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

Make mfcc generation use mfcc_size instead of feature_size 

feature_size could be different.
pull/19/head
Matthew Scholefield 2018-07-05 13:34:15 -05:00
parent 603385291f
commit 0da11b33bd
1 changed files with 7 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
precise/network_runner.py
View File

@ -22,6 +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
class Runner(metaclass=ABCMeta): class Runner(metaclass=ABCMeta):
@ -85,7 +86,7 @@ class Listener:
def __init__(self, model_name: str, chunk_size: int = -1, runner_cls: type = None): def __init__(self, model_name: str, chunk_size: int = -1, runner_cls: type = None):
self.window_audio = np.array([]) self.window_audio = np.array([])
self.pr = inject_params(model_name) self.pr = inject_params(model_name)
self.features = np.zeros((self.pr.n_features, self.pr.feature_size)) self.mfccs = np.zeros((self.pr.n_features, self.pr.n_mfcc))
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)
@ -104,7 +105,7 @@ class Listener:
def clear(self): def clear(self):
self.window_audio = np.array([]) self.window_audio = np.array([])
self.features = np.zeros((self.pr.n_features, self.pr.feature_size)) self.mfccs = np.zeros((self.pr.n_features, self.pr.n_mfcc))
def update(self, stream: Union[BinaryIO, np.ndarray, bytes]) -> float: def update(self, stream: Union[BinaryIO, np.ndarray, bytes]) -> float:
if isinstance(stream, np.ndarray): if isinstance(stream, np.ndarray):
@ -124,8 +125,8 @@ class Listener:
new_features = self.mfcc(self.window_audio, self.pr.sample_rate, self.pr.window_t, new_features = self.mfcc(self.window_audio, self.pr.sample_rate, self.pr.window_t,
self.pr.hop_t, self.pr.n_mfcc, self.pr.n_filt, self.pr.n_fft) 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.features): if len(new_features) > len(self.mfccs):
new_features = new_features[-len(self.features):] new_features = new_features[-len(self.mfccs):]
self.features = np.concatenate((self.features[len(new_features):], new_features)) self.mfccs = np.concatenate((self.mfccs[len(new_features):], new_features))
return self.runner.run(self.features) return self.runner.run(add_deltas(self.mfccs))