Jpylyzer 2.2 release candidate
We just released a first release candidate of Jpylyzer 2.2. This is quite an action-packed release, which even includes the addition of two new supported file formats! Below is an overview of the main changes.
High Throughput JPEG 2000
Jpylyzer now supports High Throughput JPEG 2000 (HTJ2K) and its associated JPH file format. Both are defined in Part 15 of the standard, which was first published in 2019. To validate a JPH file, use:
jpylyzer --format jph oj-ht-byte.jph > out-jph.xml
For a raw HTJ2K codestream, use this:
jpylyzer --format jhc oj-ht-byte_causal.jhc > out-jhc.xml
Note that not many encoders support HTJ2K at this stage, and the code has only been tested on a limited number of sample files. These files do not encompass all of HTJ2K’s features. We encourage HTJ2K users to contribute additional files to the Jpylyzer test files repository, and report any unexpected behaviour.
Additional codestream markers
This release also adds support for the Extended Capabilities (CAP) and Profile (PRF) marker segments, which were introduced in the 2019 revision of JPEG 2000 Part 1, and the Corresponding Profile (CPF) marker segment, which is specific to JPEG 2000 Part 15 (High Throughput JPEG 2000).
Reporting and validation of rsiz property
Previous versions of Jpylyzer reported the rsiz property as either a numerical value, or as a text string associated with that value. Changes to the use of the rsiz field since the first editions of the standard made this approach increasingly impractical. So, from version 2.2 onward, Jpylyzer always reports rsiz as a numerical value. In addition, it adds the new capability property, which contains the associated text string (which in most cases describes a defined profile). The coverage of these text strings / profiles is now also up to date with the 2019 version of the JPEG 2000 Part 1 standard, which means that many more profiles are reported. Finally, the validation of rsiz itself has been improved as well, as earlier Jpylyzer versions were overly restrictive on the allowed values.
Reporting of warnings to output file
File-level warnings (e.g. in case of unknown boxes) are now included in a new “warnings” element in Jpylyzer’s output. Previously, these were only written to the terminal’s stderr.
All output and extraction options can now be set directly as Python API parameters, which will be helpful for those who want to integrate Jpylyzer in their own Python projects.
Compression ratio reporting for codestreams
Previously Jpylyzer only reported the image compression ratio for JP2 files, and not for raw codestreams. This has changed, and the compression ratio is now reported for all supported formats (including codestreams).
No Python 2.7 support
From this release onward, Jpylyzer now only works with Python 3. Python 2.7 is no longer supported, as we already announced in 2019 at the launch of Jpylyzer 2.0.
Removed legacy options
--legacyout options have been removed in this release.
In addition to the above changes, this release also includes several bugfixes. There’s also a new XSD schema, and the documentation has been made up to date.
Installation with pip
As the current release candidate is a pre-release, make sure to include the
--pre option if you install it with pip. For example, for a fresh install use:
pip install jpylyzer --pre
To upgrade an existing version of Jpylyzer, use:
pip install jpylyzer --upgrade --pre
Installation from binaries
Alternatively, you can use the binaries that are available here:
As always, the Windows binaries are completely stand-alone, and don’t require Python on your machine. Linux users can use the Debian package (which does require Python).
An updated User Manual for Jpylyzer 2.2 is available here.
Any feedback on this Jpylyzer release candidate is greatly appreciated. Also, don’t hesitate to contact us if anything doesn’t behave as expected (preferrably using the issue tracker).
Thanks are due to Aous Naman (University of New South Wales) and Osamu Watanabe (Takushoku University) for their help and suggestions on High Throughput JPEG 2000, and Michael D. Smith (Wavelet Consulting LLC) for providing one of the HTJ2K test images.