tif2mrc(1) General Commands Manual tif2mrc(1)
NAME
tif2mrc - Make a MRC file from TIFF image files.
SYNOPSIS
tif2mrc [options] TIFF_input_files... MRC_output_file
DESCRIPTION
tif2mrc takes a series of TIFF files and converts them into an MRC
stack. tif2mrc can also convert a single multi-page TIFF file to an
MRC stack.
Only grayscale and color image files are supported. Grayscale images
can be 8 or 16 bit integers or 32 bit floats or integers, with 32-bit
integers converted to floats in the MRC file. Color images can be 24
or more bits as long as pixels are stored as 8-bit red-green-blue
triplets followed by the extra bytes if any (e.g., in RGBA images, the
alpha is an extra byte). Color images can also be 8-bit with an embed-
ded colormap. All color images will be converted to 24 bit RGB images,
since this is the kind of color MRC file that IMOD defines. If a
grayscale image file has multiple sample values per pixel, then each
sample will be converted to a separate section in the output file, and
the input file will be treated like a multi-page TIFF file. The multi-
ple samples can be stored either contiguously or in separate planes.
Files with data stored in LZW, ZIP, or JPEG compression can be read
when running with the TIFF libraries distributed with IMOD.
Images from different files need not all be the same size. The x and y
dimensions of the stack are set from the x and y size of the first
image, unless the -o option is used to set the output size. Images
that are larger than the output size will be cropped around all four
sides, and images that are smaller will be placed in the center of the
output area and surrounded with gray.
Most of the options have to do with the treatment of unsigned 16-bit
numbers. The standard MRC file mode for holding 16-bit integers is
mode 1, which holds signed (positive and negative) values and has a
range from -32768 to 32767. Unsigned integers, with a range from 0 to
65535, can be stored in file mode 6, which is a nonstandard mode. If
the TIFF file has a tag indicating that it contains unsigned integers,
then by default the program will preserve these values in a file of
mode 6. You can alter this behavior and store the numbers as signed
values (mode 1) in three ways: 1) with the -u option to subtract 32768,
which will preserve the dynamic range of the values; 2) with the -d
option to divide by 2, which will throw away one bit of precision but
keep the numbers positive; or 3) with the -s option to force the num-
bers to be saved unmodified in mode 1, which will be fine as long as
there are no values above 32767. Finally, if the TIFF file lacks a tag
indicating that it is unsigned but the numbers do range above 32767,
you can preserve unsigned integers and force output into a file of mode
6 with the -k option.
Since IMOD is distributed with version 4 of the TIFF library, the pro-
gram will read TIFF files bigger than 2 GB, including ones in the
BigTIFF format.
OPTIONS
-g This option will convert 24-bit color images to 8-bit grayscale
in the MRC stack by equal weighting of the three color compo-
nents. This weighting has always been used with this option to
Tif2mrc, and is also the way that 3dmod read in color data with
its -G option prior to IMOD 3.12.10.
-G This option will convert 24-bit color images to 8-bit grayscale
in the MRC stack by the NTSC weighting of the three color compo-
nents, which is used when reading color data as grayscale into
3dmod (as of IMOD 3.12.10), and when changing the mode of images
from color to grayscale with Clip.
-u With this option, the program will convert 16-bit numbers from
unsigned to signed integers by subtracting 32768. Use this
option if the TIFF files have positive numbers ranging above
32767, you want an output file of mode 1, and you do not want to
lose any of the dynamic range of the numbers.
-d With this option, the program will treat 16-bit numbers as
unsigned integers and divide them by two. The resulting numbers
will range between 0 and 32767. Use this option if the TIFF
files have positive numbers ranging above 32767, you want an
output file of mode 1, you want the numbers to remain positive,
and you do not mind losing one bit of precision.
-s With this option, the program will treat numbers as signed inte-
gers and just store them in a file of mode 1. Use this option
if an unsigned integer file contains values under 32768 and you
want a file of mode 1. The program will print the minimum and
maximum values in the output file. If the min is near -32768
and the max is near 32767, this option was chosen incorrectly.
