Chrome starting version 76 onwards, runs Audio in a separate process on Windows, Mac, and Linux platforms. Google has moved the Audio from browser process to a separate process and you can now notice “Utility: Audio Service” running in Chrome Task Manager when video or audio is playing in the browser.
With the system audio running in the browser process, if audio crashes, the whole browser will crash. If the audio hangs, there will be no audio for the total browsing session.
with this change ” the browser would restart the audio process upon its crash and seamlessly pick up the audio so that it results in a one-time glitch.”
To check if your Chrome is running audio service in a separate process
- Launch Chrome 76 stable or later
- Visit YouTube
- Play any video, press Shift+Esc on Widnows to open Task Manager in Chrome, you can see Utility: Audio service task running in TM.
Chromium team believes “moving device-facing audio functionality to a mojo component goes along with Chrome modularization effort and moving it out of browser process to a separate audio process should improve Chrome stability and availability of audio. since it will restart a failed audio stack without the whole browser restart.”
Google doesn’t want to run Audio in a dedicated process on Android and Chrome OS. The “move platform audio to seperate process” bug for Chrome 76 is closed and fixed today.
The immediate benefit of moving platform audio to a separate process is increased browser stability due to isolating the browser process from audio driver faults.
The long-term benefit is enabling audio processing logic (such as echo cancellation) to be run in the audio process, i.e. closer to the hardware and without passing audio buffers to renderer processes wherever possible, which means having a more stable audio path latency and better performance (we can better tweak priority of the audio process as well).
Other than extra process count, this change brings stability and performance improvements to Chrome, do you have any concerns with this?