By AKrishna - 8/11/2025
Hi all,
I wrote a version of the DSST (Digit Symbol Substitution Test) with a customizable display grid for smartphone use. For ease of use, the user sets the height and position of the DSST reference grid at the top of the script, then the task begins and runs for the user's desired duration. To achieve the dynamic display of the grid, a black rectangle shape is drawn and then partially covered by text stimuli with white backgrounds, so that the black background creates a grid impression.
This has worked well in the past on multiple phone models (e.g., Samsung S10). However, one of my users was testing the script on a newer Samsung Galaxy S24 (SM-S921B/DS, running Android 15) and encountered a display issue where every even number on both the reference grid and the response scale was not displayed:
After she entered her first response, the display issue changed once (and then remained in the changed state after further responses):
While testing, I tried a version where each number stimulus got a unique background color, to see whether there was any overlap or covering issue, which resulted in the following:
I have no clue what's going on. What baffles me most is that the text stimuli are being drawn at their correct sizes, as otherwise you would expect to see no gridlines (they would be covered by the bgcolor of a too-wide text stimulus). Yet completely unrelated stimuli are being covered by the background color of other stimuli (e.g. the textbox for number 1 on the bottom left taking on the blue of textbox 2 instead of its own darkred). Would appreciate any help!
The current version of the script can be found at https://mili2nd.eu/jzjc (note that this is a version with white textboxes, not the colored version shown above).
|
By Dave - 8/11/2025
+xHi all, I wrote a version of the DSST (Digit Symbol Substitution Test) with a customizable display grid for smartphone use. For ease of use, the user sets the height and position of the DSST reference grid at the top of the script, then the task begins and runs for the user's desired duration. To achieve the dynamic display of the grid, a black rectangle shape is drawn and then partially covered by text stimuli with white backgrounds, so that the black background creates a grid impression. This has worked well in the past on multiple phone models (e.g., Samsung S10). However, one of my users was testing the script on a newer Samsung Galaxy S24 (SM-S921B/DS, running Android 15) and encountered a display issue where every even number on both the reference grid and the response scale was not displayed: After she entered her first response, the display issue changed once (and then remained in the changed state after further responses): While testing, I tried a version where each number stimulus got a unique background color, to see whether there was any overlap or covering issue, which resulted in the following: I have no clue what's going on. What baffles me most is that the text stimuli are being drawn at their correct sizes, as otherwise you would expect to see no gridlines (they would be covered by the bgcolor of a too-wide text stimulus). Yet completely unrelated stimuli are being covered by the background color of other stimuli (e.g. the textbox for number 1 on the bottom left taking on the blue of textbox 2 instead of its own darkred). Would appreciate any help! The current version of the script can be found at https://mili2nd.eu/jzjc (note that this is a version with white textboxes, not the colored version shown above).
Not reproducing this on my phones running Android 15, so it's not some general compatibility issue with v6 under Android 15. (For completeness's sake, it works fine under v7 as well.) I don't have an S24 available, though. Do you know which model / variant of the S24 this was? Thre are quite a few (S24, S24+, S24 Ultra, S24 FE), and there are variants with different GPUs as well (Adreno vs Exynos). There are quite a number of reports about the S24 series having buggy graphics drivers and graphics performance issues, so my guess is this is a symptom of that, e.g. laggy driver re-draws last used texture instead of the actual one.
|
By AKrishna - 8/11/2025
Best I can tell, it's a standard Galaxy S24 (not +, FE, or Ultra), dual-sim model. It should be an international model (bought in Germany), so Xclipse 940 graphics, I assume?
The model number is SM-S921B/DS, but I can't for the life of me figure out what that actually means. Thanks, Samsung! :-/
|
By Dave - 8/11/2025
+xBest I can tell, it's a standard Galaxy S24 (not +, FE, or Ultra), dual-sim model. It should be an international model (bought in Germany), so Xclipse 940 graphics, I assume? The model number is SM-S921B/DS, but I can't for the life of me figure out what that actually means. Thanks, Samsung! :-/
Yeah, that should be the basic model with the Exynos SoC (and therefore the Xclipse GPU).
|
By AKrishna - 8/11/2025
OK, thanks for the info. So the most likely hypothesis is a driver issue on the specific phone. Bad luck for now, but at least it's something with a chance to get fixed by Samsung in future. I'll pass it on to the user!
|
|