Re: Conversion RGB > 32bit float

Posted by Rasband, Wayne (NIH/NIMH) [E] on Sep 27, 2013; 5:31pm
URL: http://imagej.273.s1.nabble.com/Quicktime-problem-tp5004918p5004965.html

On Sep 27, 2013, at 9:10 AM, Stephan Saalfeld wrote:

> Dear List,
>
> is there a philosophical reason why I cannot directly convert from
> RGB-Color to 32bit float?  It annoys me every two months or so for the
> following two reasons:
>
> 1. pragmatic: ImageProcessor.getf(...) is defined in ColorProcessor, so
> why enforcing two extra mouse-clicks
> 2. precision: whatever mapping is done, going over RGB > 8bit > 32bit
> reduces precision from float to integer

The Image>Type>32-bit command in the ImageJ 1.48 daily build now supports conversion of RGB stacks to 32-bit float. RGB to float conversions are now done directly so they are faster, more precise and require less memory.

-wayne
--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html