cbf_definition_rev


[IUCr Home Page] [CIF Home Page] [CBFlib]


Proposed Revised
DRAFT CBF/imgCIF DEFINITION
14 January 1999

Revisions
by
Herbert J. Bernstein
Bernstein + Sons, P.O. Box 177, Bellport, NY 11713-0177
yaya@bernstein-plus-sons.com

based on

DRAFT CBF DEFINITION

by
Andy Hammersley
European Synchrotron Radiation Facility, BP 200, Grenoble, 38043, CEDEX, France
hammersley@esrf.fr


This document and the CBF definitions are still subject to change. This document is a draft proposal for discussion.

This is a version of the CBF draft proposal, revised to include a coordinated pure ASCII ImgCIF definition, based on the Draft CBF Definition by Andy Hammersley, the work done at the Brookhaven imgCIF workshop, and the work on "CBFLIB: An ANSI-C API for Crystallographic Binary File" by Paul Ellis, ellis@SSRL.SLAC.STANFORD.EDU. For the binary CBF format, a "binary-string" approach, as proposed by Paul Ellis, is used, while for the ASCII imgCIF format, binary information is encoded using a variant on MIME (Multipurpose Internet Mail Extensions) format, which makes the CBF and ImgCIF formats very similar.

We have included an updated version of John Westbrook's DDL2-compliant CBF Extensions Dictionary, of Paul Ellis's CBFLIB manual, and examples of CBF/imgCIF files.

This is just a proposal. My apologies in advance, especially to Andy, John and especially to Paul for whatever I may have muddled here. Please be careful about basing any code on this until and unless there has been a general agreement.



Notices

Please read the NOTICES, which are part of this package, before making use of this software.


Most of this document is adapted from Andy's, so we follow his convention by "...[separating] the definition from comments on discussion items by using round brackets to refer to notes kept separate from the main text e.g. (1) refers to point 1 in the notes section.". We have integrated all comments to date into this document without special annotation.


A Draft Proposal
for
A Combined
Crystallographic Binary File (CBF)
and
Image-supporting Crystallographic Information File (ImgCIF)
Format

ABSTRACT

This document describes a proposal for a combined Crystallographic Binary File (CBF) and Image-supporting Crystallographic Information File (ImgCIF) format; a simple self-describing binary format for efficient transport and archiving of experimental data for the crystallographic community, and well as for the presentation of other image data, such as PICT, GIF and JPEG, within publication CIFs. With minor differences, both the binary CBF format and the ASCII ImgCIF have a similar, CIF-like structure. All the information other than actual binary data is presented as ASCII strings in both formats. The formats differ only in the handling of line termination and the actual presentation of the binary data of an image. The CBF format, presents binary information as a raw string of octets, while the ImgCIF format presents the binary information as ASCII-encoded lines. The format of the binary file, and the new CIF data-items are defined. In this document we concentrate on the representation of images per se. The CBF/imgCIF dictionary includes additional data items related to crystallographic data acquisition. Those additional data items are not discussed here.

Note:

  • All numbers are decimal unless otherwise stated.
  • The terms octet and byte refer to a group of eight bits.

1.0 INTRODUCTION

The Crystallographic Binary File (CBF) format is a complementary format to the Crystallographic Information File (CIF) [1], supporting efficient storage of large quantities of experimental data in a self-describing binary format (1). The Image-supporting Crystallographic Information File (ImgCIF) format is a proposed extension to CIF to assist in ASCII debugging and archiving of CBF files and to allow for convenient and standardized inclusion of images, such as maps, diagrams and molecular drawing into publication CIFs. It is our expectation that, for large images, the raw binary CBF format will be used both with in laboratories and for interchange among collaborating groups. For smaller chunks of binary data, either format should be be suitable, with the ASCII ImgCIF format being more appropriate for interchange and archiving.

The initial aim is to support efficient storage of raw experimental data from area-detectors (images) with no loss of information compared to existing formats. The format should be both efficient in terms of writing and reading speeds, and in terms of stored file sizes, and should be simple enough to be easily coded, or ported to new computer systems.

Flexibility and extensibility are required, and later the storage of other forms of data may be added without affecting the present definitions.

The aims are achieved by a simple file format, consisting of lines of ASCII information defining information about the binary data as CIF tag-value pairs and tables, and either raw octets of binary data in delimited sections, or ASCII-based presentations of the same binary information in similarly delimited sections.

The present version of the format only tries to deal with simple Cartesian data. This is essentially the "raw" data from detectors that is typically stored in commercial formats or individual formats internal to particular institutes, but could be other forms of data. It is hoped that CBF can replace individual laboratory or institute formats for "home" built detector systems, be used as a inter-program data exchange format, and may be offered as an output choice by a number of commercial detector manufacturers specialising in X-ray and other detector systems.

This format does not imply any particular demands on processing software nor on the manner in which such software should work. Definitions of units, coordinate systems, etc. may quite different. The clear precise definitions within CIF, and hence CBF, help, when necessary, to convert from one system to another. Whilst no strict demands are made, it is clearly to be hoped that software will make as much use as is reasonable of information relevant to the processing which is stored within the file. It is expected that processing software will give clear and informative error messages when they encounter problems within CBF's or do not support necessary mechanisms for inputting a file.

1.1 CBF and "imgCIF"

CBF and "imgCIF" are two aspects of the same format. Since CIF's are pure ASCII text files, a separate binary format is needed to be defined to allow the combination of pseudo-ASCII sections and binary data sections. The binary file format is the Crystallographic Binary File (CBF). The ASCII sections are very close to the CIF standard, but must use operating system independent "line separators". In describing the ASCII sections, we use the notation "\r\n" (for the pair of characters carriage return, line-feed) for a line terminator would allow the ASCII sections to viewed with standard system utilities on a very wide range of operating systems. However, an API to read the binary format must accept any of the following three alternative line terminators as the end of an ascii line: "\r", "\n" or "\r\n". An API to write CBF should write "\r\n" as the line terminator, if at all possible.

imgCIF is also the name of the CIF dictionary which contains the terms specific to describing the binary data (the orginal, designed by John Westbrook, without the modifications in this proposal is avaliable from http://ndbserver.rutgers.edu/NDB/mmcif. Thus a CBF or ImgCIF files uses data names from the imgCIF dictionary and other CIF dictionaries.

