I am trying to run an experiment on pavlovia. After encountering many “runs” that involved a blank white screen, I deleted my project folder from Gitlab and started again with a new project in a new folder (on my desktop - rather than a cloud drive). The new experiment is extremely simple. It presents three screens of text (black background, white text) for one second each. There are no images. There are no responses. There are no custom code components. I can run it just fine locally (the python run). It syncs to pavlovia without issue. It runs on pavlovia (in pilot mode, it produces datafiles that download at the end of a run; in run mode, it logs datafiles). The issue is that when it runs, all I see is a blank white screen. Then the run ends normally.
I have tried running in several browsers (Chrome and Safari) on different computers (both MacOS: Tahoe 26.5.1). I am using the most recent version of Psychopy (standalone 2026.1.3) .
The console log produces a warning for each of my elements - that the size is undefined. I am not new to Psychopy or Python (my JS knowledge is very basic). This is a problem that I have never encountered before. I have deleted all the ‘html’ documents from the folder and re-synced the project several times. Any thoughts on what might be going on? (Thank you in advance).
I’m posting the answer to my own question because I spent the weekend playing with this and figured out how to resolve this error - turns out that it is related to RGB codes.
I made a new toy experiment (in Builder) with one screen that displayed one line of text (“This is some text.”) for one second. There was no code, no response - just that one screen. I left all settings as default. It ran, and showed my stimulus. Then I made the first change to the default settings (the one I ALWAYS change, before I do anything else). In the screen preferences I set the background colour to black. It always shows up in Psychopy colours $(-1.0000, -1.0000, -1.0000), in RBG [0, 0, 0]. I used the colour picker and simply clicked “black” to fill in these values. That broke the display (I tried it with and without the $, with round and square brackets - none of these manipulations did anything). The program still ran but the screen was white and nothing appeared to happen (the output file still generated after 1 sec). The console log gave me the same error as above. On a whim, I then changed the background colour to RGB [1, 1, 1]. It ran, with a background colour of not noticeably different from black and displayed my text. I then went to my text object, the default text colour is “white” (named value). I changed this to Psychopy white $(1.0000, 1.0000, 1.0000). Same error as above. Experiment ‘ran’ with a blank white screen and the same console log error. I then tried several other iterations of text colours (e.g., RBG[254, 254, 254]). Nothing got rid of the console log error. I then changed the text colour back to “white” and it worked as intended. Importantly, I did NOT change the psychopy setting which defaults to psychopy normalized RGB colours to anything other than the default. I then went back to the background (which until this time had been RBG[1, 1, 1]) and changed it to named “black”. This ran fine. I added a new component (a random photo from my photos app), along with the text. All good. I then added a start and an end screen with a ‘space’ response on each. It worked as long as the text was set to “white” (any text colour with an RBG code didn’t show up, even though the background was black). I added a polygon with a “white” border and a blue $(-1.0000, -1.0000, 1.0000) fill. It ran but the fill colour was the same as the background colour (i.e., it looked like the colour was set to ‘none’). It worked when I changed it to named “blue”. I played around with this a bit more and ultimately settled on changing the colour space to hexadecimal. I tried out a bunch of bunch of random hex codes and all of them worked. If this is an error that you experience, try changing the colour codes.
I’m really not sure why or what is causing this issue - except that I can reliably make it happen and reliably make it go away. I’ve always used the RGB space in the past without issue. I’m still not sure if this is a ‘me’ problem, or a Psychopy problem. I’m using the latest version of the Psychopy Standalone (downloaded on Friday, 19 June), Mac OS 26.5.1, Chrome Version 149.0.7827.103 (Official Build) (arm64), although I had the issue on Safari also. One of my computer runs an Apple Silicone M1 chip; one runs the M4Pro. Happy to hear if anyone else has this experience (or if any of the authors have thoughts)…
