just duplicating this from the blog post
Currently calling is established (e.g. both sides connect fine) but audio is not routed (no voice heard or sent)
A shitstorm is starting in 3… 2… 1…
My prediction on this is:
- A few people will moan about the Librem5 (and Purism in general) because the functionality for making calls isn’t ready. They will kick up as big a stink as they can, and be very negative.
- A larger number of people will complain about Purism’s communication/openness because this issue has been suspected for a while, but they had to wait until now to have it confirmed.
- Yet more people will be glad that they finally have concrete information about this, and hope/expect that the problem will be fixed before they get their phones.
Personally, I think that shipping Birch as soon as possible (i.e. without waiting for the audio routing to be sorted out) was the right decision.
Me too seeing as I expect that could be corrected by software update When it is corrected, it will be a great opportunity to show that the Librem 5 doesn’t follow the traditional “you get what you get until it’s obsolete” model other cell phones do.
Providing that it is only a software fix. If so, it should be fixed in the next few days and people will be complaining about something else.
It most certainly is. AFAIR, they do something rather unusual, which is: the baseband modem only gets mic/speaker access during calls. (Which of course is darn awesome)
So, it’s not that audio doesn’t actually work or something. It’s just that bit that is supposed to re-route audio is not ready.
I thought it was trickier than that because the question was whether the (outgoing) audio was sent via USB (on the M.2 connector) or via some PCM pins on the M.2 connector - and for all I know that may differ between different makes/models of modems. Either way though the main CPU is in control of sending the audio to the modem.
However I think you are right that in some architectures, audio can go directly from the audio source (typically a microphone) to the modem, and that makes it more difficult to control, even though it may be more performant. Does that apply here?
(For incoming audio the direction of flow is reversed and the audio sink will typically be a speaker. If you want to be able record calls or operate some kind of in-device answering machine then more complex still.)
This is all speculation. It all depends on exactly what is meant by “audio routing”.
Work has continued to extend wys to instantiate PulseAudio’s loopback module—which ties the modem’s and codec’s ALSA devices together when a call is activated, and de-instantiates the module when the call is terminated. Since this causes conflicts with hægtesse, a scheme was devised to keep both hægtesse and wys from running at the same time.
Nothing more performant about dedicated access from the modem to the mic/speaker vs not. The performance difference comes from hardware vs software audio routing, and direct vs indirect routing. Note that software routing is always indirect.
If you want high performance non-dedicated routing, you implement the routing in hardware (with control in software), with a hardware audio mixer. Of course, that’s another chip on the board, so adds expense (or size).
The prototype semi-portable C65 phone has a hardware audio router, and it’s pretty sweet (reboot the phone mid phone call without dropping the call).
I don’t feel very enlightened by that.
What is “hægtesse”, or “wys” for that matter? Don’t answer that. They’re working on it. Hopefully it will be working real soon now (next few days) - and all this discussion will be irrelevant.
Apparently, lmddgtfy means “Let Me DuckDuckGo That For You”: https://github.com/myano/lmddgtfy
To find this out, I used DuckDuckGo. So I guess I could have started this reply with “lmddgtfy”.
The lack of enlightenment was about the paragraph quoted by @Caliga not the acronym lmddgtfy.
Ah. OK well at least I have enlightened myself about the acronym.
Let me just note that the blog post contains links, that possibly even point in the direction of open issues, but I’m to lazy to check with my phone