Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

JaaS: The team that builds Jitsi can now also run it for you

jitsi.org

Post reply on HN
Print view View original

81–90 of 120 posts

Re: JaaS: The team that builds Jitsi can now also run it for you

#81

Post by kwindla »

cloogshicer wrote: post #20

Does anybody know a video conferencing tool with low CPU usage while screen sharing? All the tools I've tried, including Jitsi, slow down my computer to a crawl.

As a user, I think Google Meet probably has the most cpu-friendly default settings for screensharing. They limit the framerate to 5fps, and they're careful to use either the native size of the screen or half the native size (which avoids at least some expensive scaling code paths).

This is hard/impossible to get right on every machine, because encoder implementations vary widely across all the hardware and operating system combinations. For example, using h264 instead of vp8 might help with cpu usage a lot on an older macOS machine, but make cpu usage worse on a newer macOS machine or some specific windows machine.

Here are settings we recommend to developers building apps that include screen-sharing using our video APIs: https://gist.github.com/kwindla/5d5a8190aee36dc00a5ef8e6c901...

Max-width of 640 for the camera. Max-width of 1920 for the screenshare, framerate limited to 5fps, and manually decimate the screenshare width by half if it's larger than 1920. ymmv, but these settings use ~30% of the cpu on a typical mid-range Windows or macOS machine, for the sending side of a two-person cam+screenshare call.

Re: JaaS: The team that builds Jitsi can now also run it for you

#82

Post by vharuck »

eznzt wrote: post #63
JacobSuperslav wrote: post #61

Earlier quoted context omitted.

it isn't? i used it for years and been very happy.

The marginal cost of the service they provide is like 10% of what they charge, or even less.

If you believe this, start a competing service. There is no moral or practical rule that forces prices down to cost without external reason. If Fastmail is making big bucks, and customers are happy with the price, that's good for everyone.

Re: JaaS: The team that builds Jitsi can now also run it for you

#83

Post by jefftk »

kwindla wrote: post #75
worble wrote: post #28

Looking for pricing, I get to this page: https://www.8x8.com/products/apis/video >We're charging on a Monthly Active User (MAU) >Pricing starts at $0.35/MAU and decreases based on volume. >An MAU is defined as a unique user who attended at least one meeting, with at least one other user, in the same month. To determine a unique user, we store an identifier on the device's local storage, and that will remain the same…

Pricing for real-time video is hard. There are pretty large variable costs to run a service like this. WebRTC is peer-to-peer, architecturally. But if you run the service you pay for bandwidth for users behind firewalls (because you need to forward the audio/video through TURN servers) and for calls with more than 3 or 4 people (because you need to offload some of the cpu/bandwidth from the clients). Bandwidth and me…

I'm currently using Twilio for echo.jefftk.com, and paying $0.004/minute (I'm very low volume).

I was considering switching to this, but their pricing is a terrible fit for my "most calls have mostly people who've never used it before" situation.

Re: JaaS: The team that builds Jitsi can now also run it for you

#84

Post by kwindla »

mguerville wrote: post #27
cloogshicer wrote: post #20

Does anybody know a video conferencing tool with low CPU usage while screen sharing? All the tools I've tried, including Jitsi, slow down my computer to a crawl.

I’ve had similar issues and have not found a tool that cooperated better. Here are a few things that helped though: - App police (mac app) to reduce cpu allocation for any given app - unplug external display monitor during video calls - turn off any video enhancing settings in app (mirror video, enhanced light, etc) as well as virtual background or background blur It’s crazy that I’m 2021 we don’t have video calling…

> It’s crazy that I’m 2021 we don’t have video calling that isn’t a ginormous performance hog

It sure is. On the other hand, the big tech companies did not view video calls as a critical use case until ~9 months ago.

You can see this in how much Chrome has improved in its last 3 releases. A year ago, you could play ~15 videos simultaneously in a page on Chrome before . Now, on the same machine, you can play >50. This matters for video calls because the standard stream topology these days for multi-party calls is 1 up, N down, where N is the number of participants in the call.

Hardware and OS support is critical, too, so this is an ecosystem challenge.

