Jap2.0 wrote:cyfrov wrote: response of human eyes is on the order of ~10ms
Response time is different from perception time. This is basically the 24/27/30/60/75/120/144 fps debate - 10ms is 100 fps, and people have refresh rates significantly higher than that.
You are absolutely right, the perception time is even slower, it takes a lot longer to identify and consciously process object information, so 60fps is more than enough for a reaction based game (Halo), and 30fps is enough for a thinking based game (Factorio)
The uber high refresh rates on TVs/Monitors is exclusively a marketing gimmick (I speak based on my background in designing high speed integrated circuits, including HDMI).
The technology has reached a level of saturation and commoditization, and makers are desperate to distinguish themselves from the Chinese no-name brands.
That being said, what the higher refresh rates do is they allow makers to cut corners on other parts of the display system. Simply put, they can provide a cheaper 30Hz effective display, while claiming to be the next best thing since sliced bread.
By employing temporal subsampling, they can reduce the framebuffer memory requirements in the display (save a few $). Alternatively, they can exploit oversampling techniques to maintain a color dynamic range without having to buy high resolution D/A converters (save a few more $). There are other tricks they can use, but these are the two most common.
To add to that, there is also supply chain pressure, especially when making physical things in the 10k+ unit range.
From the chip side, unless you're in the military area, there is a pressure to keep changing silicon process (e.g. 180nm -> 130nm -> 90nm -> 65nm ... 8nm) both internally, and from fabs.
Internally, to avoid price erosion, since a chip that has been in manufacture for over 5 years can see its profit margins shrink from 99% to <15%.
From fabs, since it's not always economical to be maintaining 20-30 year old technology, so they taper down their wafer capacity for a node, and require you to move to a different node if you want certain high volume production.
When faced with that push, unfortunately the change is not simple and straightforward as most code monkeys think, you can't just "recompile" your chip for a new process, you often have to start almost from scratch for the analog circuits. At which point, why not push that interface from 2Gbps to 3Gbps? at least you'll give marketing something to brag about...that's roughly how a lot of the internal conversations go...which then translates to a customer conversation that goes like this: we don't make the old chip in the quantities you want, but we can sell you the new one which let's you do more; Customer: but I don't need that; Sales: hmm, well...we can tell you how to "simplify" your system using our new chip, so that your BOM cost is roughly the same or lower
...so yeah...it's a different story when you look under the hood
