The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/fa0c0b5be00eaef35dacf77ec0b1d44868d4d5c4

commit fa0c0b5be00eaef35dacf77ec0b1d44868d4d5c4
Author: Guido Urdaneta <guidou@chromium.org>
Date: Wed Oct 03 09:54:42 2018

Let Pulse Audio select input device to use as default in AudioManagerPulse

The default input device reported by Pulse Audio is not really
a default device, but a fallback device.
Actual expected default behavior in Pulse Audio is achieved by
not specifying a device to use and letting Pulse Audio select it.
In this case Pulse first tries to use the device specified by the
PULSE_SOURCE environment variable and, if that fails, tries to use
the fallback device.

There is no API to fetch what device Pulse will actually use by default.
A consequence of this is that the groupId reported by the default
input device will not correspond to the group ID of the actual device
that will be used. This is not ideal, but it is better than the
alternative.

The groupId constrainable property still works fine with nondefault
devices.

Bug:  890010 
Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel
Change-Id: I3473df314f0a36f56094c4986bf8fc9b38265fad
Reviewed-on: https://chromium-review.googlesource.com/c/1256832
Reviewed-by: Max Morin <maxmorin@chromium.org>
Commit-Queue: Guido Urdaneta <guidou@chromium.org>
Cr-Commit-Position: refs/heads/master@{#596170}
[modify] https://crrev.com/fa0c0b5be00eaef35dacf77ec0b1d44868d4d5c4/media/audio/pulse/audio_manager_pulse.cc

Which version of Chrome 71 will this fix be on? I'm on 71.0.3559.6, right now, and apparently I'm affected by this bug, so I'd like to know when I'll be able to test if this works for me or not.

It should be available in the next dev version.

Looks like there has been a regression... as you said, it was fixed in one of the updates after the one where it was broken. Anyway, it stopped working yesterday: I now updated to 72.0.3622.0 and I'm getting the NotReadableError again for the Default device; manually specifying the only capture device (which is what Default should point to) works. Not sure if it's still a Pulse-related problem or something else.

Can you file a new bug where you can provide more info?
I'm not aware of more changes in this area, but will take a look.