5 ms·

I wonder what the latency is going to be on this. It needs to be at least 1/30th of a second round trip but I suspect the actual latency is going to be more lik

I wonder what the latency is going to be on this. It needs to be at least 1/30th of a second round trip but I suspect the actual latency is going to be more like 1s+

crazygringo 2y ago
I don't know why it would be any different from a Zoom call. In reality it's probably even faster since neither side is dealing with Wi-Fi.
- bagels 2y ago
  It doesn't say what technology is used, does it? Could be using wifi, or cell networks.
  - pclmulqdq 2y ago
    Wired, using the subsea fiber cables, your still stuck at a 60 ms RTT minimum from NYC to Dublin. Add in video transcoding time and a more realistic estimate of the networks on either side, and you're looking at latency closer to 200-400 ms. WiFi over this distance actually isn't a significant contributor to latency.
    - bagels 2y ago
      Where does it say it's using a wired connection? Agree that the latency from the distance is going to be a big factor.
    - oceanplexian 2y ago
      Actually transcoding can be done far below that with consumer hardware. I’m doing Quicksync accelerated transcoding right now for Moonlight and encoding/decoding latency is only a few ms. And Wifi is, by far the greatest source of latency. In fact I can get data to a datacenter a few hundred km away over a fiber connection faster than Wifi across the room (<5ms vs 10-20ms)
      - pclmulqdq 2y ago
        A few hundred km and 5000-7000 km under the ocean are very different. Good to know transcoding can be so fast, though. Dublin is about 10-50 times further from NY than you are from that datacenter, and the speed of light is a universal limit.
trey-jones 2y ago
I wonder the same. I already take issue with the use of "realtime" in the marketing material.
- Chilko 2y ago
  Would you consider zoom call with someone on the other side of an ocean "realtime"? If so, I don't see how this is any different. If not, that's a pretty narrow definition of realtime given the context.
- anyfoo 2y ago
  Realtime means different things in different contexts. If I can watch a webcam of freeway traffic, than 5 seconds delay is, for all intents and purposes, realtime. If we talk about medical equipment, it may be very different based on the equipment. Real time OSs advertise maximum guaranteed latencies. 0 latency does not exist in the physical world, the speed of light can’t be convinced to hurry up, unfortunately.
chgs 2y ago
Why 1/30th? Latency will be far more than that due to speed of light. My mpeg vision circuits are about 480ms end to end from Washington/New York to London. The j2k ones are faster but still in the 200ms range. In theory you could get it down to sub 100ms but at that point you are certainly having to engineer your paths correctly and use barely any compression. Uncompressed 4k is in the 10gbit range. My uhd jxs bitrates are about 1.3gbit but wouldn’t deliver sub 200 across the pond.
- mysecretaccount 2y ago
  > Why 1/30th? > Latency will be far more than that due to speed of light. 1/30th of 1s is almost exactly the round-trip time from NYC to Dublin (Wolfram Alpha says 17ms one-way).
- anyfoo 2y ago
  OP is correct, 1/30 s is a good ballpark theoretical lower bound. Assuming travel across the earth surface, light in a vacuum take about 14ms to cross the distances. Times two for round trip that’s about 1/30 s. Of course we don’t have anywhere near ideal conditions there (at a minimum, light in fiber is already slower, closer to 40ms round trip in that case, and of course network infrastructure adds orders of magnitude), but it’s a good limit.
  - oceanplexian 2y ago
    You don’t need round trip with live video, you can push frames in UDP packets from both sides.
    - IncreasePosts 2y ago
      I was thinking more about the rtt to do something in new york and see the reaction of the person in Dublin, which would probably be the major indicator of latency.
futureshock 2y ago
Imagine a transatlantic Zoom call off your laptop camera and home wifi. Not great. I have seen this Portal in person.

I wonder what the latency is going to be on this. It needs to be at least 1/30th of a second round trip but I suspect the actual latency is going to be more like 1s+

crazygringo 2y ago

I don't know why it would be any different from a Zoom call. In reality it's probably even faster since neither side is dealing with Wi-Fi.

bagels 2y ago
It doesn't say what technology is used, does it? Could be using wifi, or cell networks.
- pclmulqdq 2y ago
  Wired, using the subsea fiber cables, your still stuck at a 60 ms RTT minimum from NYC to Dublin. Add in video transcoding time and a more realistic estimate of the networks on either side, and you're looking at latency closer to 200-400 ms. WiFi over this distance actually isn't a significant contributor to latency.
  - bagels 2y ago
    Where does it say it's using a wired connection? Agree that the latency from the distance is going to be a big factor.
  - oceanplexian 2y ago
    Actually transcoding can be done far below that with consumer hardware. I’m doing Quicksync accelerated transcoding right now for Moonlight and encoding/decoding latency is only a few ms. And Wifi is, by far the greatest source of latency. In fact I can get data to a datacenter a few hundred km away over a fiber connection faster than Wifi across the room (<5ms vs 10-20ms)
    - pclmulqdq 2y ago
      A few hundred km and 5000-7000 km under the ocean are very different. Good to know transcoding can be so fast, though. Dublin is about 10-50 times further from NY than you are from that datacenter, and the speed of light is a universal limit.

trey-jones 2y ago

I wonder the same. I already take issue with the use of "realtime" in the marketing material.

Chilko 2y ago
Would you consider zoom call with someone on the other side of an ocean "realtime"? If so, I don't see how this is any different. If not, that's a pretty narrow definition of realtime given the context.
anyfoo 2y ago
Realtime means different things in different contexts. If I can watch a webcam of freeway traffic, than 5 seconds delay is, for all intents and purposes, realtime. If we talk about medical equipment, it may be very different based on the equipment. Real time OSs advertise maximum guaranteed latencies. 0 latency does not exist in the physical world, the speed of light can’t be convinced to hurry up, unfortunately.

chgs 2y ago

Why 1/30th? Latency will be far more than that due to speed of light. My mpeg vision circuits are about 480ms end to end from Washington/New York to London. The j2k ones are faster but still in the 200ms range. In theory you could get it down to sub 100ms but at that point you are certainly having to engineer your paths correctly and use barely any compression. Uncompressed 4k is in the 10gbit range. My uhd jxs bitrates are about 1.3gbit but wouldn’t deliver sub 200 across the pond.

mysecretaccount 2y ago
> Why 1/30th? > Latency will be far more than that due to speed of light. 1/30th of 1s is almost exactly the round-trip time from NYC to Dublin (Wolfram Alpha says 17ms one-way).
anyfoo 2y ago
OP is correct, 1/30 s is a good ballpark theoretical lower bound. Assuming travel across the earth surface, light in a vacuum take about 14ms to cross the distances. Times two for round trip that’s about 1/30 s. Of course we don’t have anywhere near ideal conditions there (at a minimum, light in fiber is already slower, closer to 40ms round trip in that case, and of course network infrastructure adds orders of magnitude), but it’s a good limit.
- oceanplexian 2y ago
  You don’t need round trip with live video, you can push frames in UDP packets from both sides.
  - IncreasePosts 2y ago
    I was thinking more about the rtt to do something in new york and see the reaction of the person in Dublin, which would probably be the major indicator of latency.

futureshock 2y ago

Imagine a transatlantic Zoom call off your laptop camera and home wifi. Not great. I have seen this Portal in person.