6 ms·

I immediately switched to skeptic mode when I saw no actuation points or normalization for travel distance. All this is showing is that to move further it takes

I immediately switched to skeptic mode when I saw no actuation points or normalization for travel distance. All this is showing is that to move further it takes longer. Duh. Why even bother pointing out the polling rate? That's like setting up runners at random places on the racetrack to measure who's faster. The guy/gal that starts closest to the finish line isn't necessarily faster. I'm using Cherry MX Speed switches, which have a 1.2 mm actuation point and 45 g actuation force. There are no published numbers for the Apple keyboards and the closest I could find was ~0.75 mm distance and ~60 g force. Cherry MX Reds, the most common mechanical switches, have a 2 mm distance and 45 g force. (I mention the actuation force because there is probably a slightly longer delay for higher actuation forces due to deforming the fingertip and motor unit recruitment, both of which are probably minimal.) On top of it, the author seems to be making a statement about speed being the deciding factor for gaming keyboards. If that was the case, they'd all be using ultra-low actuation switches, which they aren't. In most cases, a 'gaming keyboard' is just a mechanical keyboard with aggressive styling, higher quality components (less plastic, braided cable), macro keys, multi-key rollover, USB or audio passthrough, and backlighting. Then he uses 1 membrane gaming keyboard and 1 no name import that I can't even find the manufacturer for as his only gaming keyboards. There are no sample sizes reported, he just does his best to hit two keys at once, and uses a camera for determining when the key press starts. Every step of the process has huge flaws and he doesn't even use a reasonable set of devices. Sorry to sound harsh, but I trust zero of the conclusions/interpretations of the data.

stefco_ 9y ago
> I immediately switched to skeptic mode when I saw no actuation points or normalization for travel distance. All this is showing is that to move further it takes longer. I thought his explanation (that delay from the start of the motion) was a good one. Unless someone is starting with a key partially depressed/near the actuation point (which you certainly could do), the key travel time is going to be a part of the perceived delay. Seems fair to me, given that delay and perception are the point of the post. Now, if you think that gamers are starting with partially-depressed keys and therefore will actually experience less of a keystroke delay, that would be a valid counterargument. > Every step of the process has huge flaws and he doesn't even use a reasonable set of devices. I don't think this is being portrayed as an especially accurate study; the author includes plenty of caveats.
- Phrodo_00 9y ago
  But that's not how people game. When I play FPS on my MX Blue keyboard I don't make my fingers sit above the keys, but into the keys, depressing them almost until they're in the activation point. I'd guess people on linear switches don't do this, but they also probably slam their keys faster. The "gaming" keyboards he used were also pretty shoddy. Now, I'm not necessarily saying gaming keyboards offer better latency, because that's not why I got mine, but his methods seem weird.
  - tscs37 9y ago
    That's rather anecdotal. Counter anecdote: I do not hover my fingers, mostly because I have Cherry Clears, I hammer the key. Most people don't partially depress keys to my experience.
DigitalJack 9y ago
Do they spec the force in newtons? Without mass, g’s don’t really say much except as maybe a relative measure between things of the same mass.
- uxp 9y ago
  'g' is gram-force. It's the weight required to "click" the switch over it's actuation point. 1 Gram is equivalent to 0.980665 centinewton, so for most measurements of keyboard switches, g's and cN's are used interchangeably.
  - DigitalJack 9y ago
    Wow, ok. Never heard of g in that sense. I work in aerospace, so that is probably why.
edejong 9y ago
Your criticism is harsh by pointing out a (debatable) point, key travel, using a certain amount of rationale and suddenly concluding ‘every step’ has flaws. Point out those flaws! Furthermore, the OP is the first one doing real measurements on keyboard latency, so perhaps you should trust the article to hint towards a previous unnoticed problem. Anyways, your comment is totally missing the point and makes generalizations based on one debatable measurement error.
- nasredin 9y ago
  TFA would have been more convincing if the author used consistent capitalization of products. Also his table lacks whether the keyboard is mechanical and the type of switch used. This subject needs more research.

stefco_ 9y ago

> I immediately switched to skeptic mode when I saw no actuation points or normalization for travel distance. All this is showing is that to move further it takes longer. I thought his explanation (that delay from the start of the motion) was a good one. Unless someone is starting with a key partially depressed/near the actuation point (which you certainly could do), the key travel time is going to be a part of the perceived delay. Seems fair to me, given that delay and perception are the point of the post. Now, if you think that gamers are starting with partially-depressed keys and therefore will actually experience less of a keystroke delay, that would be a valid counterargument. > Every step of the process has huge flaws and he doesn't even use a reasonable set of devices. I don't think this is being portrayed as an especially accurate study; the author includes plenty of caveats.

Phrodo_00 9y ago
But that's not how people game. When I play FPS on my MX Blue keyboard I don't make my fingers sit above the keys, but into the keys, depressing them almost until they're in the activation point. I'd guess people on linear switches don't do this, but they also probably slam their keys faster. The "gaming" keyboards he used were also pretty shoddy. Now, I'm not necessarily saying gaming keyboards offer better latency, because that's not why I got mine, but his methods seem weird.
- tscs37 9y ago
  That's rather anecdotal. Counter anecdote: I do not hover my fingers, mostly because I have Cherry Clears, I hammer the key. Most people don't partially depress keys to my experience.

DigitalJack 9y ago

Do they spec the force in newtons? Without mass, g’s don’t really say much except as maybe a relative measure between things of the same mass.

uxp 9y ago
'g' is gram-force. It's the weight required to "click" the switch over it's actuation point. 1 Gram is equivalent to 0.980665 centinewton, so for most measurements of keyboard switches, g's and cN's are used interchangeably.
- DigitalJack 9y ago
  Wow, ok. Never heard of g in that sense. I work in aerospace, so that is probably why.

edejong 9y ago

Your criticism is harsh by pointing out a (debatable) point, key travel, using a certain amount of rationale and suddenly concluding ‘every step’ has flaws. Point out those flaws! Furthermore, the OP is the first one doing real measurements on keyboard latency, so perhaps you should trust the article to hint towards a previous unnoticed problem. Anyways, your comment is totally missing the point and makes generalizations based on one debatable measurement error.

nasredin 9y ago
TFA would have been more convincing if the author used consistent capitalization of products. Also his table lacks whether the keyboard is mechanical and the type of switch used. This subject needs more research.