-k With this option, the program will treat numbers as unsigned and
output data into a file of mode 6 even if there is no tag indi-
cating the data are unsigned.
-B # This entry controls how bytes are stored in the output file and
overrides both the default behavior of writing signed bytes and
the setting of the environment variable WRITE_MODE0_SIGNED.
Enter 0 for unsigned numbers or 1 for signed numbers. Unsigned
output may be needed for compatibility with external software
that has not accommodated to the new MRC standard. Regardless
of the representation in the file, bytes are read into IMOD pro-
grams as unsigned with a range of 0 to 255
-i Invert the order of sections in the output stack. Either the
multiple files or the pages of a a multi-page stack will be pro-
cessed in inverted order to produce the output. This option
will preserve preserve handedness if the images being stacked
are serial images obtained while cutting into a block face and
are in the same orientation as views looking down on the block.
-f If only one input file is given and it is a multi-page TIFF
file, the program will convert all of the pages in the file into
sections in an MRC file. Use this option to make the program
ignore the multiple pages and read only the first image. This
option is no longer needed to process files with thumbnail
images embedded; if there are pages with differnt sizes, only
the largest size will be converted to sections, and thus a
thumbnail image will be ignored.
-o #,# Set the size of the output to the given dimensions in X and Y.
This option has no effect for multi-page input files.
-F # Set the value to use when filling regions that have no image
data because an image is smaller than the output size. The
default is to fill with the mean value of each image. The value
must be within the valid range for the output data type (e.g.,
between 0 and 255 for byte output). For color image output, the
fill value will be used for red, green, and blue components.
-P Set the pixel spacing in the header of the output file from res-
olution values in the TIFF file, if any, regardless of the size
of these values. The default is to use a resolution value only
if the resulting pixel size is no bigger that 1 micron. This
option is one way to override that limit. To override the limit
automatically, define the environment variable
TIFF_RES_PIXEL_LIMIT to the desired limit, a floating point num-
ber in microns. Pixel size is assigned from the size found in
the first input file.
-p # Set the pixel spacing in the header of the output file to the
given value. This spacing is defined as having units of
Angstroms, but you can place arbitrary values there. This
option cannot be used with -p and the entered value supercedes
any resolution value in the TIFF file.
-b file
This option loads a background file in TIFF format. If a back-
ground file is loaded its values are subtracted from the values
in the rest of the files. The background file must have the
same data type as the images from which it is being subtracted
(8-bit or 16-bit grayscale).
-T file
When the input files are produced by TVIPS data acquisition
software (after October, 2012), this option can be used to out-
put tilt angles, one per line, to the given file. Regardless of
whether this option is used, when the program recognizes such
files, it will add a title to the header with the tilt axis
rotation angle, binning, and spot size if available.
-m Turn off file-to-memory mapping in the TIFF library. This map-
ping could be a problem working with large files on some sys-
tems. If you experience performance issues such as excessive
disk swapping, you can experiment with this setting. With map-
ping on, the virtual memory usage of the program is alarming and
the resident memory size continually increases as the file is
read, even though it is being processed in chunks. However, in
tests on Red Hat Enterprise Linux 5, reading and writing a file
had the same potential to cause other programs' memory to be
swapped to disk regardless of whether mapping was on or off,
apparently because the mapped memory is coincident with file
system cache.
-t value
Set the criterion for processing each file in chunks to the
given value in megabytes (fractional values are allowed). Whole
images will be read in, processed, and saved as a unit if they
are less than this criterion (default 100 MB). Above the crite-
rion, images will be read, processed, and saved in chunks of
about this size, with much lower memory requirements, as long as
the image matches the size of the output file and there is no
background subtraction.
FILES
If the output file already exists, it will be made a backup file by
adding ~ to its name.
AUTHORS
Jim Kremer and David Mastronarde
SEE ALSO
raw2mrc
BUGS
If a standard TIFF library is not installed, images must be uncom-
pressed and must consist of strips rather than tiles, and only a subset
of image types can be read.
Email bug reports to mast at colorado dot edu.
IMOD 4.11.0 tif2mrc(1)