I'd be surprised if the codec is a quality problem (except cellular codecs). But you can't make a high-SNR mic unless it has high selectivity. "Whole room" mics are the worst, by physics. One differential mic on each side of each speaker's face (airpods?) has huge potential.
-
-
Replying to @apenwarr @davidcrawshaw and
I would've guessed that video latency is already so high that adding the driver stack in wouldn't make a huge difference, but that might not be true? I sometimes see the latency of calls directly (when someone recursively screen shares and a hall of mirrors slowly develops)
1 reply 0 retweets 1 like -
It seems like I should be able to see the same thing and get the latency with the webcam in the loop with a single mirror, I should test that vs. an external camera to see if there's a noticeable difference
1 reply 0 retweets 1 like -
I tried this and got 23-28 frames of latency (60fps) w/Macbook webcam & 6-13 w/external camera in hangouts. I only tried once (range is from "echos"), wouldn't trust this measurement without repeating it, but it's at least not 100x worse or something.
1 reply 0 retweets 1 like -
Setup was 2 2017 13" Macbook Pros. Recorded using QT local recording. Macbooks pointed at each other, have something blocking the screen which is removed, latency is the time before that's seen, range comes from seeing this show up in progressively smaller windows.
2 replies 0 retweets 0 likes -
Have symmetric gigabit internet connection, both machines connected through the same TP-Link Archer C7, local network idle other than this test (modulo interference from other networks).
1 reply 0 retweets 1 like -
Replying to @danluu @davidcrawshaw and
Gadzooks, half a second of latency with the internal webcam? That’s insane. Is that somehow normal?
1 reply 0 retweets 0 likes -
Replying to @apenwarr @davidcrawshaw and
That's sort of double counting the latency (I think? Haven't really thought this through) since it's an entire round trip before you see the image updated. The change is so large I think it can't be 100% real and some of it is random variation, which I'd want to restest for.
1 reply 0 retweets 0 likes -
The reason I think there's noise is the change should only impact one computer, so half the pipeline, but the latency is cut in half? There's no way adding the camera reduced the latency on one host to zero. Maybe I'm not thinking about this correctly, but this seems odd to me.
2 replies 0 retweets 0 likes -
I mean, if noise dominates the measurement, then it probably can't hurt too much to use an external camera, but I'd be curious what the average difference is and how much is just random noise.
2 replies 0 retweets 0 likes
Wait, it's not impossible that halving the latency is real, the image isn't round tripped through both cams and both displays. The entire pipeline should be: http://A.cam -> A.mac.input -> http://A.to.B.network -> B.mac.output -> B.display ->
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.