What’s new in X-Ways Forensics v21.1

Volume Snapshot and File System Support

  • Better support for very large volume snapshots. More than 500 million items (i.e. files+directories) in a single volume are now possible, only in the 64-bit edition, subject to sufficient RAM, and assuming you have enough time to wait for the completion of the volume snapshot. This tested capability should add to the notion of X-Ways Forensics as a heavy lifter for storage device analysis. The ~500 million mark assumes an average filename length of 16 characters. With shorter filenames theoretically 1 billion items or more are possible. If only the space for filenames is exhausted, more files can still be included in the volume snapshot, but they will be shown with a dummy filename (a question mark character).
  • Volume snapshots of extraordinarily huge volumes now support files that are defined in the file system at offsets beyond 131 TB or have their data starting more than 131 TB into such a volume. The new limit is 262 TB.
  • Slightly accelerated volume snapshot creation for large NTFS file systems.
  • Two kinds of proactive filters, based on names and timestamps, can now be activated in the properties of a case. Proactive filters allow you to restrict the initial volume snapshot. Files that don’t pass these filters will not be included in any volume snapshot that is taken while such filters are active. Directories are still included. This pertains only to partitions/volumes and file archives that are evidence objects, and all the files that are found in them directly, following the defining data structures of the file system or the archive. It does not restrict the addition of files that are found in any other way, for example by a file header signature search or when checking files that are already contained in a volume snapshot for embedded data etc.
  • Proactive filters are special in that they can prevent files from involuntarily getting into a volume snapshot, files that you do not need or want to be there or that you are not supposed to see. Either if your task or search scope is limited to specific files whose names or timestamp ranges are known beforehand or if the evidence object (image or file archive) is so big that by avoiding hundreds of millions of other files you save time and main memory or can make the volume digestible at all (i.e. keep the volume snapshot size within the supported boundaries). The creation of the volume snapshot itself may be noticeably accelerated that way if the evidence object is an image file, plus all subsequent steps (navigating, listing, sorting, filtering, volume snapshot refinement) are less computationally expensive if you proactively prevent the inclusion of large numbers of unwanted files.

    A count of how many files are proactively omitted during the creation of the volume snapshot is displayed in the progress indicator window. After completion, the total number of such files can always be checked in the status of the volume snapshot in the dialog window for volume snapshot refinement. A warning that a proactive filter is active is output in the Messages window once per session, when a volume snapshot is taken.

  • Directory listings obtained from the operating system (“OS dir list”), which you get for example when adding a directory or a single file to a case as an evidence object, can now be made to not show any timestamps from the file system or only the modification timestamp. That is a volume snapshot option and useful if the timestamps of the files do not have the usual significance, e.g. if they reflect when you collected the files and not what timestamps they had originally at their original location.
  • In new installations the default setting of the volume snapshot option “Newly identified names as main names” is now half selected, which means only for original .eml files newly identified names (i.e. the subject lines) become the main name in the Name column, and the potentially unhelpful generic filenames according to the file system become the secondary names.
  • Ability to recognize SquashFS compressed file systems as such and treat their data like file archives. The supported compression algorithms for SquashFS in X-Ways Forensics are GZIP/zlib, LZMA, LZO and XZ.

Hash Database Support

  • Normal use of the hash database for reading purposes (to retrieve the names of matching hash set for display in the “Hash set” column of the directory browser), if it’s shared, no longer prevents other users from updating the database or replacing the database (i.e. the directory) because the hash set names will be kept in a local cache/buffer.
  • The rather simple CRC32 algorithm, is now supported in ordinary hash databases. Creating a hash database  based on CRC32 is useful (only) if you really only know the CRC32 values of files that you are looking for, no more advanced hash values and not the full original file contents, for example from encrypted zip archives as such archives have the CRC32 values of the unencrypted data in the metadata. If you find CRC32 matches and the file size is the same as known from the metadata in such an encrypted zip archive, then it is very likely that you have found an unencrypted copy of the very same file.

    If you wish to import CRC32 hash values from a text file (with “CRC32” in the first line, followed by one CRC32 value in hex ASCII per line), please note that their hex ASCII values are expected in big-endian (“human-readable”) byte order, as displayed in software like 7-Zip and WinZip and also X-Ways Forensics itself, which unlike MD5, SHA-1 etc. is not the byte order in which they are stored in binary, in X-Ways Forensics internally as well as in zip files themselves and presumably elsewhere.

