Latency: Difference between revisions
Jump to navigation
Jump to search
(Created page with " ==Device latency== <!-- [https://en.wikipedia.org/wiki/Latency_(engineering) Latency] is, well, how late something is. Devices, by default, do their own thing, and do not know each other's timing. If there was no design consideration, then you should assume any one device can deviate from the plan, and its execution of your plan will probably not be better than order of magnitude, 10 milliseconds off from your plan. ===Does that even matter?=== '''If you are...") |
m (→Device latency) |
||
| (2 intermediate revisions by the same user not shown) | |||
| Line 52: | Line 52: | ||
Electronics can do a ''lot'' better than a millisecond, | Electronics can do a ''lot'' better than a millisecond, | ||
but needs | but needs | ||
===Compared to what?=== | |||
The more separate devices you add, the more you have to synchronize, | |||
or at the very least verify. | |||
A lot of this stuff is best-effort at low levels, which is why it can be great, | |||
and is also why verifying can be hard. | |||
How do you even measure how late a USB input was? | |||
You can electronically measure when the button got pressed, | |||
but compare it to what? | |||
"Why is my 120Hz TV faster than my 390Hz monitor?" | |||
Because you confused input latency with refresh interval or even the pixel's response time. | |||
That number you give describes the middle one, suggests an upper limit to the third, and says nothing about the first. | |||
LG C2 120Hz Native Resolution @ Max Hz: 5.3 ms | |||
Acer Nitro 390Hz Native Resolution @ Max Hz: 1.8 ms | |||
--> | |||
==Audio latency== | |||
<!-- | |||
Audio latency is ''now'' usually about the latency involged in ditital audio. | |||
(tape delay was ''intentional'' latency) | |||
Audio latency originates both from the | |||
"different devices aren't necessarily synced" detail, | |||
as well as the fact that the typical solution involves a small buffer. | |||
As such, driving audio latency below a few milliseconds requires | |||
So it's ''not'' that the devices couldn't time the samples -- the timing of the sound itself is accurate to dozens of ''micro''seconds. | |||