Video summary
The PsychToolbox technology updates presented at VSS 2024 highlight significant advancements in flexible experiment design, particularly emphasizing its role as an open-source solution optimized for precise timing. While the tool is best utilized on Linux due to direct control over operating system processes that ensure reliable performance, developers have implemented various workarounds and optimizations to maximize compatibility with Windows and macOS. Beyond being a MATLAB toolbox, PsychToolbox also functions effectively with GNU Octave and integrates seamlessly with Python-based frameworks like PsychoPy, supported by a vast community evidenced by citations in over 37,000 articles as of May 2024.
A major breakthrough introduced this year is the new OpenXR extended reality driver, which significantly enhances virtual reality support through an industry-standard multivendor API compatible with most head-mounted displays on Linux and Windows. This update facilitates the use of various VR-tracked input controllers and includes initial eye-tracking capabilities for both monocular setups under any supported runtime and binocular tracking specifically for HTC Vive Pro Eye devices on Windows. Furthermore, substantial improvements have been made to the Eyelink toolbox by researchers at SR Research, enabling gaze tracking in stereoscopic displays with different stimuli per eye, calibration using animated movies, and better audio feedbacks across Linux, Windows, and macOS platforms while maintaining backward compatibility with existing codebases.
The session also addressed practical challenges such as display mirroring setups, which are crucial for environments like MRI scanners where researchers need to monitor stimulus presentation without compromising the visual timing experienced by the subject inside the scanner. These improvements allow experimenters to utilize mirror modes on standard hardware without degrading performance or requiring special equipment. Additionally, movie playback capabilities have been upgraded to efficiently handle demanding formats such as 4K resolution and HDR content via optimized hardware video decoding, enabling parallel playback of multiple high-definition movies at 60 frames per second on suitable graphics cards under both major operating systems.
Finally, the presentation covered necessary adjustments regarding software distribution following GitHub's removal of Subversion frontend support earlier in the year. Thanks to a workaround provided by The MathWorks open-source team, an automated workflow has been established for generating downloadable zip files containing the PsychoPy folder and facilitating installation via MATLAB's built-in add-on browser. Although manual steps are still required for setting up dependencies like GStreamer on Linux systems, these updates ensure that researchers can continue to build robust experiments with improved features while maintaining stability across different computing environments.
Read the full video transcript
okay so I'm going to tell you about um
what's new with psych toolbox in
particular in the past two years um
since um we presented at vsss in
2022 and Mario Kleiner the developer of
Psych Toolbox will be available
afterwards to answer any of your
questions I want to give an overview to
start with of what psych toolbox is it's
an open source toolbox for flexible
experiment design with the most uh
reliable timing um and most advanced
features um currently
available it's optimally used um with
the open source Linux operating system
and that's because its development goes
beyond um the toolbox itself Mario also
um basically optimizes Linux um to uh
give the best possible timing which is
only possible if you have control over
over the operating system um itself as
well this unfortunately isn't possible
on Windows and Macos um but Mario also
um finds workarounds and um and
optimizations to make timing as good as
possible on these operating systems as
well giving you um as many options for
your experiments as
possible like toolbx is a mat lab
toolbox but it also works well on G
octave as a free mat lab alternative for
many common use
cases py tobox has a really um wide
Community it's cited by over 37,000
articles as of May
2024 there are also many tool boxes that
build on top of Psych toolbox um one of
which is psych bench that you'll be
hearing from um in a talk later in this
session P toolbox core routines for
response collection and sound um are
also even instrumental for uh python
based toolkits like psychopy
fre I want to go over many of the
notable new features that have been
added um in the past couple of years the
biggest of which is the new open XR
extended reality driver
that gives improved virtual reality
support via this um industry standard
multivendor open XR API this means you
can use it on most available um
head-mounted displays on Linux and
windows and you can also um use many VR
tracked input
controllers there's also a high
extensibility for future enhancements um
for example in domains of mix reality um
or augmented reality where you can
combine virtual elements with your um
natural visual um um
uh
scene as well uh for um virtual reality
many hmds now come with um an ey tracker
built in and there's initial ey tracking
support for open XR as
well um you've got a basic monocular or
Cyclops eye that combines gaze vectors
for the two eyes into one um under Linux
and windows um with any VR hmd or it
tracker supported by open XR run
times there's also binocular ey tracking
under windows with supported HTC VR hmds
and both have been tested so far with
the HTC VI proi on Windows
10 more for ey tracking there's a
massive update of the ey link toolbox
that's a tool boox for Sr research ey
link gaze trackers on Linux windows and
Macos some of these um updates are gaze
tracking for stereoscopic display setups
where you have different visual stimuli
for each eye use of animated stimuli and
movies for gase track
calibration improved audio feedbacks and
various other improvements um for your
experiments and there are new demos to
show optimal use of uh the new ey like
toolbox that's been contributed mostly
by the team at Sr research so thank you
very much um to them and Mario um
contributed a lot of um improvements to
ensure backwards comp compatibility so
that your uh current code will continue
to work and to optimally integrate with
py
tobox there are also um improvements for
display mirroring setups these are for
setups where you might have a stimulus
display where you want really good
visual timing and um but also a
monitoring display if you imagine for
example a setup in an MRI scanner it's
important that this subject is getting
um the best the stimuli um that are very
well timed but you as an experimenter
that can't see that display because
you're outside of the MRI scanner you
might want to still have an idea what's
being shown um and so this this um
ability gives you the uh option to use
these mirror modes without impairing
visual stimulation timing on your
stimulus display um and also without
needing um special
Hardware there are also improvements to
movie Playback particularly in the
domain of demanding for example 4K
resolution um and HDR movie Playback um
via more efficient use of Hardware video
decoding this allows for example
parallel playback of multiple 4K HDR
movies at 60 frames per second on
suitable graphics cards under Linux and
windows another new feature is a new
method of downloading psych toolbox
unfortunately at the start of this year
gith up removed um subversion front end
supports which disabled our installer
updator and
versioning thankfully a workaround was
contributed by the math Works open
source team and so there's a new um
automated workflow for generating a
downloadable zip file with the py
toolbox folder and also options for
installation via mat Labs buil-in add-on
browser there still needs to be a manual
run of setups like toolbox um and inst
installation of dependencies like
gstreamer