2.0 A SIMPLE EXAMPLE

Before fully describing the format we start by showing a simple, but important and complete usage of the format; that of storing a single detector image in a file together with a small amount of useful auxiliary information. It is intened to be a useful example for people who like working from examples, as opposed to full definitions. It should also serve as an introduction or overview of the format defintion. This example uses CIF DDL2 based dictionary items.

The example is an image of 768 by 512 pixels stored as 16 bit unsigned integers, in little endian byte order. (This is the native byte ordering on a PC.) The pixel sizes are 100.5 by 99.5 microns. Comment lines starting with a hash sign (#) are used to explain the contents of the header. Only the ASCII part of the file is shown, but comments are used to describe the start of the binary section.

First the file is shown with the minimum of comments that a typical outputting program might add. Then it is repeated, but with "over- commenting" to explain the format.

Here is how a file might appear if listed on a PC or on a Unix system using "more":

###CBF: VERSION 0.6
# Data block for image 1
data_image_1

_entry.id 'image_1'

                                  
# Sample details
_chemical.entry_id                           'image_1'
_chemical.name_common                        'Protein X'

# Experimental details
_exptl_crystal.id                            'CX-1A'
_exptl_crystal.colour                        'pale yellow'

_diffrn.id                                    DS1
_diffrn.crystal_id                            'CX-1A' 

_diffrn_measurement.diffrn_id                 DS1
_diffrn_measurement.method                    Oscillation
_diffrn_measurement.sample_detector_distance  0.15 
                                                  
_diffrn_radiation_wavelength.id               L1 
_diffrn_radiation_wavelength.wavelength       0.7653 
_diffrn_radiation_wavelength.wt               1.0

_diffrn_radiation.diffrn_id                   DS1 
_diffrn_radiation.wavelength_id               L1 

_diffrn_source.diffrn_id                      DS1
_diffrn_source.source                         synchrotron
_diffrn_source.type                          'ESRF BM-14'

_diffrn_detector.diffrn_id                    DS1
_diffrn_detector.id                           ESRFCCD1
_diffrn_detector.detector                     CCD
_diffrn_detector.type                        'ESRF Be XRII/CCD'


_diffrn_detector_element.id                   1
_diffrn_detector_element.detector_id          ESRFCCD1


_diffrn_frame_data.id                         F1
_diffrn_frame_data.detector_element_id        1
_diffrn_frame_data.array_id                   'image_1'
_diffrn_frame_data.binary_id                  1


# Define image storage mechanism
#

loop_
_array_structure.id 
_array_structure.encoding_type        
_array_structure.compression_type     
_array_structure.byte_order           
image_1       "unsigned 16-bit integer"  none  little_endian
                                      
loop_
_array_intensities.array_id    
_array_intensities.binary_id       
_array_intensities.linearity          
_array_intensities.undefined_value    
_array_intensities.overload_value     
image_1     1    linear     0      65535

# Define dimensionality and element rastering
loop_
_array_structure_list.array_id
_array_structure_list.index
_array_structure_list.dimension
_array_structure_list.precedence
_array_structure_list.direction
image_1    1      768    1    increasing    
image_1    2      512    2    decreasing     

loop_
_array_element_size.array_id
_array_element_size.index
_array_element_size.size
image_1  1  100.5e-6
image_1  2  99.5e-6

loop_
_array_data.array_id
_array_data.binary_id
_array_data.data

image_1 1
;
--CIF-BINARY-FORMAT-SECTION--
Content-Type: application/octet-stream;
     conversions="x-CBF_PACKED"
Content-Transfer-Encoding: BINARY
X-Binary-Size: 374578
X-Binary-ID: 1
X-Binary-Element-Type: "unsigned 16-bit integer"
Content-MD5: jGmkxiOetd9T/Np4NufAmA==

START_OF_BIN
*************'9*****`********* ...
[This is where the raw binary data would be -- we can't print it here]

--CIF-BINARY-FORMAT-SECTION----
;
Here the file is shown again, but this time with many comment lines added to explain the format:
###CBF: VERSION 0.6

# This line starting with a "#" is a CIF and CBF comment line,
# but the first line with the three "#"s is a CBF identifier.
# (a "magic number")  The text "###_CBF: VERSION" identifies
# the file as a CBF and must be present as the very first line of
# every CBF file. Following "VERSION" is the version number of 
# the  file, which is the corresponding version of the CBF/imgCIF
# extensions dictionary and supporting documentation.   A version 
# 0.6 CIF should be readable by any program which fully supports 
# the version 1.0 CBF definitions.

# Comment lines and white space (blanks and new lines) may appear
# anywhere outside the binary sections.
  
# In a CIF, the descriptive tags and values may be presented in
# any convenient order, e.g. the data could come first, and the
# parameters necessary to interpret the data could come later.
# This order-independent convention holds for an imgCIF file, but
# for a CBF, all the tags and values describing binary data (i.e.
# all the tags other than those in the ARRAY_DATA category) should
# be presented before the binary data, in the form of a header.
# This does not mean that there cannot be more useful information
# after the binary data.  There could be another full header and
# more blocks of binary data.  All we are saying is that, in
# the interest of efficiency in processing a CBF, the parameters 
# that relate to a particular block of binary data must appear 
# earlier in the CBF than the block itself.

# The header begins with "data_", which is the CIF token to 
# identify a data block.  Within a data block, any given tag may 
# be presented only once, either directly with an immediately 
# following  value, or as one of the column headings for the rows
# of a table.  If you will need to resuse the same tag, you will 
# have to start a new data block.

# Data block for image 1
data_image_1

# We've chosen to call this data block 'image_1', but this was an 
# arbitary choice. Within a data block a data item may only be used 
# once.

_entry.id 'image_1'
                                  
# Sample details
_chemical.entry_id                           'image_1'
_chemical.name_common                        'Protein X'

# The apostrophes enclose the string which contains a space.
# A double quote (") could have been used, just as well.
# There is also a third way to quote string, with the string
# "\n;", i.e. with a semicolon at the beginning of a line
# which allows multi-lined strings to be presented.  We'll
# use that form of text quotation for the binary data.

