[Resolved] Missing notifications in 2.0

[nrbrook]

I have created a new branch for 2.0 so I can switch between 1.x and 2. With 2 I am not getting the calibration notifications I was in 1.x.

Here’s a sample of logs using 1.x. The log lines in bold and without a timestamp are the events I have received out of OpenEarsEventsObserver

Recognition loop started
2014-12-11 18:59:59.075 Napkin[388:53167] Starting openAudioDevice on the device.
2014-12-11 18:59:59.075 Napkin[388:53167] Audio unit wrapper successfully created.
2014-12-11 18:59:59.087 Napkin[388:53167] Set audio route to SpeakerAndMicrophone
2014-12-11 18:59:59.090 Napkin[388:53167] Restoring SmartCMN value of 36.995846
2014-12-11 18:59:59.091 Napkin[388:53167] Checking and resetting all audio session settings.
2014-12-11 18:59:59.092 Napkin[388:53167] audioCategory is correct, we will leave it as it is.
2014-12-11 18:59:59.092 Napkin[388:53167] bluetoothInput is correct, we will leave it as it is.
2014-12-11 18:59:59.093 Napkin[388:53167] Output Device: SpeakerAndMicrophone.
2014-12-11 18:59:59.093 Napkin[388:53167] categoryDefaultToSpeaker is correct, we will leave it as it is.
2014-12-11 18:59:59.093 Napkin[388:53167] preferredBufferSize is incorrect, we will change it.
2014-12-11 18:59:59.094 Napkin[388:53167] PreferredBufferSize is now on the correct setting of 0.128000.
2014-12-11 18:59:59.095 Napkin[388:53167] preferredSampleRateCheck is correct, we will leave it as it is.
2014-12-11 18:59:59.096 Napkin[388:53167] Setting the variables for the device and starting it.
2014-12-11 18:59:59.096 Napkin[388:53167] Looping through ringbuffer sections and pre-allocating them.
2014-12-11 18:59:59.394 Napkin[388:53167] Started audio output unit.
2014-12-11 18:59:59.394 Napkin[388:53167] Calibration has started
Started Calibration
Completed Calibration
2014-12-11 19:00:01.597 Napkin[388:53167] Calibration has completed
2014-12-11 19:00:01.597 Napkin[388:53167] Listening.
Started listening
2014-12-11 19:00:16.092 Napkin[388:53167] Speech detected…
2014-12-11 19:00:17.315 Napkin[388:53167] Stopping audio unit.
Detect finished
2014-12-11 19:00:17.448 Napkin[388:53167] Audio Output Unit stopped, cleaning up variable states.
2014-12-11 19:00:17.449 Napkin[388:53167] Processing speech, please wait…
2014-12-11 19:00:17.515 Napkin[388:53167] Pocketsphinx heard “___REJ_IH ___REJ_T” with a score of (0) and an utterance ID of 000000000.
2014-12-11 19:00:17.515 Napkin[388:53167] Checking and resetting all audio session settings.
2014-12-11 19:00:17.516 Napkin[388:53167] audioCategory is correct, we will leave it as it is.
2014-12-11 19:00:17.516 Napkin[388:53167] bluetoothInput is correct, we will leave it as it is.
2014-12-11 19:00:17.517 Napkin[388:53167] Output Device: SpeakerAndMicrophone.
2014-12-11 19:00:17.517 Napkin[388:53167] categoryDefaultToSpeaker is correct, we will leave it as it is.
2014-12-11 19:00:17.517 Napkin[388:53167] preferredBufferSize is incorrect, we will change it.
The received hypothesis is ___REJ_IH ___REJ_T with a score of 0 and an ID of 000000000
2014-12-11 19:00:17.518 Napkin[388:53167] PreferredBufferSize is now on the correct setting of 0.128000.
2014-12-11 19:00:17.518 Napkin[388:53167] preferredSampleRateCheck is correct, we will leave it as it is.
2014-12-11 19:00:17.518 Napkin[388:53167] Setting the variables for the device and starting it.
2014-12-11 19:00:17.519 Napkin[388:53167] Looping through ringbuffer sections and pre-allocating them.
2014-12-11 19:00:17.570 Napkin[388:53167] Started audio output unit.
2014-12-11 19:00:17.570 Napkin[388:53167] Listening.
Started listening

Here’s the 2.0 output:

Started listening
Recognition loop started
2014-12-11 18:57:10.425 Napkin[381:52278] Speech detected…
2014-12-11 18:57:29.004 Napkin[381:52277] An utterance appears to be stuck in listening mode. Exiting stuck utterance.
2014-12-11 18:57:29.005 Napkin[381:52277] End of speech detected…
Detect finished
2014-12-11 18:57:29.224 Napkin[381:52277] Pocketsphinx heard “___REJ_D” with a score of (0) and an utterance ID of 0.
The received hypothesis is ___REJ_D with a score of 0 and an ID of 0
2014-12-11 18:57:29.906 Napkin[381:52277] Speech detected…
2014-12-11 18:58:01.008 Napkin[381:52448] An utterance appears to be stuck in listening mode. Exiting stuck utterance.
2014-12-11 18:58:01.009 Napkin[381:52448] End of speech detected…
Detect finished
2014-12-11 18:58:01.363 Napkin[381:52448] Pocketsphinx heard “___REJ_S ___REJ_EH” with a score of (0) and an utterance ID of 1.
The received hypothesis is ___REJ_S ___REJ_EH with a score of 0 and an ID of 1
2014-12-11 18:58:01.519 Napkin[381:52448] Speech detected…

[Halle Winkler]

There is no calibration in 2.0. Check out the docs and class headers to check for API changes.

[nrbrook]

Ah, my mistake, I thought it still existed in the protocol.

[Halle Winkler]

No problem! I wish there was a way to raise errors for nonexistent delegate methods but they are just interpreted by the compiler as new methods if they aren’t part of a protocol. There are a couple of other positive changes in there so check it out.

[nrbrook]

You can do

– (void) method __deprecated;

But it doesn’t display a warning, so it isn’t much use…

http://stackoverflow.com/questions/11432452/attribute-deprecated-does-not-work-on-objective-c-protocol-methods

[Halle Winkler]

That would be an option, and I was also considering keeping the callback but just returning a warning that the callback no longer does anything.

[nrbrook]

You could just check on delegate set if the delegate implements the removed methods and log a warning.

[Halle Winkler]

But since they are functionally gone, it’s probably most semantically clear for them to be gone from the source, and I should just call it out specifically as one of the steps in the upgrade guide.

[Halle Winkler]

You could just check on delegate set if the delegate implements the removed methods and log a warning.

That’s a good idea, I’ll check it out for 2.01.

[Author]

Posts