def task(quit_event):
#mic = Microphone(quit_event=quit_event)
mic_index = None
for i, microphone_name in enumerate(Microphone.list_microphone_names()):
if 'seeed' in microphone_name:
mic_index = i
print("Using microphone {}".format(microphone_name))
break
if not mic_index:
print("Could not find a proper microphone")
exit()
with Microphone(device_index=mic_index) as mic:
recognizer = Recognizer()
while not quit_event.is_set():
pixel_ring.off()
print("Listening for keyword")
data = recognizer.listen(source=mic)
kw_text = recognizer.recognize_sphinx(data)
print("Heard '{}' while listening for keyword".format(kw_text))
if kw_text == name:
print('Wake up')
pixel_ring.listen()
data = recognizer.listen(mic)
pixel_ring.think()
text = recognizer.recognize_sphinx(data)
print('Done listening')
pixel_ring.off()
if text:
print('Recognized {}'.format(text))
tts.say(text)
def task(quit_event):
with Microphone(device_index=mic_index, quit_event=quit_event) as mic:
recognizer = Recognizer()
while not quit_event.is_set():
pixel_ring.off()
print("Listening for keyword")
data = recognizer.listen(source=mic)
kw_text = recognizer.recognize_sphinx(data)
print("Heard '{}' while listening for keyword".format(kw_text))
if kw_text == name:
print('Wake up')
pixel_ring.listen()
data = recognizer.listen(mic)
pixel_ring.think()
text = recognizer.recognize_sphinx(data)
print('Done listening')
pixel_ring.off()
if text:
print('Recognized {}'.format(text))
tts.say(text)
def recognizer_func(recognizer: sr.Recognizer, audio: sr.AudioData):
try:
# for testing purposes, we're just using the default API key
# to use another API key, use `r.recognize_google(audio, key="GOOGLE_SPEECH_RECOGNITION_API_KEY")`
# instead of `r.recognize_google(audio)`
print("Decoding")
print("Google Speech Recognition thinks you said: " + recognizer.recognize_sphinx(audio))
except sr.UnknownValueError:
print("Google Speech Recognition could not understand audio")
except sr.RequestError as e:
print("Could not request results from Google Speech Recognition service; {0}".format(e))
class AudioToTextSphinxPipe(PushPipe[AudioRecording, str]):
def __init__(self, postProcessCallback=None):
super().__init__(postProcessCallback=postProcessCallback)
self.sr = Recognizer()
def process(self, recording: AudioRecording,
passThrough: PushPipe.PassThrough):
print('Sample width:', recording.METADATA.getFormatByteSize())
data = AudioData(recording.data, recording.METADATA.Rate,
recording.METADATA.getFormatByteSize())
try:
return self.sr.recognize_sphinx(data, 'en-IN')
except UnknownValueError:
self.setErrored("Couldn't identify speech.")
return ''
def listen_for_hotword(self, recognizer: sr.Recognizer, audio: sr.AudioData):
"""
Callback that runs continuously in the background to listen
for hotwords using offline speech recognition.
:param recognizer: the Recognizer instance
:param audio: audio data to analyze
:return: None
"""
if self.listening_for_command:
return
try:
speech_as_text: Decoder = recognizer.recognize_sphinx(audio, keyword_entries=keywords)
print(f"Heard '{speech_as_text}'")
for keyword, sensitivity in keywords:
if keyword in speech_as_text:
self.listen_command_queue.put('command')
break
except sr.UnknownValueError:
print("Oops! Didn't catch that")
class CommandsRecognition(ApiState):
def __init__(self, callback, language="pl-PL", api_option=ApiOption.GOOGLE):
super().__init__()
self.callback_command_detected = callback
self.api_option = api_option
self.language = language
self.listen_thread = None
self.phrase_time_limit = 3
try:
self.recognizer = Recognizer()
self.microphone = Microphone()
"""
adjust the recognizer sensitivity to ambient noise and record audio
from the microphone
"""
with self.microphone as source:
self.recognizer.adjust_for_ambient_noise(source)
except OSError as ose_err:
Logger.critical("OSError: {0}".format(ose_err))
self.api_runs = False
self.api_info = GLO_MSG['MICROPHONE_FAILURE']
return
except Exception as err:
Logger.critical("Exception: {0}".format(err))
self.api_runs = False
self.api_info = GLO_MSG['MICROPHONE_FAILURE']
return
Logger.debug("Initialization of CommandsRecognition class")
self.api_info = GLO_MSG['MICROPHONE_INITIALIZED']
def validate_command(self, command):
if command.lower() in GLO_CMD.values():
detected_command_id = GET_COMMAND(command.lower())
Logger.info("GLO_CMD command available -> {0}".format(command.lower()))
self.callback_command_detected(detected_command_id)
else:
Logger.info("Detected command: {0}".format(command.lower()))
def callback_recognition(self, recognizer, audio):
try:
command = self._api_recognition(audio)
self.validate_command(command)
except UnboundLocalError as err:
Logger.warning("UnboundLocalError : {0} ".format(err))
except RequestError as err:
Logger.warning("RequestError : {0} ".format(err))
except UnknownValueError as err:
Logger.debug("UnknownValueError : {0} ".format(err))
def background_listen(self):