# Experimental details
_exptl_crystal.id                            'CX-1A'
_exptl_crystal.colour                        'pale yellow'

_diffrn.id                                    DS1
_diffrn.crystal_id                            'CX-1A' 

_diffrn_measurement.diffrn_id                 DS1
_diffrn_measurement.method                    Oscillation
_diffrn_measurement.sample_detector_distance  0.15 
                                                  
_diffrn_radiation_wavelength.id               L1 
_diffrn_radiation_wavelength.wavelength       0.7653 
_diffrn_radiation_wavelength.wt               1.0

_diffrn_radiation.diffrn_id                   DS1 
_diffrn_radiation.wavelength_id               L1 

_diffrn_source.diffrn_id                      DS1
_diffrn_source.source                         synchrotron
_diffrn_source.type                          'ESRF BM-14'

_diffrn_detector.diffrn_id                    DS1
_diffrn_detector.id                           ESRFCCD1
_diffrn_detector.detector                     CCD
_diffrn_detector.type                        'ESRF Be XRII/CCD'


_diffrn_detector_element.id                   1
_diffrn_detector_element.detector_id          ESRFCCD1


_diffrn_frame_data.id                         F1
_diffrn_frame_data.detector_element_id        1
_diffrn_frame_data.array_id                   'image_1'
_diffrn_frame_data.binary_id                  1

# Many more data items can be defined, but the above gives the idea
# of a useful minimum set (but not minimum in the sense of 
# compulsory, the above data items are optional in a CIF or CBF).
 
# Define image storage mechanism
#
# Notice that we did not include a binary ID here.  The idea of
# the ARRAY_STRUCTURE category is to present parameters which
# could be common to multiple blocks of binary data, which would 
# all have the same array ID, but would have distinct binary ID's

loop_
_array_structure.id 
_array_structure.encoding_type        
_array_structure.compression_type     
_array_structure.byte_order           
image_1      "unsigned 16-bit integer"  none  little_endian
                                      
# On the other hand, we do provide a binary ID for ARRAY_INTENSITIES,
# since there might be different paremeters for each binary block. 
# We could have left it out here, since there is only one block and
# the default binary ID happens to be 1

loop_
_array_intensities.array_id  
_array_intensities.binary_id       
_array_intensities.linearity          
_array_intensities.undefined_value    
_array_intensities.overload_value     
image_1     1   linear     0      65535

# Define dimensionality and element rastering

# Here the size of the image and the ordering (rastering) of the  data 
# elements is defined. The CIF "loop_" structure is used to
# define different dimensions. (It can be used for defining multiple
# images.)

loop_
_array_structure_list.array_id
_array_structure_list.index
_array_structure_list.dimension
_array_structure_list.precedence
_array_structure_list.direction
image_1    1      768    1    increasing    
image_1    2      512    2    decreasing     

loop_
_array_element_size.array_id
_array_element_size.index
_array_element_size.size
image_1  1  100.5e-6
image_1  2  99.5e-6


# The "array_id" identifies data items belong to the same array. 
# Here we have chosen the name "image_1", but another name could 
# have been used, so long as it's used consistently. The "index" 
# component refers to the dimension being defined, and the 
# "dimension" component defines  the number of elements in that 
# dimension. The "precedence" component defines which precedence 
# of rastering of the data. In this case the first dimension is the faster 
# changing dimension. The "direction" component tells us the 
# direction in which the data rasters within a dimension. Here the 
# data  rasters faster from minimum elements towards the maximum 
# element ("increasing") in the first dimension, and more 
# slowly from the maximum element towards the minimum element in 
# the second dimension. (This is the default rastering order.)

# The storage of the binary data is now fully defined.

# Further data items could be defined, but  we are ready to
# present the data.  That is done with the ARRAY_DATA category.
# The start of this category marks the end of the header
# (Well, almost the end, there is a bit more header information
# below).

loop_
_array_data.array_id
_array_data.binary_id
_array_data.data

image_1 1

# The binary data itself will come just a little further down,
# as the essential part of the value of _array_data.data, which 
# begins as semicolon-quoted text.  The line immediately after 
# the line with the semicolon is a MIME boundary marker.  As for
# all MIME boundary markers, it begins with "--".  The next
# few lines are MIME headers, describing some useful information
# we will need to process the binary section.  MIME headers can
# appear in different orders, and can be very confusing (look
# at the raw contents of a email message with attachments), but
# there is only a few headers which is have to be understood to
# process a CBF: 
#
#      The "Content-Type" header may be any of discrete types 
#      permitted in RFC 2045; "application/octet-stream" is 
#      recommended.  If an octet stream was compressed, the 
#      compression should be specified by the parameter 
#      'conversions="x-CBF_PACKED"' or by specifying 
#      one of the other compression types.
#          
#      The "Content-Transfer-Encoding" header should be 'BINARY' 
#      for a CBF.  We'll consider the other values used for imgCIF 
#      below.
#                           
#      The "X-Binary-Size" header specifies the size of the
#      binary data in octets.  If compression was used, this size 
#      is the  size after compression, including any book-keeping
#      fields, but not the 8 bytes for the compression type.
#
#      The "X-Binary-Element-Type" header specifies the 
#      type of binary data in the octets, using the same 
#      descriptive phrases as in _array_structure.encoding_type. 
#      The default value is "unsigned 32-bit integer".
#
# The MIME header items are followed by an empty line and then by
# a special sequence marked here as 'START_OF_BIN', consisting of
# Control-L, Control-Z, Control-D to stop printing on most systems,
# and then a single binary flag  character of hexadecimal value D5 
# (213 decimal).  The binary data follows immediately after this 
# flag character.
#
;
--CIF-BINARY-FORMAT-SECTION--
Content-Type: application/octet-stream;
     conversions="x-CBF_PACKED"
Content-Transfer-Encoding: BINARY
X-Binary-Size: 374578
X-Binary-ID: 1
X-Binary-Element-Type: "unsigned 16-bit integer"
Content-MD5: jGmkxkrpnizOetd9T/Np4NufAmA==

START_OF_BIN
*************'9*****`********* ...
[This is where the raw binary data would be -- we can't print it here]

--CIF-BINARY-FORMAT-SECTION----
;

# After the last octet (i.e. byte) of the binary data, there is a
# special trailer "\n--CIF-BINARY-FORMAT-SECTION----\n;"
# which repeats the initial bounday marker with an extra "--"
# at the end (a MIME convention for the last boundary marker), and
# then the closing semicolon quote for a text section.  This
# is essential in an imgCIF, and we include it in a CBF for 
# consistency.

