit floating point, *linear* RGB data? If I do the dumb thing, then Photoshop CS3 Extended interprets the data with a 2.2 gamma. However, ImageMagick seems to display the image without the gamma correction.

Looking at the TIFF 6 spec, the default TransferFunction seems to be 2.2 gamma correction if unspecified for RGB PhotometricInterpretation? However, one can't easily provide a linear TransferFunction for 32 bits per sample data.