def RappSpeechDetectionSphinx(self,
audioString,
audioFile,
language,
audioSource,
words=None,
sentences=None,
grammar=None):
path = audioFileORaudioString(audioString, audioFile)
ch = RappPlatformAPI(address=addr)
response = ch.speechRecognitionSphinx(path,
language=language,
audio_source=audioSource,
words=words,
sentences=sentences,
grammar=grammar)
#f=open('/home/panos/Desktop/Swagger/finalTest/test_v3/RappGoogle.wav','r')
return response
class voiceCommand:
def __init__(self):
self.frames = []
self.pa = pyaudio.PyAudio()
self.stream = self.open_mic_stream()
self.tap_threshold, self.noise_threshold = self.calibrate()
self.hear = False
self.ch = RappPlatformAPI()
topic = rospy.get_param("rec_topic")
self.publisher = rospy.Publisher(topic, String, queue_size=10)
def stop(self):
self.stream.close()
#finds device inputs
def find_input_device(self):
device_index = None
for i in range(self.pa.get_device_count()):
devinfo = self.pa.get_device_info_by_index(i)
print("Device %d: %s" % (i, devinfo["name"]))
for keyword in ["mic", "input"]:
if keyword in devinfo["name"].lower():
print("Found an input: device %d - %s" %
(i, devinfo["name"]))
device_index = i
return device_index
if device_index == None:
print("No preferred input found; using default input device.")
return device_index
#opens stream
def open_mic_stream(self):
device_index = self.find_input_device()
stream = self.pa.open(format=FORMAT,
channels=CHANNELS,
rate=RATE,
input=True,
input_device_index=device_index,
frames_per_buffer=INPUT_FRAMES_PER_BLOCK)
return stream
def tapDetected(self):
print "Tap!"
def unpack(self, block):
count = len(block) / 2
format = "%dh" % (count)
shorts = struct.unpack(format, block)
return shorts, count
def get_rms(self, block):
# RMS amplitude is defined as the square root of the
# mean over time of the square of the amplitude.
# so we need to convert this string of bytes into
# a string of 16-bit samples...
# we will get one short out for each
# two chars in the string.
shorts, count = self.unpack(block)
# iterate over the block.
sum_squares = 0.0
for sample in shorts:
# sample is a signed short in +/- 32768.
# normalize it to 1.0
n = sample * SHORT_NORMALIZE
sum_squares += n * n
return math.sqrt(sum_squares / count)
#computeThreshold listens to stream for $seconds seconds
#and computes the threshold using RMS
def computeThreshold(self, seconds):
block_sum = 0
counter = 0
while seconds > 0:
block_sum += self.get_rms(self.stream.read(INPUT_FRAMES_PER_BLOCK))
seconds -= INPUT_BLOCK_TIME
counter += 1
return block_sum / counter
#calibrate compute silence and voice thresholds
def calibrate(self):
print("Noise calibration.BE QUIET.")
noise_threshold = self.computeThreshold(2)
print("Noise calibration COMPLETE.")
print("Command calibration.")
tap_threshold = self.computeThreshold(2)
print("Command calibration complete.")
return tap_threshold, noise_threshold
#write2WAV saves self.frames as file.wav
def write2WAV(self):
waveFile = wave.open(PATH + 'file.wav', 'w')
waveFile.setnchannels(CHANNELS)
waveFile.setsampwidth(self.pa.get_sample_size(FORMAT))
waveFile.setframerate(RATE)
waveFile.writeframes(b''.join(self.frames))
waveFile.close()
self.frames = []
def setHear(self, value):
self.hear = value
def listen(self):
try:
downCounter = 0
upCounter = 0
self.frames = []
flag = False
while self.hear and (not rospy.is_shutdown()):
block = self.stream.read(INPUT_FRAMES_PER_BLOCK)
amp = self.get_rms(block)
if amp > self.tap_threshold + self.noise_threshold:
flag = True #heard something for first time
#self.tapDetected()
downCounter = 0
upCounter += 1
else:
downCounter += 1
if flag:
self.frames.append(block)
else:
if len(self.frames) > 3:
self.frames.pop(0)
self.frames.append(block)
if upCounter > 40: #Too long record 40*0.05 = 2sec
print("TOO LONG RECORD")
break
if downCounter > 20:
if upCounter > 1:
self.stream.stop_stream()
self.write2WAV()
print('File recorded.')