OVERVIEW OF THE FORMAT

This section describes the major "components" of the CBF format.
  1. CBF is a binary file, containing self-describing array data e.g. one or more images, and auxiliary data e.g. describing the experiment.

  2. Except for the handling of line terminators, the way binary data is presented, and more liberal rules in ordinger information, an ASCII imgCIF file is the same as a CBF binary file.

  3. A CBF consists of pseudo-ASCII text header sections, which are "lines" of no more than 80 ASCII characters separated by "line separators" which are the pair of ASCII characters carriage return and line-feed (ASCII 13, ASCII 10), followed by zero, one, or more binary sections presented as "binary strings". This structure may be repeated.

  4. An imgCIF consists of ASCII lines of no more than 80 characters using the the normal line termination conventions of the current system (e.g. ASCII 10 in UNIX) with MIME-encoded binary strings at any appropriate point in the file.

  5. The very start of the file has an identification item (magic number) (2). This item also describes the CBF version or level. The identifier is:
    ###CBF: VERSION
    
    which must always be present so that a program can easily identify whether or not a file is a CBF, by simply inputting the first 15 characters. (The space is a blank (ASCII 32) and not a tab. All identifier characters are uppercase only.)

    The first hash means that this line within a CIF would be a comment line, but the three hashes mean that this is a line describing the binary file layout for CBF. (All CBF internal identifiers start with the three hashes, and all other must immediately follow a "line separator".) No whitespace may precede the first hash sign.

    Following the file identifier is the version number of the file. e.g. the full line might appear as:

    ###CBF: VERSION 0.6
    
    The version number must be separated from the file identifier characters by whitespace e.g. a blank (ASCII 32).

    The version number is defined as a major version number and minor version number separated by the decimal point. A change in the major version may well mean that a program for the previous version cannot input the new version as some major change has occurred to CBF (3). A change in the minor version may also mean incompatibility, if the CBF has been written using some new feature. e.g. a new form of linearity scaling may be specified and this would be considered a minor version change. A file containing the new feature would not be readable by a program supporting only an older version of the format.

    Note: Until we reach major version 1 (the first official release), the rules are a little more relaxed. While there will be some effort at upwards compatability, in order to ensure a reasonable agreed specification without too many strange artifacts, changes between minor versions may, unfortunately, introduce incompatabilities which require program changes to still read CBFs compliant with an earlier draft, e.g. the change in the "magic number" and from binary sections to binary strings in going to version 0.3, and a removal of the redundant parts of the binary header in going to version 0.6. Naturally, such changes should be sufficiently well documented to allow for conversions.>>>

  6. Header Information:

    1. The start of an header section is delimited by the usual CIF "data_" token. Optionally, the formerly specified header identifier,
      ###_START_OF_HEADER
      
      may be used before the "data_" taken, followed by the carriage return, line-feed pair, as an aid in debugging, but it is no longer required. (Naturally, another carriage return, line-feed pair should immediately precedes this and all other CBF identifiers, with the exception of the CBF file identifier which is at the very start of the file.)

    2. A header section, including the identification items which delimit it, uses only ASCII characters, and is divided into "lines". The "line separator" symbols, "\r\n" (carriage return, line-feed) are the same regardless of the operating system on which the file is written. (This is an importance difference with CIF, but must be so, as the file contains binary data, so cannot be translated from one O.S. to another, which is the case for ASCII text files.) While a properly functioning CBF API should write the full "\r\n" line separator, it should recognize any of three sequences "\r", "\n", "\r\n" as valid line separators, so that hand-edited headers will not be rejected.

    3. The header section within the delimiting identification items obeys all CIF rules [1] with the exception of the line separators.

      e.g.

      • "Lines" are a maximum of 80 characters long. (For CBF it is probably best to allow for this maximum to be larger.)

      • The tokens "data_" and "loop_" have special meaning to CIF, and should not be used except in their indicated places. The tokens "save_", "stop_" and "global_" also have special meaning to CIF's parent language, STAR, and also should not be used.

      • All data names (tags) start with an underscore character.

      • The hash symbol (#) (outside a character string) means that all text up to the line separator is a comment.

      • Whitespace outside of character strings is not significant.

      • Data names are case insensitive.

      • The data item follows the data name separator, and may be of one of two types: character string (char) or number (numb). (The type is specified for each data name.)

      • Character strings may be delimited with single of double quotes, or blocks of text may be delimited by semi-colons occurring as the first character on a line.

      • The "loop_" mechanism allows a data name to have multiple values. Immediately following the "loop_", one or more data names are listed without their values, as column headings. Then one or more rows of values are given.

      Any CIF data name may occur within the header section.

    4. A single header section may contain one or more data blocks (CIF terminology).

    5. The end of the header information is marked by coming to the tags from the "ARRAY_DATA" category. The formerly specifier special identifier:
      ###_END_OF_HEADER
      
      followed by carriage return, line-feed, may be used as well as an aid to debugging, but it is not required.

  7. The header information must contain sufficient data names to fully describe the binary data section(s) which follow(s).

  8. Binary Information:

    Note: Under CBFlib "binary sections" have been replaced by "binary strings" values within a data name/value pair. The structure of the proposed "binary string" is similar to the former binary sections, but there are significant differences.

    1. Before getting to the binary data, itself, there are some preliminaries to allow a smooth transition from the conventions of CIF to those of raw streams of "octets" (8-bit bytes). The binary data is given as the essential part of a specially formatted semicolon-delimited CIF multi-line text string. This text string is the value associated with the tag "_array_data.data".

    2. Within that text string, the conventions developed for transmitting email messages including binary attachments are followed. There is secondary ASCII header information, formatted as Multipurpose Internet Mail Extensions (MIME) headers (see RFCs 2045-49 by Freed, et. al). The bounday marker for the beginning of all this is the special string


      --CIF-BINARY-FORMAT-SECTION--


      at the beginning of a line. The initial "--" says that this is a MIME boundary. We cannot put "###" in front of it and conform to MIME conventions. Immediately after the boundary marker are MIME headers, describing some useful information we will need to process the binary section. MIME headers can appear in different orders, and can be very confusing (look at the raw contents of a email message with attachments), but there is only a few headers with a narrow range of values which is have to be understood to process a CBF (as opposed of an imgCIF, for which the headers can be more varied):

      • The "Content-Type" header may be any of discrete types permitted in RFC 2045; "application/octet-stream" is recommended. If an octet stream was compressed, the compression should be specified by the parameter 'conversions="x-CBF_PACKED"' or by specifying one of the other compression types.