self.listen_thread = self.recognizer.listen_in_background(
source=self.microphone, callback=self.callback_recognition, phrase_time_limit=self.phrase_time_limit)
def _api_recognition(self, audio):
if self.api_option is ApiOption.GOOGLE:
return self.recognizer.recognize_google(audio, language=self.language)
elif self.api_option is ApiOption.GOOGLE_CLOUD:
# Support languages: https://cloud.google.com/speech-to-text/docs/languages
return self.recognizer.recognize_google_cloud(audio, credentials_json='', language=self.language)
elif self.api_option is ApiOption.SPHINX:
# Support languages : https://sourceforge.net/projects/cmusphinx/files/Acoustic%20and%20Language%20Models/
return self.recognizer.recognize_sphinx(audio, language=self.language)
elif self.api_option is ApiOption.WIT:
# Support languages : https://wit.ai/faq, login required
return self.recognizer.recognize_wit(audio, key='',)
elif self.api_option is ApiOption.AZURE_BING:
# Support languages : https://docs.microsoft.com/en-us/azure/cognitive-services/bing-web-search/language-support, login required
self.recognizer.recognize_bing(audio, key='', language=self.language)
elif self.api_option is ApiOption.LEX:
# Support languages: ONLY ENG -> https://docs.aws.amazon.com/lex/latest/dg/gl-limits.html
return self.recognizer.recognize_lex(audio)
elif self.api_option is ApiOption.HOUNDIFY:
# Support languages: ONLY ENG, login required
return self.recognizer.recognize_houndify(audio, client_id='', client_key='')
elif self.api_option is ApiOption.IBM:
# Support languages : https://www.ibm.com/watson/services/language-translator/, login required
return self.recognizer.recognize_ibm(audio, username='', password='', language=self.language)
else:
Logger.error("Api recognition option is not defined")
class Speech(object):
def __init__(self):
rospy.init_node("speech")
self.pub = rospy.Publisher("speech_recognizer",
String,
latch=True,
queue_size=1)
self.recognizer = Recognizer()
# self.recognizer.energy_threshold = 1000
# self.recognizer.pause_threshold = .7
self.microphone = Microphone()
@staticmethod
def check_internet_connection():
connection = httplib2.HTTPConnectionWithTimeout("www.google.com",
timeout=5)
try:
connection.request("HEAD", "/")
connection.close()
return True
except httplib2.HttpLib2Error:
connection.close()
return False
def __recognize_helper(self, recognized_speech):
'''
Once speech obtained do TODO:something with it
'''
print(recognized_speech)
if recognized_speech != "":
rospy.loginfo("You said: " + recognized_speech)
self.pub.publish(recognized_speech)
def __microphone_helper(self, ):
rospy.loginfo('Listening...')
with self.microphone as source:
# If phase_time_limit is not set to 5 it will
# take a really long time every 3-4th attempt
audio = self.recognizer.listen(source, phrase_time_limit=5)
rospy.loginfo('Got a sound; recognizing...')
return audio
def recognize_google(self):
with self.microphone as source:
self.recognizer.adjust_for_ambient_noise(source)
try:
while not rospy.is_shutdown():
audio = self.__microphone_helper()
recognized_speech = ""
if Speech.check_internet_connection():
try:
recognized_speech = self.recognizer.recognize_google(
audio)
except sr.UnknownValueError:
rospy.logerr("Could not understand audio.")
except sr.RequestError:
rospy.logerr("Could not request results.")
else:
rospy.logerr("No internet conneciton for Google API")
self.__recognize_helper(recognized_speech)
except Exception as exc:
rospy.logerr(exc)
def recognize_sphynix(self):
with self.microphone as source:
self.recognizer.adjust_for_ambient_noise(source)
try:
while not rospy.is_shutdown():
audio = self.__microphone_helper()
recognized_speech = ""
try:
recognized_speech = self.recognizer.recognize_sphinx(audio)
except sr.UnknownValueError:
rospy.logerr("Could not understand audio.")
except sr.RequestError:
rospy.logerr("""Could not request results. Do you have
pocket Sphynx installed?""")
self.__recognize_helper(recognized_speech)
except Exception as exc:
rospy.logerr(exc)
# On enregistre le son
with Microphone() as source:
print("Réglage du bruit ambiant... Patientez...")
recognizer.adjust_for_ambient_noise(source)
print("Vous pouvez parler...")
recorded_audio = recognizer.listen(source)
print("Enregistrement terminé !")
with open("record.wav", "wb") as f:
f.write(recorded_audio.get_wav_data())
# Reconnaissance de l'audio
try:
print("Reconnaissance du texte...")
googletext = recognizer.recognize_google(recorded_audio, language="fr-FR")
print("Google pense que vous avez dit {}".format(googletext))
except UnknownValueError:
print("L'audio n'as pas été compris par Google")
except Exception as ex:
print(ex)
try:
print("Reconnaissance du texte...")
sphinxtext = recognizer.recognize_sphinx(recorded_audio, language="fr-FR")
print("Sphinx pense que vous avez dit {}".format(sphinxtext))
except UnknownValueError:
print("L'audio n'as pas été compris par Sphinx")
except Exception as ex:
print(ex)