Add a new parameter to method synthesizeForCache

[lolodomo]

Related to openhab/openhab-addons#14564 (comment)

Signed-off-by: Laurent Garnier [email protected]

[lolodomo]

@dalgwen to discuss this proposal

[dalgwen]

Is this modification only for the workaround around the limitation of the servlet based audio sinks ? Or it is also for the MaryTTS service and its custom behaviour with cache files ?

Another though : couldn't this boolean be a part of the AudioFormat object ?
It could then join the other matching capabilities process (when openHAB try to match the audio format requested with the TTS capabilities, i.e. wav, mp3, etc) :
The TTS service could declare if it supports FixedLength, as the AudioSink already does with the getSupportedStreams method.
Then when openHAB matches (getBestMatch), the requested AudioFormat can be filled with the information that the sink needs a FixedLengthAudioStream.

(And the cache could intercept the requested AudioFormat and disable the need of a FixedLengthAudioStream compatible response, because it knows that it will create a file itself)

[lolodomo]

Is this modification only for the workaround around the limitation of the servlet based audio sinks ? Or it is also for the MaryTTS service and its custom behaviour with cache files ?

It was really for a workaround around the limitation of the servlet based audio sinks.
As you have already (almost) a solution for that, my proposal could be then be forgiven.
Priority to your solution. I will close this PR as soon as you submit your PR.

I have not yet looked at the MaryTTS service.

[dalgwen]

I made a PR that could solve the need for a workaround for the AudioServlet based sink. Hope this will help for your use case.

[lolodomo]

I made a PR that could solve the need for a workaround for the AudioServlet based sink. Hope this will help for your use case.

Yes, very probably.