      • The "Content-Transfer-Encoding" header should be 'BINARY' for a CBF. We'll consider the other values used for imgCIF below.

      • The "X-Binary-Size" header specifies the size of the binary data in octets. If compression was used, this size is the size after compression, including any book-keeping fields, but not the 8 bytes for the compression type.

      • The "X-Binary-Element-Type" header specifies the type of binary data in the octets, using the same descriptive phrases as in _array_structure.encoding_type. The default value is "unsigned 32-bit integer".

      The MIME header items are followed by an empty line and then by a special sequence marked here as 'START_OF_BIN', consisting of Control-L, Control-Z, Control-D to stop printing on most systems, and then a single binary flag character of hexadecimal value D5 (213 decimal). The binary data follows immediately after this flag character.


      In general, if the value given for "Content-Transfer-Encoding" is one of the real encodings: "BASE64", "QUOTED-PRINTABLE", "X-BASE8", "X-BASE10" or "X-BASE16", this file is an imgCIF.

      For either a CBF or an imgCIF the optional "Content-MD5" header provides a much more sophisticated check on the integrity of the binary data.


      In a CBF, the raw binary data begins after an empty line terminating the MIME headers and after the START_OF_BIN identifier. "START_OF_BIN" contains bytes to separate the "ASCII" lines from the binary data, bytes to try to stop the listing of the header, bytes which define the binary identifier which should match the "binary_id" defined in the header, and bytes which define the length of the binary section.


      Octet Hex Decimal Purpose
      1 0C 12 (ctrl-L) End the current page
      2 1A 26 (ctrl-Z) Stop listings in MS-DOS
      3 04 04 (Ctrl-D) Stop listings in UNIX
      4 D5 213 Binary section begins
      5..5+n-1     Binary data (n octets)


      Only bytes 5..5+n-1 are encoded for an imgCIF file using the indicated Content-Transfer-Encoding.

      Note: Earlier versions of the specification included three 8-byte words of information in binary which replicated information now available in the MIME header:

      5..12     Binary Section Identifier
      (See _array_data.binary_id)
      64-bit, little endian
      13..20     the size (n) of the
      binary section in octets
      (i.e. the offset from octet
      29 to the first byte following
      the data)
      21..28     Compression type:
      CBF_NONE 0x0040 (64)
      CBF_CANONICAL 0x0050 (80)
      CBF_PACKED 0x0060 (96)
      ...  
      followed by binary data. These three 8-byte words are no longer included when a MIME header is provided. In addition, in still earlier versions, the size given in the second 8-byte word was n+8, rather than n.


      The binary characters serve specific purposes:


      • The Control-L will terminate the current page in listings on most operating systems.


      • The Control-Z will stop the listing of the file on MS-DOS type operating systems.


      • The Control-D will stop the listing of the file on Unix type operating systems.


      • The unsigned byte value 213 (decimal) is binary 11010101. (Octal 325, and hexadecimal D5). This has the eighth bit set so can be used for error checking on 7-bit transmission. It is also asymmetric, but with the first bit also set in the case that the bit order could be reversed (which is not a known concern).


      • (The carriage return, line-feed pair before the START_OF_BIN and other lines can also be used to check that the file has not been corrupted e.g. by being sent by ftp in ASCII mode.)



    3. The "line separator" immediately precedes the "start of binary identifier", but blank spaces may be added prior to the preceding "line separator" if desired (e.g. to force word or block alignment).


    4. The binary data does not have to completely fill the bytes defined by the byte length value, but clearly cannot be greater than this value (except when the value zero has been stored, which means that the size is unknown, and no other headers follow). The values of any unused bytes are undefined.


    5. At exactly the byte following the full binary section as defined by the length value is the end of binary section identifier. This consists of the carriage return / line feed pair followed by:


      --CIF-BINARY-FORMAT-SECTION--
      ;


      with each of these lines followed by the carriage return / line feed pair. This brings us back into a normal CIF environment


      The first "line separator" separates the binary data from the pseudo-ASCII line.


      This identifier is in a sense redundant since the binary data length value tells the a program how many bytes to jump over to the end of the binary data. However, this redundancy has been deliberately added for error checking, and for possible file recovery in the case of a corrupted file.


      This identifier must be present at the end of every block of binary data.


  9. Whitespace may be used within the pseudo-ASCII sections prior to the "start of binary section" identifier to align the start binary data sections to word or block boundaries. Similar use may be made of unused bytes in binary sections. However, no blank lines should be introduced among the MIME headers, since that would terminate processing of those headers and start the scan for binary data.


    However, in general no guarantee is made of block nor word alignment in a CBF of unknown origin.


  10. The end of the file need not be not explicitly indicated, but including a comment of the form:


    ###_END_OF_CBF


    (including the carriage return, line-feed pair) can help in debugging.


  11. All binary data described in a single data block should follow the header section prior to another data block, or the end of the file, so allow for the most efficient processing of CBF files. However, since binary strings can be parsed anywhere within the context of a CBF or imgCIf file, it is recommended that processing software from CBF accept such strings in any order and it is mandatory that processing software for imgCIF accept such string in any order.


    The binary identifier values used within a given data block section, and hence the binary data must be unique for any given array_id, and, it would be best to make them truly unique.


    A different data block may reuse binary identifier values.


    (This allows concatenation of files without re-numbering the binary identifiers, and provides a certain level of localization of data within the file, to avoid programs having to search potentially huge files for missing binary sections.)


  12. The recommended file extension for a CBF is: cbf
    This allows users to recognise file types easily, and gives programs a chance to "know" the file type without having to prompt the user. Although they should check for at least the file identifier to ensure that the file type is indeed a CBF.


  13. The recommended file extensions for imgCIF are: icf or cif
    (use of "cif" subject to IUCr approval).


  14. CBF format files are binary files and when ftp is used to transfer files between different computer systems "binary" or "image" mode transfer should be selected.


  15. imgCIF files are ASCII files and when ftp is used to transfer files between different computer systems "ascii" transfer should be selected.

3.1 SIMPLE EXAMPLE OF THE ORDERING OF IDENTIFIERS

Here only the ASCII part of the file structuring identifiers is shown. The CIF data items are not shown, apart from the "data_" identifier which indicates the beginning of a data block.

This shows the structuring of a simple example e.g. one header section followed by one binary section. Such as could be used to store a single image.