  • Option to define the block size for block hash databases. 512 bytes is still the default and recommended unless you are certain of what you are doing. A larger block size of 4 KB for example can be compatible with volumes/partitions that have a cluster size of 4 KB and hard disks with a sector size of 4 KB physically and logically, but thwarts any attempt to find the data that you are looking if the clusters in the target file system are not aligned at 4 KB boundaries themselves from the point of view of the evidence object. The latter may be the case for example because the file system has an irregularly sized header area before the first cluster (like FAT) or because you apply the block-wise hashing (only) at the level of a partitionable storage device in which the partitions are not aligned at a 4 KB boundary. The good news, however, is that, just like the file header signature search, block-wise hashing is applied specifically to partitions if partitions are known on a partitionable storage device (or image thereof), and only the area outside of known and explorable partitions is processed at the level of the partitionable storage device.
  • Block hash matches are now displayed with their sizes in the search hit column.
  • PhotoDNA matches (notably multiple matches for the same picture) can now optionally be output as labels. This is useful if you need to see all matches and/or if you wish to see PhotoDNA matches in the same place as ordinary hash database matches, which can also be output as labels.

User Interface

  • You can change the order of labels in either the dialog window for label management or the filter dialog, if labels in that dialog window are not sorted by name, using the arrow buttons. Changing the order there now has an immediate effect on the order in which labels are listed in the Labels column. That way you can make sure that the labels that are most important to you are listed first.
  • Label names in the Labels column can now optionally be truncated, so that more label names fit into the cells of the directory browser. This is a notation setting. Half-checked means that truncations are marked with an ellipsis.
  • Reorganized and tidied up the extended dialog window for labeling.
  • The option “dynamic e-mail and date columns” now properly controls visibility of the “Content created” column.
  • Date filter setting to focus on files of which certain timestamps are not know at all (usually because they were not set e.g. in the file system).
  • More consistent and thorough error and plausibility checks of user-provided file masks.
  • Option to potentially improve synchronization of multiple gallery threads.
  • If you need to call external programs from within X-Ways Forensics with certain parameters in addition to the name of the file that they should open, you can now specify those parameters in the same line of Programs.txt, delimited from the path of the executable file with a tab. The name of the file will be appended at the end, after your own parameters, unless you include the placeholder %1 anywhere in your list of parameters. That placeholder will be replaced with the filename.
  • To associate a portable installation of X-Ways Forensics or X-Ways Investigator and its icon with .xfc case files on a particular machine, you could consciously run the application at least once explicitly as administrator and end it while any of the customizable standard paths is located on the same drive letter as your Windows installation, to give the application a hint that you are the owner of that Windows system and feel comfortable that data is written to it. That’s either the path from where you run the application, the path where to create and expect case files, the path where to create and expect image files, or the path where to create temporary files.
  • Revised compression / data density chart for .e01 evidence files. Among other improvements, the chart window now scales with the user DPI settings.
  • If Preview mode is combined with Details mode, and the lower half of the data window is moved to the right-hand side, preview and details are now split vertically instead of horizontally, with the preview appearing above the details.
  • The description cell in Details mode is now always quite detailed regardless of the notation settings for the Description column.
  • The video files from which to extract still images are now targeted with a comma-delimited type list instead of a filename mask.
  • Ukrainian and Russian translations of the user interface updated.

Picture Content Analysis

  • A new Excire version is available for download now, and required for use with X-Ways Forensics 21.1. The search for “similar” pictures was revised, and the accuracy of content detection has been improved. The number of pictures that get more than one wrong keyword assigned (false positives) has been reduced by 75%. The number of pictures with no wrong keyword has been doubled.
  • The new Excire version has dropped 69 keywords from its detection capabilities that yielded less reliable results. None of these keywords are very important. Support for 87 new keywords was added, including one that was previously requested from law enforcement/government agencies (identification documents), plus various body parts (i.e. not complete people).

