https://web.archive.org/web/20260622172525/https://www.displ...
> DisplayMate is the Worldwide Leader in Video Diagnostics and the World's most advanced Display Calibration and Optimization Software.
I recall when the first LCD TVs came out and I wanted to get a cheap 1080 one. The problem was that almost every TV said 1080 even if it was just 720 upscaled.
I put a uniform image in BMP on a memory stick - every column being R, then G, then B repeated. As the image was exactly 1080, viewing it in full-screen on a 1080 screen gave a sort of uniform grey color. On an upscaled monitor you could see very visible banding.
This would not have worked had the image been JPG.
I’m sure an LLM could offer an explanation, but I genuinely have never understood why overscan is even a thing on HDMI ports.
Most times I plug my AppleTV or laptop into a TV using HDMI by default it’s cutting off the edge of the image. When is it ever a good thing to cut off the edges of the incoming digital image and why does this seem to be the default behaviour on most TVs?
Edit: I asked an LLM and it says some content includes junk on the edge of images because it was never meant to be visible, so TVs enable it by default to cut down complaints/support requests. Apparently, even though HDMI supports a way to signal to TVs to disable overscan, many TVs ignore it.
I’m in my 40’s and have yet to personally encounter content where overscan should be used.
Maybe there were particularly crappy and prevalent analogue to digital converters in the early days of HDMI, but TV manufactures just never stopped doing it even though it’s almost never an issue any more. So we probably have a situation now where probably everyone has either manually changed their TV from its default or more likely are seeing an over scanned and thus non-pixel aligned image.
Remember that an HDMI port may be an input for another device where overscan is necessary to prevent the “garbage” at the edges of the screen. An example of this would be a DVR that outputs HDMI to your TV but perhaps records video that needs to have overscan set. Or VHS as an input. Those are just examples.
JPGs are lossy encodings; my use-case would definitely not work with JPGs, not matter how high you put the quality.
For testing a display (like displaymate does) you literally want to light up specific pixels with specific colors. You can't do that with JPGs.
Instead, you need to export the JPEG at a resampled resolution that's a multiple of your target resolution, such that each pre-transform source-image pixel gets mapped to its own entirely-independent JPEG color block.
Most obvious (though perhaps not optimal?) approach: nearest-neighbour upscale your image by 8x, and then save as JPEG with 100% quality (which will create 8x8 blocks with 4:4:4 subsampling.)
Yeah, I don't think GP did it that way.
Yeah you are, you go girl.