###CBF: VERSION 0.3

data_

### ... various CIF tags and values here

loop_
array_data.id
array_data.binary_id
array_data.data

image_1 1
;
--CIF-BINARY-FORMAT-SECTION--
Content-Type: application/octet-stream;
     conversions="x-CBF_PACKED"
Content-Transfer-Encoding: BINARY
X-Binary-ID: 1
Content-MD5: jGmkxiOetd9T/Np4NufAmA==

START_OF_BIN
*************'9*****`********* ...
[This is where the raw binary data would be -- we can't print it here]

--CIF-BINARY-FORMAT-SECTION----
;
###_END_OF_CBF

3.2 MORE COMPLICATED EXAMPLE OF THE ORDERING OF IDENTIFIERS

Here only the ASCII part of the file structuring identifiers is shown. The CIF data items are not shown, apart from the "data_" identifier which indicates the beginning of a data block.

This shows the a possible structuring of a more complicated example. Two header sections, the first contains two data blocks and defines three binary sections. CIF comment lines, starting with a hash (#) are used to example the structure.

###CBF: VERSION 0.6
# CBF file written by cbflib v0.6

# A comment cannot appear before the file identifier, but can appear
# anywhere else, except within the binary sections.

# Here the first data block starts
data_

### ... various CIF tags and values here
###     but none that define array data items


# The "data_" identifier finishes the first data block and starts the
# second
data_

### ... various CIF tags and values here
###     including ones that define array data items

loop_
array_data.array_id
array_data.binary_id
array_data.data

image_1 1
;
--CIF-BINARY-FORMAT-SECTION--
Content-Type: application/octet-stream;
     conversions="x-CBF_PACKED"
Content-Transfer-Encoding: BINARY
X-Binary-Size: 3745758
X-Binary-ID: 1
X-Binary-Element-Type: "signed 32-bit integer"
Content-MD5: 1zsJjWPfol2GYl2V+QSXrw==

START_OF_BIN
<D5>^P<B8>P^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ ...
[This is where the raw binary data would be -- we can't print it here]

--CIF-BINARY-FORMAT-SECTION----
;

# Following the "end of binary" identifier the file is pseudo-ASCII
# again, so comments are valid up to the next "start of binary"
# identifier.  Note that we have bumped the binary ID.

image_1 2
;
--CIF-BINARY-FORMAT-SECTION--
Content-Type: application/octet-stream;
     conversions="x-CBF_PACKED"
Content-Transfer-Encoding: BINARY
X-Binary-Size: 3745758
X-Binary-ID: 2
X-Binary-Element-Type: "signed 32-bit integer"
Content-MD5: xR5kxiOetd9T/Nr5vMfAmA==

START_OF_BIN
<D5>^P<B8>P^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ ...
[This is where the raw binary data would be -- we can't print it here]

--CIF-BINARY-FORMAT-SECTION----
;

# Third binary section, note that we have a new array id.

image_2 3
;
--CIF-BINARY-FORMAT-SECTION--
Content-Type: application/octet-stream;
     conversions="x-CBF_PACKED"
Content-Transfer-Encoding: BINARY
X-Binary-ID: 3
Content-MD5: yS5kxiOetd9T/NrqTLfAmA==

START_OF_BIN
*************'9*****`********* ...
[This is where the raw binary data would be -- we can't print it here]

--CIF-BINARY-FORMAT-SECTION----
;


# Second Header section

data_

### ... various CIF tags and values here
###     including ones that define array data items

# Since we only have one block left, we won't use a loop


array_data.id         image
array_data.binary_id  1
array_data.data

# Note that I can put a comment here
;
--CIF-BINARY-FORMAT-SECTION--
Content-Type: application/octet-stream;
     conversions="x-CBF_PACKED"
Content-Transfer-Encoding: BINARY
X-Binary-ID: 1
Content-MD5: fooxiOetd9T/serNufAmA==

START_OF_BIN
*************'9*****`********* ...
[This is where the raw binary data would be -- we can't print it here]

--CIF-BINARY-FORMAT-SECTION----
;

###_END_OF_CBF

DATA NAME CATEGORIES

John Westbrook has proposed a number of data name categories as part of his DDL2 based "imgCIF" dictionary. This category list may be expanded to cover a structuring of the often multiple data-sets which might be used in a structurial investigation. Here we only consider the categories concerned with storing an image (or other N-dimensional topographically regular cartesian grid).

The _array_* categories cover all data names concerned with the storage of images or regular array data.

Data names from any of the existing categories may be relevant as auxiliary information in the header section, but data names from the _diffrn_ category, are likely to be the most relevant, and a number of new data names in this category are necessary.

The "array" Class of Binary Data

The "array" class is used to store regular arrays of data values, such as 1-D histograms, area-detector data, series of area-detector data, and volume data. Normally such data is regularly spaced in space or time, however spatial distorted data could nevertheless be stored in such a format. There is only one data "value" stored per lattice position, although that value may be of type complex.

The "array" class is defined by data names from the ARRAY_STRUCTURE and ARRAY_STRUCTURE_LIST categories.

Here is a short summary of the data names and their purposes.

  • _array_structure.id: Alpha numeric identifier for the array structure
  • _array_structure.compression_type: Type of data compression used
  • _array_structure.byte_order: Order of bytes for multi-byte integer or reals
  • _array_structure.encoding_type: Native data type used to store elements.

    e.g. "unsigned_16_bit_integer" is used if the stored image was 16 bit unsigned integer values, regardless of any compression scheme used.

"Array" Dimensions and Element Rastering and Orientation

The array dimension sizes, i.e. the number of elements in each dimension are defined by _array_structure_list.dimension. Which takes an integer value. This is used in a loop together with the _array_structure_list.index item to define the different dimensions for one or more arrays.

Fundamental to treating a long line of data values as a 2-D image or an N-dimensional volume or hyper-volume is the knowledge of the manner in which the values need to be wrapped. For the raster orientation to be meaningful we define the sense of the view:

For a detector image the sense of the view is defined as that looking from the crystal towards the detector.

(For the present we consider only an equatorial plane geometry, with 2-theta = 0; the detector as being vertically mounted.)

The rastering is defined by the three data names _array_structure_list.index, _array_structure_list.precedence, and _array_structure_list.direction data names.

index refers to the dimension index i.e. In an image 1 refers to the X-direction (horizontal), 2 refers to the Y-direction (vertical).

precedence refers to the order in which the data in wrapped.

direction refers the direction of the rastering for that index.

