"tee" bash command for live streaming applications

timbiker

Dear all,

first time I post here, trying to find the best place to search help on my matter.

My basic problem is that: I want to make a "tee" command which would keep on working as a charm if one or multiple of the output pipes breaks or encounters any issue.

The background is the following:

I use a bunch of homemade bash script for some video recording/live streaming...

On key part of the video/audio path is that it is first grabbed from a firewire capture card, then go to a "tee" and from there is split to multiple application:

- Full quality dv recording

- live encoding and streaming to internet

- feed monitoring

As I can start and stop these "services" independently, I need to make sure that whatever happens, the "tee" keeps to taking the video from the capturing process and feed them to the services still alive without delaying (would cause frame drop) or anything similar.

Currently I use a modified version from the coreutils 5.0 (file attached, as well as diff from coreutils).

It works pretty well but I still see some cases where if my internet line drops (making the live stream encoding to pause or to break, the recording is also impacted.

However my C knowledge are not sufficient to see what more to do in this tee.c to reallly make it robust.

I expect something to do with signal handling but...

Could anyone give a hint, or point me to a better place if appropriate ?

Many thanks

Timothé

timbiker

renamed attachment in .txt...

romuffel

Hi Timothé,

Welcome to the forum!

Your use case with the "tee" command for video/audio processing sounds interesting and challenging. From what you've described, it seems like you need a more resilient solution that can handle interruptions in one of the output pipes without affecting the overall performance of your streaming and recording.

One approach to improve robustness might be to implement better error handling and recovery mechanisms. This could involve:

Signal Handling: Implementing signal handling in your custom "tee" command to manage interruptions and ensure that the process continues to function smoothly.
Buffering: Introducing additional buffering to handle temporary disruptions without causing frame drops.
Failover Mechanisms: Setting up failover mechanisms for critical components to ensure that if one output fails, the others continue to operate seamlessly.

If you're looking for a more advanced and reliable solution for live streaming and video processing, you might want to explore specialized services. Cogniteq offers professional live streaming solutions that could provide the robustness and flexibility you need for your setup.

Hope this helps, and good luck with your project!