[R] ISO Eric Kort (rtiff)

[cgw at witthoft.com]

Thanks, Brian.  I can see where to mod readTiff to return the original data ranges; and where to mod writeTiff so it writes files with something better than the current 0:255 resolution range.

I have found an additional problem with readTiff, so is there someone I can write to about it?  
What I found was, for some tiff images created within my company, readTiff does not convert the source data correctly.  The files contain greyscale data, 16 bits per pixel (i.e. 0 to 4095).  Whether it's a mistake in the way the tags were originally written to the file, or a mistake in the way that readTiff interprets the libtiff outputs, I don't know, but readTiff only reads the upper byte of each pixel.  This produces data with a range of 0 to 15 (prior to being autoscaled into pixmap's [0,1] space ).   I can dig up the values returned by tools like tiffdump, so if someone out there in R-help land can point me to the pertinent values, I'll do all I can to help solve this problem.


Thanks again

Carl

May 4, 2010 01:46:05 AM, ripley at stats.ox.ac.uk wrote:

===========================================

On Mon, 3 May 2010, Carl Witthoft wrote:

> I wanted to ask Eric a question or two about the rtiff package, but his 
> listed email address bounces w/ 550 error.   Does anyone know how to reach 
> him, or whether he's actively maintaining rtiff?

He is not.  The latest version of rtiff was done by the CRAN team to 
fix a number of errors and keep it building on the CRAN platfoms -- 
you will see it was packaged by me.

> If anyone's interested, my primary desire is for rtiff (or other tool) to 
> provide me with the raw range of pixel values in a tiff file.  rtiff dumps 
> straight into a pixmap object, so the data are autoscaled into [0,1] .

It is very simple package, easy for you to modify -- this could be 
done in readTiff in a couple of minutes.

An alternative is BioC package EBImage (provided your ImageMagick 
installation supports TIFF).

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595