Real-time stream of audio from your microphone

This is an example of using your own Microphone to continuously transcribe what is being uttered, while it is being uttered. Whenever the recognizer detects a silence in the audio stream from your microphone, the generator will return is_last=True, and the full transcription from the secondary model.

from danspeech import Recognizer
from danspeech.pretrained_models import CPUStreamingRNN, TestModel
from danspeech.audio.resources import Microphone
from danspeech.language_models import DSL3gram

print("Loading model...")
model = CPUStreamingRNN()

mic_list = Microphone.list_microphone_names()
mic_list_with_numbers = list(zip(range(len(mic_list)), mic_list))
print("Available microphones: {0}".format(mic_list_with_numbers))
mic_number = input("Pick the number of the microphone you would like to use: ")
m = Microphone(sampling_rate=16000, device_index=int(mic_number))

r = Recognizer()

print("Adjusting energy level...")
with m as source:
    r.adjust_for_ambient_noise(source, duration=1)


seconday_model = TestModel()
r = Recognizer(model=model)
try:
    lm = DSL3gram()
    r.update_decoder(lm=lm)
except ImportError:
    print("ctcdecode not installed. Using greedy decoding.")


r.enable_real_time_streaming(streaming_model=model, string_parts=False, secondary_model=seconday_model)
generator = r.real_time_streaming(source=m)

iterating_transcript = ""
print("Speak!")
while True:
    is_last, trans = next(generator)

    # If the transcription is empty, it means that the energy level required for data
    # was passed, but nothing was predicted.
    if is_last and trans:
        print("Final: " + trans)
        iterating_transcript = ""
        continue

    if trans:
        iterating_transcript += trans
        print(iterating_transcript)
        continue