self.stream.start_stream()
time.sleep(0.1)
downCounter = 0
upCounter = 0
self.hear = False
else:
downCounter = 0
upCounter = 0
self.frames = []
flag = False
except KeyboardInterrupt:
self.stop()
return False
return True
def voiceRec(self):
response = self.ch.speechRecognitionSphinx(PATH + "file.wav",
"nao_wav_1_ch", 'en',
dictionary)
str_response = str(response.get('words'))
if len(response.get('words')) != 1:
return ''
str_response = str(response.get('words')[0])
return str_response
class SpeechDetectionSphinxTests(unittest.TestCase):
def setUp(self):
rospack = rospkg.RosPack()
self.pkgDir = rospack.get_path('rapp_testing_tools')
self.ch = RappPlatformAPI()
def test_ogg_oxi(self):
audioFile = path.join(self.pkgDir, 'test_data',
'speech_detection_samples', 'recording_oxi.ogg')
response = self.ch.speechRecognitionSphinx(audioFile, 'nao_ogg', 'el',\
[u'ναι', u'οχι'])
valid_words_found = [u'οχι']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_ogg_no(self):
audioFile = path.join(self.pkgDir, 'test_data',
'speech_detection_samples', 'recording_no.ogg')
response = self.ch.speechRecognitionSphinx(audioFile, 'nao_ogg', 'en',\
[u'yes', u'no'])
valid_words_found = [u'no']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_wav_1_ch_yes_no(self):
audioFile = path.join(self.pkgDir, 'test_data', 'yes-no.wav')
response = self.ch.speechRecognitionSphinx(audioFile, 'nao_wav_1_ch', 'en',\
[u'yes', u'no'])
valid_words_found = [u'yes', u'no']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_wav_1_ch_nai_oxi(self):
audioFile = path.join(self.pkgDir, 'test_data', 'nai-oxi-test.wav')
response = self.ch.speechRecognitionSphinx(audioFile, 'nao_wav_1_ch', 'el',\
[u'ναι', u'οχι', u'ισως'])
valid_words_found = [u'ναι', u'οχι', u'ισως']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_headset_nai_oxi(self):
audioFile = path.join(self.pkgDir, 'test_data', 'microphone_nai.wav')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', 'el',\
[u'ναι', u'οχι'])
valid_words_found = [u'ναι']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_speech_erroneous(self):
audioFile = path.join(self.pkgDir, 'test_data', 'microphone_nai.wav')
response = self.ch.speechRecognitionSphinx('', 'headset', 'el',
[u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx([], 'headset', 'el',
[u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(3, 'headset', 'el',
[u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, '', 'el',
[u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, [], 'el',
[u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 3, 'el',
[u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', '',
[u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', [],
[u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', 3,
[u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', 'el',
3)
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', 'el',
'')
self.assertNotEqual(response['error'], u'')
class SpeechDetectionSphinxTests(unittest.TestCase):
def setUp(self):
rospack = rospkg.RosPack()
self.pkgDir = rospack.get_path('rapp_testing_tools')
self.ch = RappPlatformAPI()
def test_ogg_oxi(self):
audioFile = path.join(self.pkgDir, 'test_data',
'speech_detection_samples', 'recording_oxi.ogg')
response = self.ch.speechRecognitionSphinx(audioFile, 'nao_ogg', 'el',\
[u'ναι', u'οχι'])
valid_words_found = [u'οχι']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_ogg_no(self):
audioFile = path.join(self.pkgDir, 'test_data',
'speech_detection_samples', 'recording_no.ogg')
response = self.ch.speechRecognitionSphinx(audioFile, 'nao_ogg', 'en',\
[u'yes', u'no'])
valid_words_found = [u'no']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_wav_1_ch_yes_no(self):
audioFile = path.join(self.pkgDir, 'test_data',
'yes-no.wav')
response = self.ch.speechRecognitionSphinx(audioFile, 'nao_wav_1_ch', 'en',\
[u'yes', u'no'])
valid_words_found = [u'yes', u'no']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_wav_1_ch_nai_oxi(self):
audioFile = path.join(self.pkgDir, 'test_data',
'nai-oxi-test.wav')
response = self.ch.speechRecognitionSphinx(audioFile, 'nao_wav_1_ch', 'el',\
[u'ναι', u'οχι', u'ισως'])
valid_words_found = [u'ναι', u'οχι', u'ισως']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_headset_nai_oxi(self):
audioFile = path.join(self.pkgDir, 'test_data',
'microphone_nai.wav')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', 'el',\
[u'ναι', u'οχι'])
valid_words_found = [u'ναι']
self.assertEqual(response['error'], u'')
self.assertEqual(response['words'], valid_words_found)
def test_speech_erroneous(self):
audioFile = path.join(self.pkgDir, 'test_data', 'microphone_nai.wav')
response = self.ch.speechRecognitionSphinx('', 'headset', 'el', [u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx([], 'headset', 'el', [u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(3, 'headset', 'el', [u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, '', 'el', [u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, [], 'el', [u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 3, 'el', [u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', '', [u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', [], [u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', 3, [u'ναι', u'οχι'])
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', 'el', 3)
self.assertNotEqual(response['error'], u'')
response = self.ch.speechRecognitionSphinx(audioFile, 'headset', 'el', '')
self.assertNotEqual(response['error'], u'')
