I have been using the brilliant credit card calibration for years and it’s been working fine. However, I made the (possibly ill-advised) decision to finally upgrade to the 2024 version of Psychopy, which has been creating problems.
One problem is with the credit card calibration. When I use the in-build Screen_Scale routine, which is virtually identical to the screen-scale routine I have been using for years, it works fine on my computer, but doesn’t work online. It doesn’t generate any error messages, but there’s seemingly no response to the arrow keys.
After a bit of playing around with the code, increasing the size of the step change during the calibration (i.e. increasing dBase value), it’s clear that it does work, but the response to the arrow key is just very very very slow.
I have tried going down to older versions of PsychoPy, all to no avail.
Is this a known issue with the newer versions of PsychoPy? Is there some hidden setting that I need to enable or disable?
I’m reluctant to change my Screen Scale code at this point (which uses version 2023.1.3 but is also available in 2020.2)
However, I also use it in Blind Spot which I’ve just updated to 2024.2.2
Please could you check here: screen_scale [PsychoPy]
and let me know whether you have any issues with the arrow keys?
I could easily update it to use a different method for looking for key presses but I’d like to see whether the current method works first (it works on Chrome in Windows for me).
Thank you for responding - this issue has been absolutely doing my head in.
Unfortunately the linked experiment has the same issues for me on Edge (our university forces us to use Edge browser) - I tried the linked experiment in Chrome and Firefox and there it works fine, but many of our students will similarly be using Edge, due to University policies.
I had already noticed that the screen scale code used 2023 version of PsychoPy and consequently in the PsychoPy 2024 version I had set the “use PsychoPy version” setting to 2023.2.3. However, I don’t know how good this version emulator is.
I suppose I could go back to my original PsychoPy version where everything worked (2021.2.3), but my new study needs the better TextBox functionality of newer versions of PsychoPy and also includes the presentation of a grid of objects, for which I had been using your points-on-a-grid code, which also doesn’t work in my old PsychoPy version.
I tried to drag with the mouse version, i.e. holding mouse button and then dragging, which doesn’t really work (large changes in credit card pic size after a delay and inconsistent responses to further attempts at adjustment). However, if I click with the mouse button on location on screen, it will change the size so that nearest corner is on the clicked location. But here still behaviour is inconsistent, with mouse click sometimes eliciting response, and sometimes not.
Bit difficult to explain in words, sorry. Basically, it’s not so much that it doesn’t work at all, it just very inconsistently and somewhat unpredictably responds to mouse clicks or drags.
Slightly off topic - I really like the next verification square - I use a verification line in my own exp, where participants have to measure it and enter its length (with credit card calibration restarting if entered length is too far from true length, suggesting incorrect or ignored credit card calibration).
Hmmm… shouldn’t really be the case to be honest (have a fairly up-to-date system), and I didn’t have issues with older versions of PsychoPy. Has PsychoPy become that much more resource-hungry?
Graphics card is Intel(R) Iris(R) Xe Graphics, which supports OpenGl 4.6, comfortably exceeding the minimum requirement for PsychoPy.
I would hazard a guess that many students and potential participants will be using older devices…
Will try to access the studies on my personal laptop at home tonight, as that laptop is definitely not slow and/or overloaded & report back.
EDIT: Hold on - you may be on to something - just looked at my Edge settings and enabled graphics acceleration - the mouse version seems to work perfectly now, as does the version using the arrow keys. Trust me, am vigorously slapping my forehead. So sorry for taking up your time with something that turned out to be an absolutely silly thing on my end.
This may be a step too far, but is it in any way possible to force the enabling of graphics acceleration using code in PsychoPy? Just to make sure that for participants this is automatically enabled.
I don’t believe this is ever going to be possible in a browser, no. The browser would/should not allow a web page to change the settings of a browser or computer (think of the mischief that malicious web developers could get up to).
Hopefully the code we use to detect that this was off is still working (even that isn’t guaranteed - browsers often try to hide what they do/don’t support) so you could instead include a check at the start of the task asking people to adjust their browser settings. You should have a column in your data file saying that the hardware acceleration was disabled (since PsychoJS 2023.1). PsychoJS would also have tried to alert the user with a dialog that says “It appears that hardware acceleration is either not supported by your browser or currently switched off…” but maybe the user also has popups turned off so didn’t see it.
I supposed you could go further and add message into the study itself rather than a dialog so that it can’t be blocked
I have been using PsychoPy version 2024.2.1 (installed using installer), and did not have a column in my data file reporting on hardware acceleration and did not have a pop-up in study on Pavlovia (but have adblock, so pop-up likely disabled). However, looking at the JS code associated with my study, the above code is not included in the JS file (not sure however, whether this would be in the experiment JS file or in one of the separate PsychoJsS files)
I did however find this online: psychojs/docs/core_PsychoJS.js.html at main · psychopy/psychojs · GitHub which seems identical to the code you posted, so will try putting this code explicitly in my study at the beginning (already have various checks for e.g. mobile devices etc. so should be able to add it there).
