[cm] Decreasing delay

neutone_sdk/cached_mel_spec.py

@@ -45,9 +45,12 @@ def __init__(
             n_mels=n_mels,
             center=False,
         )
-        self.queue = CircularInplaceTensorQueue(n_ch, self.n_fft, use_debug_mode)
-        self.register_buffer("padding", tr.zeros((n_ch, self.n_fft)))
-        self.queue.push(self.padding)
+        self.padding_n_samples = self.n_fft - self.hop_len
+        self.cache = CircularInplaceTensorQueue(
+            n_ch, self.padding_n_samples, use_debug_mode
+        )
+        self.register_buffer("padding", tr.zeros((n_ch, self.padding_n_samples)))
+        self.cache.push(self.padding)
 
     def forward(self, x: Tensor) -> Tensor:
         if self.use_debug_mode:
@@ -61,30 +64,30 @@ def forward(self, x: Tensor) -> Tensor:
         padded_x = tr.cat([self.padding, x], dim=1)
         # log.info(f"padded_x = {padded_x}")
         spec = self.mel_spec(padded_x)
-        spec = spec[:, :, :n_frames]
+        spec = spec[:, :, -n_frames:]
 
-        lookahead_idx = min(n_samples, self.n_fft)
-        self.queue.push(x[:, -lookahead_idx:])
-        self.queue.fill(self.padding)
+        padding_idx = min(n_samples, self.padding_n_samples)
+        self.cache.push(x[:, -padding_idx:])
+        self.cache.fill(self.padding)
         return spec
 
     @tr.jit.export
     def get_delay_samples(self) -> int:
-        return self.n_fft // 2
+        return (self.n_fft // 2) - self.hop_len
 
     @tr.jit.export
     def get_delay_frames(self) -> int:
         return self.get_delay_samples() // self.hop_len
 
     @tr.jit.export
     def reset(self) -> None:
-        self.queue.reset()
+        self.cache.reset()
         self.padding.zero_()
-        self.queue.push(self.padding)
+        self.cache.push(self.padding)
 
 
 def test_cached_mel_spec():
-    tr.set_printoptions(precision=2)
+    tr.set_printoptions(precision=1)
     tr.random.manual_seed(0)
 
     sr = 44100
@@ -127,7 +130,9 @@ def test_cached_mel_spec():
         curr_idx += chunk_size
     if curr_idx < total_n_samples:
         chunks.append(audio[:, curr_idx:])
-    chunks.append(tr.zeros(n_ch, cached_mel_spec.n_fft))
+    chunks.append(
+        tr.zeros(n_ch, cached_mel_spec.get_delay_samples() + cached_mel_spec.hop_len)
+    )
 
     spec_chunks = []
     for chunk in chunks: