optimize(f0): move some f0s into rvc.f0

rvc/f0/__init__.py

@@ -0,0 +1,4 @@
+from .dio import Dio
+from .harvest import Harvest
+from .pm import PM
+from .f0 import F0Predictor

rvc/f0/dio.py

@@ -6,27 +6,15 @@ import pyworld
 from .f0 import F0Predictor
 
 
-class DioF0Predictor(F0Predictor):
+class Dio(F0Predictor):
     def __init__(self, hop_length=512, f0_min=50, f0_max=1100, sampling_rate=44100):
         super().__init__(hop_length, f0_min, f0_max, sampling_rate)
 
-    def compute_f0(self, wav: np.ndarray[Any, np.dtype], p_len: Optional[int] = None):
-        if p_len is None:
-            p_len = wav.shape[0] // self.hop_length
-        f0, t = pyworld.dio(
-            wav.astype(np.double),
-            fs=self.sampling_rate,
-            f0_floor=self.f0_min,
-            f0_ceil=self.f0_max,
-            frame_period=1000 * self.hop_length / self.sampling_rate,
-        )
-        f0 = pyworld.stonemask(wav.astype(np.double), f0, t, self.sampling_rate)
-        for index, pitch in enumerate(f0):
-            f0[index] = round(pitch, 1)
-        return self.interpolate_f0(self.resize_f0(f0, p_len))[0]
-
-    def compute_f0_uv(
-        self, wav: np.ndarray[Any, np.dtype], p_len: Optional[int] = None
+    def compute_f0(
+        self,
+        wav: np.ndarray[Any, np.dtype],
+        p_len: Optional[int] = None,
+        filter_radius: Optional[int] = None,
     ):
         if p_len is None:
             p_len = wav.shape[0] // self.hop_length
@@ -40,4 +28,4 @@ class DioF0Predictor(F0Predictor):
         f0 = pyworld.stonemask(wav.astype(np.double), f0, t, self.sampling_rate)
         for index, pitch in enumerate(f0):
             f0[index] = round(pitch, 1)
-        return self.interpolate_f0(self.resize_f0(f0, p_len))
+        return self.interpolate_f0(self.resize_f0(f0, p_len))[0]

rvc/f0/f0.py

@@ -11,11 +11,10 @@ class F0Predictor(object):
         self.sampling_rate = sampling_rate
 
     def compute_f0(
-        self, wav: np.ndarray[Any, np.dtype], p_len: Optional[int] = None
-    ): ...
-
-    def compute_f0_uv(
-        self, wav: np.ndarray[Any, np.dtype], p_len: Optional[int] = None
+        self,
+        wav: np.ndarray[Any, np.dtype],
+        p_len: Optional[int] = None,
+        filter_radius: Optional[int] = None,
     ): ...
 
     def interpolate_f0(self, f0: np.ndarray[Any, np.dtype]):

rvc/f0/harvest.py

@@ -2,38 +2,31 @@ from typing import Any, Optional
 
 import numpy as np
 import pyworld
+from scipy import signal
 
 from .f0 import F0Predictor
 
 
-class HarvestF0Predictor(F0Predictor):
+class Harvest(F0Predictor):
     def __init__(self, hop_length=512, f0_min=50, f0_max=1100, sampling_rate=44100):
         super().__init__(hop_length, f0_min, f0_max, sampling_rate)
 
-    def compute_f0(self, wav: np.ndarray[Any, np.dtype], p_len: Optional[int] = None):
-        if p_len is None:
-            p_len = wav.shape[0] // self.hop_length
-        f0, t = pyworld.harvest(
-            wav.astype(np.double),
-            fs=self.sampling_rate,
-            f0_ceil=self.f0_max,
-            f0_floor=self.f0_min,
-            frame_period=1000 * self.hop_length / self.sampling_rate,
-        )
-        f0 = pyworld.stonemask(wav.astype(np.double), f0, t, self.fs)
-        return self.interpolate_f0(self.resize_f0(f0, p_len))[0]
-
-    def compute_f0_uv(
-        self, wav: np.ndarray[Any, np.dtype], p_len: Optional[int] = None
+    def compute_f0(
+        self,
+        wav: np.ndarray[Any, np.dtype],
+        p_len: Optional[int] = None,
+        filter_radius: Optional[int] = None,
     ):
         if p_len is None:
             p_len = wav.shape[0] // self.hop_length
         f0, t = pyworld.harvest(
             wav.astype(np.double),
             fs=self.sampling_rate,
-            f0_floor=self.f0_min,
             f0_ceil=self.f0_max,
+            f0_floor=self.f0_min,
             frame_period=1000 * self.hop_length / self.sampling_rate,
         )
         f0 = pyworld.stonemask(wav.astype(np.double), f0, t, self.sampling_rate)
-        return self.interpolate_f0(self.resize_f0(f0, p_len))
+        if filter_radius is not None and filter_radius > 2:
+            f0 = signal.medfilt(f0, 3)
+        return self.interpolate_f0(self.resize_f0(f0, p_len))[0]

