Update audio2text/a2t.py
Browse files- audio2text/a2t.py +6 -5
audio2text/a2t.py
CHANGED
@@ -19,6 +19,12 @@ class A2T:
|
|
19 |
return transcribed_text
|
20 |
|
21 |
def __preprocces(self, raw: np.ndarray, sampling_rate: int):
|
|
|
|
|
|
|
|
|
|
|
|
|
22 |
chunk = raw.astype(np.float32, order='C') / 32768.0
|
23 |
|
24 |
print(f"Chunk : {chunk} max chunk : {np.max(chunk)}")
|
@@ -26,11 +32,6 @@ class A2T:
|
|
26 |
if len(chunk.shape) > 1:
|
27 |
chunk = librosa.to_mono(chunk.T)
|
28 |
|
29 |
-
if sampling_rate != SAMPLING_RATE:
|
30 |
-
chunk = librosa.resample(chunk, orig_sr=sampling_rate, target_sr=SAMPLING_RATE)
|
31 |
-
|
32 |
-
print(f"Sampling rate : {chunk} max chunk : {np.max(chunk)}")
|
33 |
-
|
34 |
chunk = chunk[:SAMPLING_RATE*LIMIT]
|
35 |
|
36 |
print(f"Chunk cut : {chunk} max chunk : {np.max(chunk)}")
|
|
|
19 |
return transcribed_text
|
20 |
|
21 |
def __preprocces(self, raw: np.ndarray, sampling_rate: int):
|
22 |
+
|
23 |
+
if sampling_rate != SAMPLING_RATE:
|
24 |
+
raw = librosa.resample(raw, orig_sr=sampling_rate, target_sr=SAMPLING_RATE)
|
25 |
+
|
26 |
+
print(f"Sampling rate : {raw} max chunk : {np.max(raw)}")
|
27 |
+
|
28 |
chunk = raw.astype(np.float32, order='C') / 32768.0
|
29 |
|
30 |
print(f"Chunk : {chunk} max chunk : {np.max(chunk)}")
|
|
|
32 |
if len(chunk.shape) > 1:
|
33 |
chunk = librosa.to_mono(chunk.T)
|
34 |
|
|
|
|
|
|
|
|
|
|
|
35 |
chunk = chunk[:SAMPLING_RATE*LIMIT]
|
36 |
|
37 |
print(f"Chunk cut : {chunk} max chunk : {np.max(chunk)}")
|