We define a preferred rastering orientation, which is the default if the keyword is not defined. This is with the start in the upper-left-hand corner and the fastest changing direction for the rastering horizontally, and the slower change from top to bottom.

(Note: With off-line scanners the rastering type depending on which way round the imaging plate or film is entered into the scanner. Care may need to be taken to make this consistent.)

"Array_Structure" Examples

To define an image array of 1300 times 1200 elements, with the raster faster in the first dimension, from left to right, and slower in the second dimension from top to bottom, the following header section might be used:

# Define image size and rastering
loop_
_array_structure_list.array_id
_array_structure_list.index
_array_structure_list.dimension
_array_structure_list.precedence
_array_structure_list.direction
image_1    1      1300    1    increasing
image_1    2      1200    2    decreasing
To define two arrays, the first a volume of 100 times 100 times 50 elements, fastest changing in the first dimension, from left to right, changing from bottom to top in the second dimension, and slowest changing in the third dimension from front to back; the second an image of 1024 times 1280 pixels, with the second dimension changing fastest from top to bottom, and the first dimension changing slower from left to right; the following header section might be used:

# Define array sizes and rasterings
loop_
_array_structure_list.array_id
_array_structure_list.index
_array_structure_list.dimension
_array_structure.precedence
_array_structure.direction
volume_a    1      100    1    increasing
volume_a    2      100    2    increasing
volume_a    3       50    3    increasing
slice_1     1      1024   2    increasing
slice_1     2      1280   1    decreasing

"Array" Element Intensity Scaling

Existing data storage formats use a wide variety of methods for storing physical intensities as element values. The simplest is a linear relationship, but square root and logarithm scaling methods have attractions and are used. Additionally some formats use a lower dynamic range to store the vast majority of element values, and use some other mechanism to store the elements which over-flow this limited dynamic range. The problem of limited dynamic range storage is solved by the data compression methods byte_offsets and predictor_huffman (see next Section), but the possibility of defining non-linear scaling must also be provided.

The _array_intensities.linearity data item specifies how the intensity scaling is defined. Apart from linear scaling, which is specified by the value linear, two other methods are available to specify the scaling.

One is to refer to the detector system, and then knowledge of the manufacturers method will either be known or not by a program. This has the advantage that any system can be easily accommodated, but requires external knowledge of the scaling system.

The recommended alternative is to define a number of standard intensity linearity scaling methods, with additional data items when needed. A number of standard methods are defined by _array_intensities.linearity values: offset, scaling_offset, sqrt_scaled, and logarithmic_scaled. The "offset" methods require the data item _array_intensities.offset to be defined, and the "scaling" methods require the data item _array_intensities.scaling to be defined. The above scaling methods allow the element values to be converted to a linear scale, but do not necessarily relate the linear intensities to physical units. When appropriate the data item _array_intensities.gain can be defined. Dividing the linearized intensities by the value of _array_intensities.gain should produce counts. Two special optional data flag values may be defined which both refer to the values of the "raw" stored intensities in the file (after decompression if necessary), and not to the linearized scaled values. _array_intensities.undefined_value specifies a value which indicates that the element value is not known. This may be due to data missing e.g. a circular image stored in a square array, or where the data values are flagged as missing e.g. behind a beam-stop. _array_intensities.overload_value indicates the intensity value at which and above, values are considered unreliable. This is usually due to saturation.

"Array_intensities" Example

To define the characteristics of image_1 as linear with a gain of 1.2, and an undefined value of 0, and a saturated (overloaded) value of 65535, the following header section might be used:
# Define image intensity scaling
loop_
_array_intensities.array_id
_array_intensities.binary_id
_array_intensities.linearity
_array_intensities.gain
_array_intensities.undefined_value
_array_intensities.overload_value
image_1    1    linear   1.2    0   65535

DATA COMPRESSION

One of the primary aims of imgCIF / CBF is to allow efficient storage, and efficient reading and writing of data, so data compression is of great interest. Despite the extra CPU over-heads it can very often be faster to compress data prior to storage, as much smaller amounts of data need to be written to disk, and disk I/O is relatively slow. However, optimum data compression can result in complicated algorithms, and be highly data specific.

In CBFlib version 0.1, Paul Ellis has coded two lossless compression algorithms: canonical and packed.

Canonical-code compression

The canonical-code compression scheme encodes errors in two ways: directly or indirectly. Errors are coded directly using a symbol corresponding to the error value. Errors are coded indirectly using a symbol for the number of bits in the (signed) error, followed by the error iteslf.

At the start of the compression, CBFLIB constructs a table containing a set of symbols, one for each of the 2^n direct codes from -(2^(n-1)) .. 2^(n-1) -1, one for a stop code, and one for each of the maxbits -n indirect codes, where n is chosen at compress time and maxbits is the maximum number of bits in an error. CBFLIB then assigns to each symbol a bit-code, using a shorter bit code for the more common symbols and a longer bit code for the less common symbols. The bit-code lengths are calculated using a Huffman-type algorithm, and the actual bit-codes are constructed using the canonical-code algorithm described by Moffat, et al. (International Journal of High Speed Electronics and Systems, Vol 8, No 1 (1997) 179-231).

The structure of the compressed data is:

Byte Value
1 .. 8 Number of elements (64-bit little-endian number)
9 .. 16 Minimum element
17 .. 24 Maximum element
25 .. 32 Repeat length (currently unused)
33 Number of bits directly coded, n
34 Maximum number of bits encoded, maxbits
35 .. 35+2^n-1 Number of bits in each direct code
35+2^n Number of bits in the stop code
35+2^n+1 .. 35+2^n+maxbits-n Number of bits in each indirect code
35+2^n+maxbits-n+1 .. Coded data

CCP4-style compression

The CCP4-style compression writes the errors in blocks . Each block begins with a 6-bit code. The number of errors in the block is 2^n, where n is the value in bits 0 .. 2. Bits 3 .. 5 encode the number of bits in each error:

Value in
bits 3 .. 5
Number of bits
in each error
0 0
1 4
2 5
3 6
4 7
5 8
6 16
7 65


The structure of the compressed data is:
Byte Value
1 .. 8 Number of elements (64-bit little-endian number)
9 .. 16 Minumum element (currently unused)

17 .. 24 Maximum element (currently unused)

25 .. 32 Repeat length (used, starting with version 0.2)

33 .. Coded data

Additional Compression Schemes

In addition, Andy Hammersley has proposed two types of lossless data compression algorithms for CBF version 1.0. In later versions other types including lossy algorithms may be added.

