This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(208, 0%, 30%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EB%3C/text%3E%3C/svg%3E)
I am using start_continuous_recognition to start the recognition, but for some reason speech_end_detected is not being triggered so stop_continuous_recognition is never being called.
It used to work, but for some reason it isn't anymore. I was using version 1.24.2, but I also tried 1.29.0 and 1.30.0 with but the issue persists.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(195.2, 4%, 28%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EVM%3C/text%3E%3C/svg%3E)
Hello @Bryan , Thanks for using Microsoft Q&A Platform.
Is it possible to share the code and the error details with us so that we can reproduce the same on our end.
Yes, of course I can share the code. I am only having issues with the stop_cb function being triggered. It seems that the "ending" event never happens.
speech_config = SpeechConfig(subscription=AZURE_KEY, region=AZURE_REGION)
CACHE = {}
def azure_transcribe_from_file(filepath: str, language: str) -> list:
"""The transcribe
Args:
filepath (str): The filepath to the audio file for transcription.
language (str): The language chosen by the user.
Raises:
TimeoutException: A custom timeout exception for long-running transcriptions.
Returns:
list: A list of the events during the transcription.
"""
azure_start = timer()
CACHE[filepath] = False
speech_config.output_format = OutputFormat.Detailed
audio_config = AudioConfig(filename=filepath)
speech_recognizer = SpeechRecognizer(
speech_config=speech_config, audio_config=audio_config, language=language
)
events = []
def recognized_cb(event: dict) -> None:
"""A callback to load the results into the events list.
Args:
event (dict): The event during the transciption.
"""
text = event.result.text
if text:
logger.info("Azure: Recognized-> %s", event.result.text)
events.append(loads(event.result.json))
else:
logger.warning(
"Azure: No text for this event: %s", loads(event.result.json)
)
def stop_cb(event: dict) -> None:
"""A callback that stops continuous recognition upon receiving an `event`
Args:
event (dict): The event during the transciption.
"""
logger.info("Azure: CLOSING on %s", event.result.json)
CACHE[filepath] = True
speech_recognizer.stop_continuous_recognition()
# Setting speech callbacks
speech_recognizer.recognized.connect(recognized_cb)
speech_recognizer.speech_end_detected.connect(stop_cb)
logger.info("Azure: starting speech recognition.")
speech_recognizer.start_continuous_recognition()
while not CACHE[filepath] or not events:
if (timer() - azure_start) > SERVICE_TIMEOUT:
speech_recognizer.stop_continuous_recognition()
raise TimeoutException("Azure exceeded timeout.")
continue
logger.info("Azure: finished speech recognition: %s", events)
return events
@VasaviLankipalle-MSFT have you had any time to test this?
I was able to resolve the issue. I made two changes.
I removed this line:
logger.info("Azure: CLOSING on %s", event.result.json)
Trying to get the result.json in the closing event was messing it up.
Also I moved the speech_recognizer.stop_continuous_recognition() to after the while loop.