  - encoder pipelines don't do low-resolution streams very well, which matters for large video calls
  - encoder pipelines don't do variable framerate, variable resolution, large resolution/low-framerate, and "non-standard" resolutions as well as they could
  - hardware support for newer codecs (vp9, h265, av1) would help
  - decoder pipelines aren't great at decoding lots of small videos in parallel
Everyone focused on single-stream video playback for the last few years, which led to really amazing improvements watching video. Nobody complains anymore about their fan spinning up every time they watch Netflix. Now we have to put the same effort into the building blocks needed for video calls.

Re: JaaS: The team that builds Jitsi can now also run it for you

#85

Post by Spare_account »

philipps wrote: post #65
worble wrote: post #28

Looking for pricing, I get to this page: https://www.8x8.com/products/apis/video >We're charging on a Monthly Active User (MAU) >Pricing starts at $0.35/MAU and decreases based on volume. >An MAU is defined as a unique user who attended at least one meeting, with at least one other user, in the same month. To determine a unique user, we store an identifier on the device's local storage, and that will remain the same…

I would still like to see an answer to the questions above. In particular, does a user who connects on three devices count as three MAUs?

If MAUs is hard, then don't use MAU as a billing metric? Charge for concurrent connections or by the minute maybe?

Re: JaaS: The team that builds Jitsi can now also run it for you

#86

Post by aloer »

robertkrahn01 wrote: post #8

CoScreen, an app sharing and pair programming tool [1] runs on Jitsi / JaaS. I work on that project. We're really happy with the service since a reliable and fast global infrastructure is key for what we offer. [1] https://www.coscreen.co

Can you tell us a little bit about the technology you use to make the virtual screen work on MacOS?

Someone recently posted something requiring an “HDMI dummy plug” (a fake display adapter?) to create a virtual screen. How did you solve it on a pure software level?

I would like to understand better what kind of api access your solution requires and I have trouble finding the right words to google.

- I assume you have to use undocumented internal APIs?

- do you have to inject code into higher privileged processes such as the Dock or window manager?

Edit: your page looks broken on mobile Safari with content blockers enabled via Firefox focus. The whole part above fold stays white

Re: JaaS: The team that builds Jitsi can now also run it for you

#87

Post by marcinzm »

eznzt wrote: post #74
tomashubelbauer wrote: post #67

Earlier quoted context omitted.

Do you go to restaurants expecting to pay only the total cost of ingredients for your meal? Fastmail provides an excellent service and if anything I feel as though I underpay for what I'm getting. I'm happy to know they have a healthy margin to pay their developers from and improve the product.

I expect to pay a reasonable price. Being overcharged by 90% is not reasonable.

Restaurants charge 4-5x the cost of ingredients. So I take it you never eat out?

Re: JaaS: The team that builds Jitsi can now also run it for you

#88

Post by eznzt »

marcinzm wrote: post #87
eznzt wrote: post #74

Earlier quoted context omitted.

I expect to pay a reasonable price. Being overcharged by 90% is not reasonable.

Restaurants charge 4-5x the cost of ingredients. So I take it you never eat out?

I have worked in one and no they don't. Most restaurants barely break even by the end of the month once you factor in all costs (like employees, rent...), which I did in the example of fastmail.

Re: JaaS: The team that builds Jitsi can now also run it for you

#89

Post by marcinzm »

eznzt wrote: post #88
marcinzm wrote: post #87

Earlier quoted context omitted.

Restaurants charge 4-5x the cost of ingredients. So I take it you never eat out?

I have worked in one and no they don't. Most restaurants barely break even by the end of the month once you factor in all costs (like employees, rent...), which I did in the example of fastmail.

You said marginal cost which is not all costs. Marginal cost is equivalent to ingredients and does not include fixed costs which likely make up the majority fo their expenses.

Re: JaaS: The team that builds Jitsi can now also run it for you

#90

Post by sudosysgen »

jefftk wrote: post #83
kwindla wrote: post #75

Earlier quoted context omitted.

Pricing for real-time video is hard. There are pretty large variable costs to run a service like this. WebRTC is peer-to-peer, architecturally. But if you run the service you pay for bandwidth for users behind firewalls (because you need to forward the audio/video through TURN servers) and for calls with more than 3 or 4 people (because you need to offload some of the cpu/bandwidth from the clients). Bandwidth and me…

I'm currently using Twilio for echo.jefftk.com, and paying $0.004/minute (I'm very low volume). I was considering switching to this, but their pricing is a terrible fit for my "most calls have mostly people who've never used it before" situation.

There are other JaaS providers that might have better pricing for you
Post reply on HN