File Format Support

  • Report HTML files can now be generated automatically for the Windows Registry hive files NTUSER.DAT, SYSTEM, SOFTWARE, SECURITY, and SAM as part of metadata extraction, based on “Reg Report *.txt” definition files that you have in your installation directory (a number of which are preinstalled). The HTML files are added to the volume snapshot as child objects. The benefit is that they can serve as human-readable previews of selected interesting values, and they contain some encoded text in plain text such as UserAssist entries, so that the logical search can find them. Lots of timestamps from the processed registry hives will be added to the event list at the same time. This all happens if the user also chose to generate HTML previews for browser databases etc. and/or to populate the event list with internal timestamps in files.
  • Checks certain temporary files of MS Edge for embedded pictures automatically as part of the “File header signature search in files not processed above” procedure. The file mask for this procedure is reset in this release for that purpose.
  • Extracts Microsoft Teams messages stored in certain PST archives that were exported via the Admin Center of Microsoft 365.
  • Ability to extract e-mail messages from OLM databases of Microsoft Outlook for Mac.
  • Extracts plain text attachments from original .eml files and MBOX e-mail archives as child objects.
  • Ability to decode .json files for logical searches, indexing, and Text Preview mode, including files with specially encoded Unicode characters from the Basic Multilingual Plane (e.g. Chinese).
  • Metadata extraction from WEBP files extended. In particular, output of Exif metadata in WEBP pictures in addition to XMP metadata has been introduced.
  • Support for some more TIFF picture variants with the internal graphics display library.
  • “Social media” used to be one of multiple possible values of the so-called processing state in the Summary table of JPEG files in Details mode. This origin of photos is now brought to the user’s attention via the so-called software class.
  • 28 software classes are currently supported for JPEG and WEBP pictures: AI generated, Adobe, Amazon (for photos from their shopping web site), Android, Apple, Beautifier, Bing, Camera, ContentGeneral, Editor, Facebook/Instagram, Firmware, General, Google/Picasa, LinkedIn, MSN, PHP, Pinterest, Scanner, Screenshot, Misc social media, Stock (in the sense of stock photos), Twitter (X), Video still, Website builder, WhatsApp, Windows, WordPress.
  • About 75% of all JPEG and PNG (plus some WEBP) pictures now get a software class assigned.
  • More definitions of photo generating devices. In particular the Galaxy S23 and S24 generator signatures were updated.
  • The Summary table was revised.
  • Various special properties that are detected in pictures are now referenced in Details mode with “remark” numbers. A new text file “Remarks.txt” is included, which documents those numbers and may offer a rudimentary explanation.
  • Improved output of metadata for ICC color profiles
  • The output of QWORD values in the registry viewer was previously only for 32 bits. It now covers the complete 64 bits.

X-Tension API

  • The X-Tension API got two additional functions: XWF_Mount() and XWF_Unmount(). If your X-Tensions need to give external programs read access to many or large files in a volume snapshot, it may be faster to mount the volume snapshot as a drive letter than to copy those files to a path that is accessible to those external programs.
  • When X-Tensions add directories to a case as an evidence object, they can choose to have X-Ways Forensics ignore any of the four regular timestamps of NTFS, to prevent their inclusion in the volume snapshot if they are of no value.
  • The X-Tension API has spawned a notable bundle of commercial 3rd-party modules called Exponent™ that integrate very well in X-Ways Forensics and significantly extend its functionality in particular with access to acquired smartphone data and mailboxes. Exponent is available for purchase directly from X-Ways. For details please see below!

Miscellaneous

  • The “Capture Processes” command for Windows live systems was revised. The ability to take window screenshots of various applications, especially Internet browsers and certain Microsoft applications, was considerably improved. Also, users now have some more control over what information is included in the tab-delimited list of windows, e.g. comprehensive lists of child windows and (also new) hash values of screenshots.
  • When interpreting a file as a raw image that does not have a multiple of the presumed sector size as the file size, the extra data at the end that doesn’t add to another full sector is now included, unlike in previous versions, which affects hash computation and potentially file carving. You will still get a warning about the unexpected file size when interpreting such an image, unless you have suppressed it for an evidence object. You may also get read error messages when operations that are applied sector-wise try to read the last (incomplete) “sector”.
  • error.log file entries are now stored in UTF-8 instead of the ANSI code page active in Windows.
  • Improved error message when encountering non-standard internal timestamps in .e01 evidence files.
  • The program help and the user manual were updated.
  • Many minor improvements.