fix

libdetectability/detectability.py

@@ -6,7 +6,14 @@
 from .internal.outer_middle_ear_filter import outer_middle_ear_filter
 
 class Detectability:
-    def __init__(self, frame_size=2048, sampling_rate=48000, taps=64, dbspl=94.0, spl=1.0, relax_threshold=False, norm="backward"):
+    def __init__(self, 
+                 frame_size=2048, 
+                 sampling_rate=48000, 
+                 taps=64, 
+                 dbspl=94.0, 
+                 spl=1.0, 
+                 relax_threshold=False, 
+                 norm="backward"):
         assert frame_size % 2 == 0, "only evenly-sized frames are supported"
         self.frame_size = frame_size
         self.freq_size = frame_size // 2 + 1
@@ -18,7 +25,9 @@ def __init__(self, frame_size=2048, sampling_rate=48000, taps=64, dbspl=94.0, sp
 
         # prealloc
         self.g = np.power(np.abs(gammatone_filterbank(self.taps, self.frame_size, self.sampling_rate)), 2.0)
-        self.h = np.power(np.abs(outer_middle_ear_filter(self.frame_size, self.spl, self.dbspl, self.sampling_rate, relax_threshold=relax_threshold)), 2.0)
+        self.h = np.power(np.abs(
+            outer_middle_ear_filter(self.frame_size, self.spl, self.dbspl, self.sampling_rate, relax_threshold=relax_threshold)
+            ), 2.0)
         self.leff = min(float(self.frame_size) / float(sampling_rate) / 0.30, 1.0)
 
         # calibration
@@ -53,8 +62,10 @@ def _detectability(self, s, m, cs, ca):
         return cs * self.leff * (s / (m + ca)).sum()
 
     def frame(self, reference, test):
-        assert reference.size == self.frame_size and test.size == self.frame_size, f"input frame size different the specified upon construction"
-        assert len(reference.shape) == 1 and len(test.shape) == 1, f"only support for one-dimensional inputs"
+        assert reference.size == self.frame_size and test.size == self.frame_size, \
+                f"input frame size different the specified upon construction"
+        assert len(reference.shape) == 1 and len(test.shape) == 1, \
+                f"only support for one-dimensional inputs"
 
         e = self._spectrum(test - reference)
         x = self._spectrum(reference)
@@ -64,8 +75,10 @@ def frame(self, reference, test):
         return self._detectability(e, x, self.cs, self.ca)
 
     def gain(self, reference):
-        assert reference.size == self.frame_size, f"input frame size different the specified upon construction"
-        assert len(reference.shape) == 1, f"only support for one-dimensional inputs"
+        assert reference.size == self.frame_size, \
+                f"input frame size different the specified upon construction"
+        assert len(reference.shape) == 1, \
+                f"only support for one-dimensional inputs"
 
         x = self._spectrum(reference)
         x = self._masker_power_array(x)
@@ -75,7 +88,16 @@ def gain(self, reference):
         return np.sqrt(G.sum(axis=0))
 
 class DetectabilityLoss(tc.nn.Module):
-    def __init__(self, frame_size=2048, sampling_rate=48000, taps=32, dbspl=94.0, spl=1.0, relax_threshold=True, norm = "backward", reduction="meanlog", eps=1e-8):
+    def __init__(self, 
+                 frame_size=2048, 
+                 sampling_rate=48000, 
+                 taps=64, 
+                 dbspl=94.0, 
+                 spl=1.0, 
+                 relax_threshold=False, 
+                 norm = "backward", 
+                 reduction="meanlog", 
+                 eps=1e-8):
         super(DetectabilityLoss, self).__init__()
         self.detectability = Detectability(frame_size=frame_size, sampling_rate=sampling_rate, taps=taps, dbspl=dbspl, \
                 spl=spl, relax_threshold=relax_threshold, norm=norm)

libdetectability/test.py

@@ -18,6 +18,8 @@ def test_cost():
     y = tc.concatenate((y, y))
 
     tcv = tcd.frame(x, y)
+    print(tcv)
+    print(npv)
 
     assert tcv[0] == pytest.approx(npv)
     assert tcv[1] == pytest.approx(npv)