PWP and 32 bit floating point numbers

[Bernard]

Merging several bracketed photos into a HDR image produces 32 bit floating point numbers representing pixel values, for example in a '.EXR' file.
It would be interesting to continue editing of these images in 32 bit mode, before converting them to LDR 8 or 16 bits.

Are there any plans to adapt some PWP transforms to 32 bit floating point mode ?

Bernard

[jsachs]

In my opinion, 32-bit floating point is extreme overkill for working with photographic images. 16 bits per channel is more than enough to represent any scene - especially when you take into account that no monitor or printer can display more than 8 bits per channel.

[afx]

... no monitor or printer can display more than 8 bits per channel.

Hmm,
check out high end Eizos, NECs and the latest offerings from Dell and LG.
Attached to a Windows system via DisplayPort with the right cards (FirePro for example), they have a 10 bit data path.

cheers
afx

[jsachs]

10 is still a lot less than 16 -- mostly the extra bits are useful for loading lookup tables since the eye can barely distinguish 200 different gray levels under ideal conditions.

[Bernard]

Jonathan,
I don't disagree with you, however bracketing and HDR is sometimes the only way to represent a scene with extreme contrast.
The result may be a '.EXR' file in 32 bit floating point, therefore my question ...
Bernard

[jsachs]

Understood.

We might consider for the future versions reading EXR files and converting them to 16-bit.

Most HDR programs can output 16-bit files as well which can represent an extra 8 stops of dynamic range over 8-bit files.

[Bernard]

That would be nice !
Thanks,
Bernard