I'm trying to detect a keyword from a .wav file using Pocketsphinx, specifically with the decoder class. When I give it this .wav file and print what it detects it isnt even close. Here is the code:
import pocketsphinx as ps
import requests
import json
import sys, os
import subprocess
model_path = ps.get_model_path()
data_path = ps.get_data_path()
print("start")
print(os.getcwd())
subprocess.call("sox -V4 /home/miro/client_audio.wav -r 16000 -c 1 client_audio.wav", shell=True)
config = ps.Decoder.default_config()
config.set_string('-kws', 'keyphrase.list')
config.set_string('-hmm', os.path.join(model_path, 'en-us'))
config.set_string('-lm', os.path.join(model_path, 'en-us.lm.bin'))
config.set_string('-dict', os.path.join(model_path, 'cmudict-en-us.dict'))
stream = open("client_audio.wav", "rb")
decoder = ps.Decoder(config)
decoder.start_utt()
while True:
buf = stream.read(1024)
if buf:
decoder.process_raw(buf, False, False)
else:
break
if decoder.hyp() != None:
# print ([(seg.word, seg.prob, seg.start_frame, seg.end_frame) for seg in decoder.seg()])
words=[]
[words.append(seg.word) for seg in decoder.seg()]
print(words)
decoder.end_utt()
decoder.start_utt()
It prints this:
['<s>', "it's"]
Does anyone know why this is?
Credit goes to Nikolay Shmyrev!
The correct code is:
import pocketsphinx as ps
import requests
import json
import sys, os
import subprocess
model_path = ps.get_model_path()
data_path = ps.get_data_path()
print("start")
print(os.getcwd())
subprocess.call("sox -V4 /home/miro/client_audio.wav -r 16000 -c 1 client_audio.wav", shell=True)
config = ps.Decoder.default_config()
config.set_string('-kws', 'keyphrase.list')
config.set_string('-hmm', os.path.join(model_path, 'en-us'))
config.set_string('-dict', os.path.join(model_path, 'cmudict-en-us.dict'))
stream = open("client_audio.wav", "rb")
decoder = ps.Decoder(config)
decoder.start_utt()
while True:
buf = stream.read(1024)
if buf:
decoder.process_raw(buf, False, False)
else:
break
if decoder.hyp() != None:
# print ([(seg.word, seg.prob, seg.start_frame, seg.end_frame) for seg in decoder.seg()])
words=[]
[words.append(seg.word) for seg in decoder.seg()]
print(words)
decoder.end_utt()
decoder.start_utt()