Converting a file to different format¶
The bfconvert command line tool can be used to convert files between supported formats.
bfconvert with no options displays a summary of available options.
To convert a file to single output file (e.g. TIFF):
bfconvert /path/to/input output.tiff
The output file format is determined by the extension of the output file, e.g. .tiff for TIFF files, .ome.tiff for OME-TIFF, .png for PNG.
- -option KEY VALUE¶
Passes options expressed as key/value pairs:
bfconvert -option key value /path/to/input /path/to/output
e.g. additional writer options, see Additional reader and writer options:
bfconvert -option ometiff.companion converted.companion.ome input.fake converted.ome.tiff
New in version 5.4.0.
Do not flatten resolutions into individual series. This option is mandatory to read images with pyramidal levels using the sub-resolution API and generate an output image with sub-resolutions. As of Bio-Formats 6.0.0, only the OME-TIFF output format properly supports this option:
bfconvert -noflat /path/to/input output-first-series.ome.tiff
New in version 6.0.0.
- -series SERIES¶
All images in the input file are converted by default. To convert only one series:
bfconvert -series 0 /path/to/input output-first-series.tiff
- -timepoint TIMEPOINT¶
To convert only one timepoint:
bfconvert -timepoint 0 /path/to/input output-first-timepoint.tiff
- -channel CHANNEL¶
To convert only one channel:
bfconvert -channel 0 /path/to/input output-first-channel.tiff
- -z Z¶
To convert only one Z section:
bfconvert -z 0 /path/to/input output-first-z.tiff
- -range START END¶
To convert images between certain indices (inclusive):
bfconvert -range 0 2 /path/to/input output-first-3-images.tiff
- -tilex TILEX, -tiley TILEY¶
All images larger than 4096×4096 will be saved as a set of tiles if the output format supports doing so. The default tile size is determined by the input format, and can be overridden like this:
bfconvert -tilex 512 -tiley 512 /path/to/input output-512x512-tiles.tiff
-tilexis the width in pixels of each tile;
-tileyis the height in pixels of each tile. The last row and column of tiles may be slightly smaller if the image width and height are not multiples of the specified tile width and height. Note that specifying
-tileywill cause tiles to be written even if the image is smaller than 4096×4096.
Also note that the specified tile size will affect performance. If large amounts of data are being processed, it is a good idea to try converting a single tile with a few different tile sizes using the
-cropoption. This gives an idea of what the most performant size will be.
- -crop X,Y,WIDTH,HEIGHT¶
For very large images, it may also be useful to convert a small tile from the image instead of reading everything into memory. To convert the upper-left-most 512×512 tile from the images:
bfconvert -crop 0,0,512,512 /path/to/file output-512x512-crop.tiff
The parameter to
-cropis of the format
x,y,width,height. The (x, y) coordinate (0, 0) is the upper-left corner of the image;
x + widthmust be less than or equal to the image width and
y + heightmust be less than or equal to the image height.
Images can also be written to multiple files by specifying a pattern string in the output file. For example, to write one series, timepoint, channel, and Z section per file:
bfconvert /path/to/input output_series_%s_Z%z_C%c_T%t.tiff
%s is the series index,
%z is the Z section index,
%c is the
channel index, and
%t is the timepoint index (all indices begin at 0).
For large images in particular, it can also be useful to write each tile to a separate file:
bfconvert -tilex 512 -tiley 512 /path/to/input output_tile_%x_%y_%m.jpg
%x is the row index of the tile,
%y is the column
index of the tile, and
%m is the overall tile index. As above, all
indices begin at 0. Note that if
%y is included in the file
name pattern, then the other must be included too. The only exception is if
%m was also included in the pattern.
Note for Windows Users: The command interpreter for batch files needs the
characters to be doubled in order to process the sequencing variables
correctly. So in Windows, the above example would read:
bfconvert /path/to/input output_series_%%s_Z%%z_C%%c_T%%t.tif
- -compression COMPRESSION¶
By default, all images will be written uncompressed. Supported compression modes vary based upon the output format, but when multiple modes are available the compression can be changed using the
-compressionoption. For example, to use LZW compression in a TIFF file:
bfconvert -compression LZW /path/to/input output-lzw.tiff
If the specified output file already exists, bfconvert will prompt to overwrite the file. When running bfconvert non-interactively, it may be useful to always allow bfconvert to overwrite the output file:
bfconvert -overwrite /path/to/input /path/to/output
To always exit without overwriting:
bfconvert -nooverwrite /path/to/input /path/to/output
To disable the conversion of lookup tables, leaving the output file without any lookup tables:
bfconvert -nolookup /path/to/input /path/to/output
New in version 5.2.1.
This option forces the writing of a BigTiff file:
bfconvert -bigtiff /path/to/input output.ome.tiff
New in version 5.1.2.
-bigtiffoption is not necessary if a BigTiff extension is used for the output file, e.g.:
bfconvert /path/to/input output.ome.btf
This option disables the automatic switching to BigTiff based upon the number of pixel bytes (TIFF files larger than 4GB):
bfconvert -nobigtiff /path/to/input output.ome.tiff
New in version 6.4.0.
-nobigtiffwill disable writing BigTiff when the output format is less than 4GB. It will not be able to write standard Tiff files greater than 4GB. An example of when it might be used would be when converting using a compression codec that reduces the size of the output file, e.g.:
bfconvert -nobigtiff -compression LZW /path/to/input output.ome.btf
This option is used alongside a pattern string when writing an image to multiple files. When set this will enforce zero padding on the filename indexes set in the provided pattern string:
bfconvert /path/to/input output_xy%sz%zc%ct%t.ome.tif -padded
New in version 5.2.2.
- -pyramid-resolutions RESOLUTIONS¶
- -pyramid-scale SCALE¶
-noflatby default, each series of the converted file will contain the same number of resolutions as in the input file. The
-pyramid-resolutionsoption allows to set the number of expected resolutions in the output file for each series. If the target number of resolutions is greater than the actual number of sub-resolutions present in the input file, additional pyramidal levels will be calculated using the downsampling factor specified by the
bfconvert -noflat -pyramid-resolutions 4 -pyramid-scale 2 /path/to/input out.ome.tiff
New in version 6.0.0.
This option will cache the initialized reader under the same directory as the input file after initialization:
bfconvert -cache /path/to/input output.ome.tiff
New in version 6.2.0.
- -cache-dir DIRECTORY¶
This option is to be used in conjunction with
-cache. When used it specifies the directory to store the cached initialized reader. If unspecified, the cached reader will be stored under the same folder as the image file:
bfconvert -cache-dir /path/to/store/cached/reader /path/to/input output.ome.tiff
New in version 6.2.0.
Do not preserve the OME-XML StructuredAnnotation elements:
bfconvert -no-sas /path/to/input output.ome.tiff
New in version 6.2.0.
Do not assume that planes are written in sequential order:
bfconvert -no-sequential /path/to/input output.ome.tiff
New in version 6.8.0.
- -swap DIMENSIONORDER¶
Overrides the default input dimension order:
bfconvert -swap XYZTC /path/to/input output.ome.tiff
New in version 6.9.0.
- -fill UNSIGNED_BYTE¶
Sets the fill value to use for undefined pixels. Valid values are 0-255, with 0 representing a black pixel and 255 representing a white pixel. The value set here will be applied to all bytes in an undefined pixel, e.g. setting to 128 (0x80) on a uint16 image will result in a pixel value of 32896 (0x8080). The default fill value is typically 0, but some readers may have different defaults.
New in version 6.13.0.