The first algorithm is referred to as byte_offsets and has been chosen for the following characteristics: it is very simple, may be easily implemented, and can easily lead to faster reading and writing to hard disk as the arithmetic complication is very small. This algorithm can never achieve better than a factor of two compression relative to 16-bit raw data, but for most diffraction data the compression will indeed be very close to a factor 2.

The second algorithm is referred to as predictor_huffman and has been chosen as it can achieve close to optimum compression on typical diffraction patterns, with a relatively fast algorithm, whilst avoiding patent problems and licensing fees. This will typically provide a compression ratio between 2.5 and 3 on well exposed diffraction images, and will achieve greater ratios on more weakly exposed data e.g. 4 - 5 on "thin phi-slicing" images. Normally, this would be a two pass algorithm; 1st pass to define symbol probabilities; second pass to entropy encode the data symbols. However, the Huffman algorithm makes it possible to use a fixed table of symbol codes, so faster single pass compression may be implemented with a small loss in compression ratio. With very fast cpus this approach may provide faster hard disk reading and writing than the "byte_offsets" algorithm owing to the smaller amounts of data to be stored.

There are practical disadvantages to data compression: the value of a particular element cannot be obtained without calculating the values of all previous elements, and there is no simple relationship between element position and stored bytes. If generally the whole array is required this disadvantage does not apply. These disadvantages can be reduced by compressing separately different regions of the arrays, which is an approach available in TIFF, but this adds to the complexity reading and writing images.

For simple predictor algorithms such as the byte_offsets algorithm a simple alternative is an optional data item, which defines a look-up table of element addresses, values, and byte positions within the compressed data, and it is suggested that this approach is followed.

THE "BYTE_OFFSETS" ALGORITHM

The byte_offsets algorithm will typically result in close to a factor of two reduction in data storage size relative to typical 2-byte diffraction images. It should give similar gains in disk I/O and network transfer. It also has the advantage that integer values up to 32 bits (31 bits unsigned) may be stored efficiently without the need for special over-load tables. It is a fixed algorithm which does not need to calculate any image statistics, so is fast.

The algorithm works because of the following property of almost all diffraction data and much other image data: The value of one element tends to be close to the value of the adjacent elements, and the vast majority of the differences use little of the full dynamic range. However, noise in experimental data means that run-length encoding is not useful (unless the image is separated into different bit-planes). If a variable length code is used to store the differences, with the number of bits used being inversely proportional to the probability of occurrence, then compression ratios of 2.5 to 3.0 may be achieved. However, the optimum encoding becomes dependent of the exact properties of the image, and in particular on the noise. Here a lower compression ratio is achieved, but the resulting algorithm is much simpler and more robust.

The byte_offsets algorithm is the following:

  1. The first element of the array is stored as a 4-byte signed two's integer regardless of the raw array element type. The byte order for this and all subsequent multi-byte integers is little_endian regardless of the native computer architecture i.e. the first byte is the least significant, and the last byte the most. This value is the first reference value ("previous element") for calculating pixel to pixel differences.

  2. For all elements, including the first element, the value of the previous element is subtracted to produce the difference. For the first element on a line the value to subtract is the value of the first element of the previous line. For the first element of a subsequent image (or plane) the value to subtract is the value of the first element of the previous image (or plane).

  3. If the difference is less than +-127, then one byte is used to store the difference as a signed two's complement integer, otherwise the byte is set to -128 (80 in hex) and if the difference is less than +-32767, then the next two bytes are used to store the difference as a signed two byte two's complement integer, otherwise -32768 (8000 in hex, which will be output as 00 80 in little-endian format) is written into the two bytes and the following 4-bytes store the difference as a full signed 32-bit two's complement integer.

  4. The array element order follows the normal ordering as defined by the _array_structure_list entries index, precedence and direction.

It may be noted that one element value may require up to 7 bytes for storage, however for almost all 16-bit experimental data the vast majority of element values will be within +-127 units of the previous element and so only require 1 byte for storage and a compression factor of close to 2 is achieved.

The PREDICTOR_HUFFMAN ALGORITHM

Section to be added.

OTHER SECTIONS

Other sections will be added.

9.0 REFERENCES

1. S R Hall, F H Allen, and I D Brown, "The Crystallographic Information File (CIF): a New Standard Archive File for Crystallography", Acta Cryst., A47, 655-685 (1991)

10.0 NOTES

(1) A pure ASCII CIF based format has been considered inappropriate given the enormous size of many raw experimental data-sets and the desire for efficient storage, and reading and writing. However, an ASCII format is helpful for debugging software and in understanding what has been written in a CBF when problems arise, and there are other CIF application for which a convenience binary format should be useful (e.g. illustrations in a manuscript).

(2) Some simple method of checking whether the file is a CBF or not is needed. Ideally this would be right at the start of the file. Thus, a program only needs to read in n bytes and should then know immediately if the file is of the right type or not. Andy though this identifier should be some straightforward and clear ASCII string. With the use of binary strings and MIME conventions identification of a CBF versus a CIF is less critical than it was before, but the distinct header as a simple ASCII string is still a good idea for the sake of the most efficient processing of large files.

The underscore character has been used to avoid any ambiguity in the spaces.

(Such an identifier should be long enough that it is highly unlikely to occur randomly, and if it is ASCII text, should be very slightly obscure, again to reduce the chances that it is found accidently. Hence I added the three hashes, but some other form may be equally valid.)

(3) The format should maintain backward compatibility e.g. a version 1.0 file can be read in by a version 1.1, 3.0, etc. program, but to allow future extensions the reverse cannot be guaranteed to be true. However, prior to actual adoption of version 1.0, we are not yet trying to ensure full upwards compatibility, just that the effort to convert won't be unreasonable.


Examples of CBF and imgCIF Files


This page was produced on 23 April 2001
by Herbert J. Bernstein (email: yaya@bernstein-plus-sons.com),
based on the 14 November 1998 and 8 July 1998 versions and the page produced by Andy Hammersley (E-mail: hammersley@esrf.fr).


Subpages (1): examples
Comments