I'm definitely not certain, but I'm beginning to wonder if it could be a driver issue with interpretation of the EDID. The EDID design has a section of supported timings called Standard Timings (and another called Established Timings, but let's ignore that one). The 1920x1280 resolution doesn't appear in the Standard Timings. It only appears in a different section called the Detailed Timings Descriptors.
I found a comment on EDID at Dr. HDMI's site (http://www.drhdmi.eu/dictionary/edid.html) that states: "Some graphics card drivers have historically coped poorly with the EDID, using only its standard timing descriptors rather than its Detailed Timing Descriptors (DTDs)." That certainly seems to match the symptom (although, there are probably other situations that can exhibit that same symptom). If that is what is happening here, it occurred to me, with the ability to use a binary EDID file in place of an actual EDID read in Linux, I could synthesize an EDID binary that had the 1920x1280 resolution in the Standard Timings section, so it became a choice with the others for the graphics controller driver.
With that in mind I looked for a way to edit the EDID and found wxedid. This utility helpfully explains the format as you go.
In the Standard Timings, each timing is specified with the horizontal resolution, the vertical refresh rate, and the pixel aspect ratio. The vertical resolution is calculated based on the horizontal resolution and the aspect ratio. Unfortunately, only 2 bits are used for the aspect ratio, and 3:2 isn't one of the choices (16:10, 4:3, 5:4, and 16:9 are the options). This means that 1920x1280 cannot be specified in this group. The closest would be 1920x1200 (16:10) or 1920x1440 (4:3).
So, if this is the issue I'm certainly not going to be able to get it fixed before my project needs to be done.
If I can't get the driver to accept the resolution through some avenue (I"m still holding out hope), then the only other option is to figure out how to modify the software I'm using to adjust the aspect ratio, so at least the scaled 1920x1080 doesn't look stretched vertically.
Please let me know if there are any other options I can try to get the pixel-correct mode in place.
Rick
I found a comment on EDID at Dr. HDMI's site (http://www.drhdmi.eu/dictionary/edid.html) that states: "Some graphics card drivers have historically coped poorly with the EDID, using only its standard timing descriptors rather than its Detailed Timing Descriptors (DTDs)." That certainly seems to match the symptom (although, there are probably other situations that can exhibit that same symptom). If that is what is happening here, it occurred to me, with the ability to use a binary EDID file in place of an actual EDID read in Linux, I could synthesize an EDID binary that had the 1920x1280 resolution in the Standard Timings section, so it became a choice with the others for the graphics controller driver.
With that in mind I looked for a way to edit the EDID and found wxedid. This utility helpfully explains the format as you go.
In the Standard Timings, each timing is specified with the horizontal resolution, the vertical refresh rate, and the pixel aspect ratio. The vertical resolution is calculated based on the horizontal resolution and the aspect ratio. Unfortunately, only 2 bits are used for the aspect ratio, and 3:2 isn't one of the choices (16:10, 4:3, 5:4, and 16:9 are the options). This means that 1920x1280 cannot be specified in this group. The closest would be 1920x1200 (16:10) or 1920x1440 (4:3).
So, if this is the issue I'm certainly not going to be able to get it fixed before my project needs to be done.
If I can't get the driver to accept the resolution through some avenue (I"m still holding out hope), then the only other option is to figure out how to modify the software I'm using to adjust the aspect ratio, so at least the scaled 1920x1080 doesn't look stretched vertically.
Please let me know if there are any other options I can try to get the pixel-correct mode in place.
Rick
Statistics: Posted by rtillery — Sun Mar 09, 2025 7:15 am