mrc2tif(1) General Commands Manual mrc2tif(1)
NAME
mrc2tif - Create TIFF, JPEG, or PNG files from an MRC image file.
SYNOPSIS
mrc2tif [options] mrc_input_file tiff_file_root_name
DESCRIPTION
mrc2tif usually creates a series of TIFF files with the prefix
tiff_file_root_name and with the suffix .nnn.tif, where nnn is the z
number. Byte, signed and unsigned 16-bit integer, float, and RGB color
files (modes 0, 1, 6, 2, 16) can be converted. The input file need not
be an MRC file; it can be any kind of file readable by other IMOD pro-
grams, plus the image file types readable by 3dmod using Qt libraries,
of which the most useful are PNG, JPEG, and BMP. There are options to
create JPEG or PNG output files instead of TIFF. The data will be
written to a TIFF file in the same mode as in the input file by
default, but there are options that allow the integer and float modes
to be scaled to bytes; the same options allow scaling of byte and RGB
values. If the input file has only one section, then the output file
will simply be the tiff_file_root_name with no suffix added.
When the output mode is integer or float, the TIFF file will contain
the minimum and maximum value for the data in the file. When a TIFF
stack is created, each image will have the same overall minimum and
maximum, rather than values specific to that image. This way, 3dmod
will be able to read in the data with proper scaling. However, Photo-
shop pays no attention to these values when displaying integer data, so
when the integer data have a relatively small range, the image will
initially appear black or very dark. The simplest solution for this
problem is to apply the scaling options described next.
When either the -S, the -C, or the -a option is entered, integer and
floating point data will be scaled to bytes and any values outside the
range of 0 to 255 will be truncated. If the input data is already in
bytes (either byte data or RGB data), these options will cause the data
to be scaled, and values outside the range truncated. To have the
range of the input data mapped into the full range of byte data, just
enter "-C 0,255". To have a subset of the input range mapped to 0 to
255, either enter the limits of the subset of the range directly with
the -S option, or adjust the contrast in 3dmod to give the desired
expansion of the dynamic range for that subset of the image data, and
enter the contrast settings withs the -C option.
Output to JPEG or PNG files is selected by entering the -j or the -p
option. There are some restrictions in these cases. An output stack
cannot be created, only a series of single-image files. The type of
compression cannot be specified, but the quality or degree of compres-
sion can. Tiling cannot be specified. Only byte and RGB output is
available. Integer and floating point files are converted to byte
based either on the minimum and maximum densities in the input file or
on the values entered with -S, -C, or -a. In addition to the fact that
JPEG images cannot be bigger than 65535 pixels in X or Y, problems may
be encountered with images much bigger than 1 gigapixel.
Since IMOD is distributed with version 4 of the TIFF library, this pro-
gram will write files bigger than 4 GB in the BigTIFF format, and con-
tinue to write smaller files in classical TIFF format.
When images are written to TIFF files with ZIP or LZW compression, the
program will use compress the data in parallel threads, with the number
of threads depending on the image size. However, if the image is
larger than a certain threshold, it is processed in chunks instead of
read in all at once, and the parallel compression routine cannot be
used. The default threshold is 400 MB. To ensure parallel compres-
sion, you can either raise this threshold with the -t option or over-
ride the default behavior with "-O 1". Note that with parallel com-
pression, memory usage will be about 2.2 times the size of the image.
OPTIONS
-s Create a single output file, a TIFF stack with all of the sec-
tions in the input file.
-j Produce JPEG files instead of TIFF files. See above for the
restrictions with this kind of output.
-p Produce PNG files instead of TIFF files. See above for the
restrictions.
-c value
Compress the data in TIFF files with the type of compression
indicated. The most useful compression types are LZW (value 5),
Zip or Deflate (value 8), and JPEG (value 7). For these types,
the value can be given as a text string: "lzw", "zip", or
"jpeg". The other allowed values are 32773 for Packbits and
32946 for a possibly different Deflate. These are available on
all operating systems when running with the TIFF libraries dis-
tributed with IMOD. Also note that JPEG compression is not
available with 16-bit data.
-q value
Set the quality for JPEG or ZIP compression (ZIP compression is
used in PNG output). For JPEG, the quality ranges from 0 to 100
with a default of 75; higher numbers preserve more information
but with less compression. For ZIP/PNG, which is a lossless
compression, the quality factor ranges from 1 to 9 with a
default of 6 and specifies a tradeoff between speed and amount
of compression. Higher numbers giving somewhat more compression
for more computational time. Lower numbers give somewhat less
compression with potentially much less computational time.
-S min,min
Specify an initial intensity scaling that will scale "min" to 0
and "max" to 255, just as is achieved with the -s options to
3dmod and mrcbyte.
-C black,white
Specify intensity scaling to match the scaling achieved with the
given black and white contrast settings in 3dmod, with the -c
option to mrcbyte, or the "-contrast" option to newstack. If
both -S and -C are entered, the final scaling will be the prod-
uct of these two scalings. If -S is entered alone, the
black/white values are 0/255 (no scaling for contrast). If -C
is entered alone, the initial scaling will be from the minimum
in the input file to 0, and from the maximum in the file to 255.
If you have integer data and you want to convert it to bytes,
you will generally want to apply some scaling with black/white
values to preserve the dynamic range of the data of interest.
-a mean,SD
Scale data for conversion to bytes so that it has the given mean
and standard deviation. Enter 0 for either value to use the
default for auto-contrasting in 3dmod, which is a mean of 150
and an SD of 40. This option cannot be entered with -S or -C,
since it supercedes the scaling specified by them.
-z min,max
Starting and ending Z slices in file to output, numbered from
0.
-i value
Initial file number. The default is the number of the first Z
slice output, where Z is numbered from 0.
-r value
Set the resolution setting of the output file in dots per inch
to the given value.
-P Use the pixel spacing in the MRC file header to set the resolu-
tion in the output file. The resolution will be set in dots per
inch and will typically be a very large number, but this value
allows the pixel spacing to be recovered if the file is con-
verted back with Tif2mrc.
-T xsize,ysize or size
Organize output data in tiles instead of strips, which can allow
faster access to subareas of very large images. Enter one num-
ber for the tile size in both dimensions, or two numbers to
specify a different size in X and Y. The TIFF library requires
that all tiles be the same size, which means that the image file
will be padded so that tiles on the bottom and right sides have
the same sizes as the rest. The program will pick an actual
tile size near the entered number that is a multiple of 16 and
that minimizes the amount of padding required.
-t value
Set the criterion for processing each image in strips 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 400 MB). Above the crite-
rion, images will be read, processed, and saved in strips
(unless "-O 1" is entered), with much lower memory requirements.
-O value
Override the default behavior on whether to compress data in
parallel. Specifically, a value of 0 will force data not to be
compressed in parallel if it ordinarily would be. A value of 1
will force large images to be compressed in parallel if they
would ordinarily be read in chunks instead.
-o Use old IMOD code instead of libtiff for writing the files (for
testing purposes).
AUTHORS
Jim Kremer, David Mastronarde
SEE ALSO
clip, newstack
BUGS
Complex and long integer data modes are not supported.
Email bug reports to mast at colorado dot edu.
IMOD 4.11.0 mrc2tif(1)