rvc/f0/pm.py

@@ -0,0 +1,39 @@
+from typing import Any, Optional
+
+import numpy as np
+import parselmouth
+
+from .f0 import F0Predictor
+
+
+class PM(F0Predictor):
+    def __init__(self, hop_length=512, f0_min=50, f0_max=1100, sampling_rate=44100):
+        super().__init__(hop_length, f0_min, f0_max, sampling_rate)
+
+    def compute_f0(
+        self,
+        wav: np.ndarray[Any, np.dtype],
+        p_len: Optional[int] = None,
+        filter_radius: Optional[int] = None,
+    ):
+        x = wav
+        if p_len is None:
+            p_len = x.shape[0] // self.hop_length
+        else:
+            assert abs(p_len - x.shape[0] // self.hop_length) < 4, "pad length error"
+        time_step = self.hop_length / self.sampling_rate * 1000
+        f0 = (
+            parselmouth.Sound(x, self.sampling_rate)
+            .to_pitch_ac(
+                time_step=time_step / 1000,
+                voicing_threshold=0.6,
+                pitch_floor=self.f0_min,
+                pitch_ceiling=self.f0_max,
+            )
+            .selected_array["frequency"]
+        )
+
+        pad_size = (p_len - len(f0) + 1) // 2
+        if pad_size > 0 or p_len - len(f0) - pad_size > 0:
+            f0 = np.pad(f0, [[pad_size, p_len - len(f0) - pad_size]], mode="constant")
+        return self.interpolate_f0(f0)[0]

rvc/onnx/f0/__init__.py

@@ -1,4 +0,0 @@
-from .dio import DioF0Predictor
-from .harvest import HarvestF0Predictor
-from .pm import PMF0Predictor
-from .f0 import F0Predictor

rvc/onnx/f0/pm.py

@@ -1,61 +0,0 @@
-from typing import Any, Optional
-
-import numpy as np
-import parselmouth
-
-from .f0 import F0Predictor
-
-
-class PMF0Predictor(F0Predictor):
-    def __init__(self, hop_length=512, f0_min=50, f0_max=1100, sampling_rate=44100):
-        super().__init__(hop_length, f0_min, f0_max, sampling_rate)
-
-    def compute_f0(self, wav: np.ndarray[Any, np.dtype], p_len: Optional[int] = None):
-        x = wav
-        if p_len is None:
-            p_len = x.shape[0] // self.hop_length
-        else:
-            assert abs(p_len - x.shape[0] // self.hop_length) < 4, "pad length error"
-        time_step = self.hop_length / self.sampling_rate * 1000
-        f0 = (
-            parselmouth.Sound(x, self.sampling_rate)
-            .to_pitch_ac(
-                time_step=time_step / 1000,
-                voicing_threshold=0.6,
-                pitch_floor=self.f0_min,
-                pitch_ceiling=self.f0_max,
-            )
-            .selected_array["frequency"]
-        )
-
-        pad_size = (p_len - len(f0) + 1) // 2
-        if pad_size > 0 or p_len - len(f0) - pad_size > 0:
-            f0 = np.pad(f0, [[pad_size, p_len - len(f0) - pad_size]], mode="constant")
-        f0, uv = self.interpolate_f0(f0)
-        return f0
-
-    def compute_f0_uv(
-        self, wav: np.ndarray[Any, np.dtype], p_len: Optional[int] = None
-    ):
-        x = wav
-        if p_len is None:
-            p_len = x.shape[0] // self.hop_length
-        else:
-            assert abs(p_len - x.shape[0] // self.hop_length) < 4, "pad length error"
-        time_step = self.hop_length / self.sampling_rate * 1000
-        f0 = (
-            parselmouth.Sound(x, self.sampling_rate)
-            .to_pitch_ac(
-                time_step=time_step / 1000,
-                voicing_threshold=0.6,
-                pitch_floor=self.f0_min,
-                pitch_ceiling=self.f0_max,
-            )
-            .selected_array["frequency"]
-        )
-
-        pad_size = (p_len - len(f0) + 1) // 2
-        if pad_size > 0 or p_len - len(f0) - pad_size > 0:
-            f0 = np.pad(f0, [[pad_size, p_len - len(f0) - pad_size]], mode="constant")
-        f0, uv = self.interpolate_f0(f0)
-        return f0, uv

rvc/onnx/infer.py

@@ -5,10 +5,10 @@ import librosa
 import numpy as np
 import onnxruntime
 
-from .f0 import (
-    PMF0Predictor,
-    HarvestF0Predictor,
-    DioF0Predictor,
+from rvc.f0 import (
+    PM,
+    Harvest,
+    Dio,
     F0Predictor,
 )
 
@@ -52,9 +52,9 @@ class ContentVec(Model):
 
 
 predictors: typing.Dict[str, F0Predictor] = {
-    "pm": PMF0Predictor,
-    "harvest": HarvestF0Predictor,
-    "dio": DioF0Predictor,
+    "pm": PM,
+    "harvest": Harvest,
+    "dio": Dio,
 }