Updating audio sinks

[lolodomo]

Core PR openhab/openhab-core#3461 improved audio feature and allows a simplification for audio sinks and support for more audio streams.

• allplay => PR [allplay] Support for more audio streams through the HTTP audio servlet #15201
• amplipi => PR [amplipi] Support for more audio streams through the HTTP audio servlet #15199
• chromecast => PR [chromecast] Support for more audio streams through the audio servlet #15180
• doorbird => PR [doorbird] Support for more audio streams #15189
• freebox => PR [freebox] Support for more audio streams through the HTTP audio servlet #15121
• freeboxos: => PR [freeboxos] Align audio sink to freebox binding #15207
• heos => PR [heos] Support for more audio streams through the HTTP audio servlet #15196
• kodi => PR [kodi] Support for more audio streams through the HTTP audio servlet #15192
• onkyo => PR [onkyo] Support for more audio streams through the HTTP audio servlet #15117
• pulseaudio => PR [pulseaudio] Make the process method asynchronous #15179
• sonos => PR [sonos] Support for more audio streams through the HTTP audio servlet #15116
• squeezebox => PR [squeezebox] Support for more audio streams through the audio servlet #15194
• upnpcontrol => PR [upnpcontrol] Support for more audio streams through the HTTP audio s… #15122

There is a pending discussion with @dalgwen if URLAudioStream should be proxied or not by the openHAB server. I am not very favorable to that option.

[lolodomo]

@dalgwen : you said in another issue that you will provide a PR for Chromecast, doorbell and pulseaudio. So I do not update these ones. It will be interesting to see if your proposed "pattern" is identical to mine. You are welcome to comment the PRs I have already submitted.

[dalgwen]

@lolodomo
I think you will want to kill me, because I have another pending pull request for an audio enhancement in core.
When I made the mimic TTS simplification relatives to core-3461, I found that I couldn't make all the simplification I wanted (I wanted to get rid of FixedLengthInputStream but couldn't) because the length() information was important further in the sink process method I tried (pulseaudio).
And in mimic I don't want to use FixedLength, because the stream is not cloneable, as it is not a file (which is a huge part of the audio simplification goal...)
So as we discussed in one previous comment some time ago, there is a real need to split the length information from the FixedLength class... new PR...
The PR is simple, but time is lacking before the 4.0 release...

[lolodomo]

I just reviewed your new PR.

And in mimic I don't want to use FixedLength, because the stream is not cloneable, as it is not a file (which is a huge part of the audio simplification goal...)

In my VoiceRSS PR, when TTS cache is disabled, I simply provide an AudioStream and it seemed to work well with the Sonos sink.
So, considering your new PR, I will have to also implement SizeableAudioStream as the service knows the stream length.

[dalgwen]

In my VoiceRSS PR, when TTS cache is disabled, I simply provide an AudioStream and it seemed to work well with the Sonos sink. So, considering your new PR, I will have to also implement SizeableAudioStream as the service knows the stream length.

Not for all sink, but the size information could be valuable for some (the pulseaudio binding tries to get duration information from it). It is not mandatory, but length detection could fail and default to the stream consumption time. Not a big deal, it is still better than before.
I also know that, if we have the information, we set the length in the HTTP response header of the http server (I don't know if any binding using the http server really needs it).

[lolodomo]

@dalgwen : your last core PR has just been merged. So we can start again updates of audio sinks and TTS services.

[dalgwen]

Great, it was fast !

I'm taking doorbird and pulseaudio in top priority, as I am the sink's creator for both of them. Maybe today or tomorrow.
I have something nearly ready on chromecast too.

If I have time, I could take squeezebox and kodi, as I use them at home (but not for their sink capability)

[lolodomo]

@clinique : I don't know what Freebox player do you own but in case you have one that is not a Freebox Revolution and that supports AirPlay, you are welcome to test the audio sink feature with the freeboxos binding.

[lolodomo]

@jlaur : in case you have some time, can you please help merging all the remaining opened PRs listed in this issue ?

[jlaur]

@jlaur : in case you have some time, can you please help merging all the remaining opened PRs listed in this issue ?

I'm on vacation until Sunday, so can't promise anything. I'm mostly on the phone, and it won't be sufficient for this. I might fire up my backup laptop one evening, but can't promise.

[kaikreuzer]

I don't want that @jlaur spoils his vacation too much - I'll take care of the remaining PRs!

[kaikreuzer]

All merged!

[lolodomo]

Thank you @jlaur and @kaikreuzer

[openhab-bot]

This issue has been mentioned on openHAB Community. There might be relevant details there:

https://community.openhab.org/t/webaudio-not-working-when-trying-to-stream-webradio/148610/14