Contents || Index
02/05/2008

 

File Utilities Tool

This tool has miscillaneous functions for examining various types of files.  Use the <control> key to get a description of each function.

 


Image load control parameters

Choice of sort methods:


Image write control parameter for saving floating point images.


" This affects the saving of double precision floating point images only.

4 byte floats = single precision floating point = ieee-single
8 byte floats = double precision floating point = ieee-double.

If set to ieee-single, and
IF ALL OF THE PIXELS ARE WITHIN THE RANGE OF 4 BYTE FLOATS,
will change the 8 byte float image to a 4 byte float image, and then
save the image using 4 byte floating point numbers.
This saves space and is analogous to the way Lispix saves integer images, except
that with 8-bit floats, 4 bytes of precision are lost. (No data is lost in the case of integer images.)
If one or more of the pixels are outside the range of 4 byte floats,
will then Lispix will write the image as-is (as an 8 byte float image).
If set to ieee-double, 8 byte floats,
Lispix will leave 8 byte floating images as-is, saving all of the bytes.

Set to ieee-double if you wish to keep 8 bytes of precision."


For floats, if a double float image has pixels confined to the range of about +- 10^38, then it will be changed to a single float image and sagved that way, IF this parameter is set to ieee-single  This is NOT the default setting for Lispix.  If the parameter is set to ieee-double, the default in Lispix, then double precision floating point images are saved as-is, and not reduced to single precision floating point images.

Before writing an image array, Lispix copies the pixel data to an array with the smallest data type for all of the pixels.  For example, if the pixels in an integer image range only from 0 to 255, then the image will be saved as an 8 bit byte image rather than as a 16 or 32 bit image.  Setting the parameter to ieee-single has Lispix reduce floating point images in an analogous way.  Single precision floating point images can't be reduced, and so stay as they are.  Double precision floating point images are reduced to single precision (3.5 bytes of precision is still a lot of precision for much image processing) if all of the pixels are within the dynamic range of single precision floating point numbers.  If one or more pixels is outside this range, double precision images remain as-is, despite the setting of this parameter to ieee-single.

 


Image File Format Buttons


Spectrum Files