MycroftAI
/
mycroft-precise
mirror of https://github.com/MycroftAI/mycroft-precise.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Use vectorize_raw instead of rewriting mfcc call

pull/19/head
Matthew Scholefield 2018-07-05 13:37:50 -05:00
parent 0da11b33bd
commit 779ddd2d02
1 changed files with 3 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
precise/network_runner.py
View File

@ -22,7 +22,7 @@ import numpy as np
from precise.model import load_precise_model from precise.model import load_precise_model
from precise.params import inject_params from precise.params import inject_params
from precise.util import buffer_to_audio from precise.util import buffer_to_audio
from precise.vectorization import add_deltas from precise.vectorization import vectorize_raw
class Runner(metaclass=ABCMeta): class Runner(metaclass=ABCMeta):
@ -90,7 +90,6 @@ class Listener:
self.chunk_size = chunk_size self.chunk_size = chunk_size
runner_cls = runner_cls or self.find_runner(model_name) runner_cls = runner_cls or self.find_runner(model_name)
self.runner = runner_cls(model_name) self.runner = runner_cls(model_name)
self.mfcc = import_module('speechpy.feature').mfcc
@staticmethod @staticmethod
def find_runner(model_name: str) -> Type[Runner]: def find_runner(model_name: str) -> Type[Runner]:
@ -122,11 +121,10 @@ class Listener:
self.window_audio = np.concatenate((self.window_audio, buffer_audio)) self.window_audio = np.concatenate((self.window_audio, buffer_audio))
if len(self.window_audio) >= self.pr.window_samples: if len(self.window_audio) >= self.pr.window_samples:
new_features = self.mfcc(self.window_audio, self.pr.sample_rate, self.pr.window_t, new_features = vectorize_raw(self.window_audio)
self.pr.hop_t, self.pr.n_mfcc, self.pr.n_filt, self.pr.n_fft)
self.window_audio = self.window_audio[len(new_features) * self.pr.hop_samples:] self.window_audio = self.window_audio[len(new_features) * self.pr.hop_samples:]
if len(new_features) > len(self.mfccs): if len(new_features) > len(self.mfccs):
new_features = new_features[-len(self.mfccs):] new_features = new_features[-len(self.mfccs):]
self.mfccs = np.concatenate((self.mfccs[len(new_features):], new_features)) self.mfccs = np.concatenate((self.mfccs[len(new_features):], new_features))
return self.runner.run(add_deltas(self.mfccs)) return self.runner.run(self.mfccs)