People still buy it because it's digital so it must be better.
I have dealt with a pregnancy test only once in my life, and that was in acting as an interpreter (me: English, wife: Chinese and some English, couple: Chinese, very little English. Not easy!) Reading the instructions is one thing, but figuring out whether the result was positive was another matter. (Although I suspect it was an edge case--the woman's period showed up soon thereafter, I suspect reality was a very early miscarriage and the result was about at the dividing line.)
The number of broken HDMI cables (as fraction of cables rented out) is way bigger than for any other connector, suggesting it is completely unsuitable and a broken design.
Whenever I can I go for SDI video, I do. You plug it in and it works. Why "consumer" techology has to be so much more pain than pro stuff makes me wonder.
That's entirely the fault of your crappy smart display with some crappy OS and has entirely nothing to do with HDMI as a standard.
I would think as a plug and play standard for A/V stuff, HDMI is one of the farthest along the "just works" spectrum for the vast majority of people. Occasionally I see a device where there's something stupid like switching to a different HDMI source doesn't switch the audio source and you have to use some dumb OSD menu with many nested levels to get to the audio sources, but again, that's not HDMI's fault.
I have had quite a few broken HDMI cables in lecture halls at uni and in meeting rooms at various work places, but I think that's the reality of any connector that gets plugged and unplugged tens of times per day (especially by people who don't care and don't have to pay for them when they break). They just need to replace the cables more often.
Could I interest you in all the new features you could enable by instead tunneling video over HDMI Ethernet Channel?
I ruled out the cable, display and laptop by swapping components one at a time.
Sure yeah, but I don't buy it. If you create a standard that is too complicated or too feature-creeped to be implemented fully and that lack of full implementation means the fundamental role of the standard breaks down, that standard might be part of the problem.
I too could envision a solution that theoretically works perfectly, and all people are doing it wrong if it doesn't. But such standards have to be made with reality in mind. USB-C is another one of those. Cool – now I have a ton of USB-C cables none of which tell me on the cable what capabilities they have. One can't support USB-power delivery, the other doesn't work with video up to certain resolutions, etc.
I get that more data means higher frequency and that this directly translates to more problems, but nobody (at least no consumer) asked for the complexity of the HDMI spec. We want to connect a cable and see the picture in 99.99% of the cases. If that doesn't work 100% of the times the standard is at fault. The base functionality of the thing needs to be so dumb and so clear that it just works, even if the other side doesn't even know what an EDID is. That was the task and the result is catastrophic failure.
It's not that the cables support varying specs (which I actually have no problem with--you shouldn't have to pay for features you don't need, and some features trade off vs cable length), but that we have no easy way to find them out or test them.
I get that HDMI is higher frequency and smaller faults show earlier, but the plug is just inadequate. The plugs are levered off by the stiff cable, the thickness of the cables would require at least something like a Neutrik-D-norm connector, but they do as if something smaller is ok. By this point I am just glad that the receiving side seems to be sturdier 90% of the times, but by this point I also wonder why the heck we don't just use BNC connectors and coaxial cables..
Grrr...
But there is a reason nobody puts analogical signals in cables anymore. Beyond some bandwidth, the only way to keep cables reasonably priced and thin is to have software error correction.
HDMI is just that - it's the direct evolution of VGA signaling, with each color channel pushing pixels left-to-right top-to-bottom, it even has blanking periods (periods where there's no pixel info transmitted, used to steer back the electron beam on CRTs to the start of the row/column), same EDID format negotiation over I2C, the works.
What makes it crap is the absolute flood of cheap garbage HDMI cables/repeaters/KVMs which barely work even at the best of times and shouldn't be even allowed to be solved, as they are out of spec, but online vendors have flooded their stores with this cheap no-name garbage for some reason.
Unfortunately, the apparent build quality of the cable, or the price mean nothing when trying to get a working one.
HDMI tries to be a video link, an audio hub, a remote‑control bus, and a content‑police checkpoint all at once. Strip out the DRM, kill the optional‑but‑mandatory feature soup, and let the cable do its one job: move bits from A to B. I had Apple laptops not working with 3-digit Pro A/V gear from reputable vendors because HDCP. This is fucking bullshit. By this point I am starting to consider analog video superior to whatever this is supposed to be.
Um, you just hit why HDMI sucks. You have a "default broken" state that is required by the standard.
Look, every single interface could have been an evolution of Ethernet (and mostly ARE--HDMI and USB-C are basically enshittified Ethernet). But they weren't because everybody wants to put their fingers in the pie and take out a chunk for profit by being a rent-seeking middleman.
It is not only HDMI; many other common things are too complicated (than they should be) and often have other problems too (you mention some of them, such as HDCP, but there is many more problem than that).
(Someone else wrote: "The sad fact of the matter is that people play politics with standards to gain commercial advantage, and the result is that end users suffer the consequences. This is the case with character encoding for computer systems, and it is even more the case with HDTV.")
From the consumer perspective, two cables are more complicated.
Recently I dusted off this project and bought myself a new FPGA board, with a HDMI output. Getting that thing to work was a nightmare. Not only did I have to configure the HDMI transciever, I had to build proper EDID support, and get I2C to cooperate, and get the timing of the signals exactly right. And even then, out of the 5 screens I could get my hands on, only 2 worked, one of which would inexplicably go to sleep despite displaying video on the screen.
Getting HDMI working was more painful and took about 10x as long as the rest of the original project did, and even then, it was a partial success.
The interesting thing about it was that we had very little RAM but the FPGA was quite fast, meaning instead of a framebuffer we scanned out tiles and sprites to the scanline at realtime, at a relatively high resolution (1024x768).
This combined with the low color depth (the 'DAC' was a resistor ladder per color channel, giving us 2 bits per channel, or 6 bits in total), meant we had that dithered, Transport Tycoon-y look, with tilemaps and animaton, so it looked quite okay by the standards of 2d games.
I think even that is more than it should be. I had made up my own kind of video connection (which is not implemented yet), which only has video (the audio is a separate cable, with balanced analog audio, although the standard arrangement of the connectors is supposed to be defined so that you can clip them together when wanted and unclip them if you want to move them separately), and does not negotiate the resolution (for several reasons).
> Last time I plugged in an HDMI source and the darn "smart" television showed the image for 0.5 seconds before displaying a menu that asks me to press a button on the remote to show the image.
At least a part of that is a problem with the smart TV (which has many other problems, too) rather than with HDMI itself, although HDMI also has many problems.
SDI is better in some ways, but it is not good enough either, I think.