Add sign triggering

corrscope/gui/__init__.py

@@ -854,6 +854,7 @@ class ChannelModel(qc.QAbstractTableModel):
         Column("trigger__responsiveness", float, None, "Buffer\nResponsiveness"),
         Column("trigger__edge_direction", plus_minus_one, None),
         Column("trigger__edge_strength", float, None),
+        Column("trigger__sign_strength", float, None),
         Column("trigger__slope_strength", float, None),
         Column("trigger__slope_width", float, None),
     ]

corrscope/gui/view_mainwindow.py

@@ -272,6 +272,22 @@ class MainWindow(QWidget):
         tr = self.tr
 
         with self.add_tab_stretch(s, tr("&Trigger"), layout=QVBoxLayout) as tab:
+            with append_widget(
+                s,
+                QGroupBox,
+                title=tr("Input Data Preprocessing"),
+                layout=QFormLayout,
+            ):
+                with add_row(
+                    s,
+                    tr("Sign Triggering\n(for triangle waves)"),
+                    BoundDoubleSpinBox,
+                    name="trigger__sign_strength",
+                    minimum=0,
+                    singleStep=0.25,
+                ):
+                    pass
+
             with append_widget(
                 s, QGroupBox, title=tr("Wave Alignment"), layout=QFormLayout
             ):
@@ -291,14 +307,6 @@ class MainWindow(QWidget):
                     singleStep=0.1,
                 ):
                     pass
-                with add_row(
-                    s,
-                    tr("Mean Responsiveness"),
-                    BoundDoubleSpinBox,
-                    name="trigger__mean_responsiveness",
-                ) as w:  # type: BoundDoubleSpinBox
-                    w.setMaximum(1.0)
-                    w.setSingleStep(0.1)
                 with add_row(s, BoundCheckBox, Both) as (self.trigger__pitch_tracking):
                     assert isinstance(self.trigger__pitch_tracking, QWidget)
 

corrscope/triggers.py

@@ -9,12 +9,22 @@ import corrscope.utils.scipy.windows as windows
 from corrscope.config import KeywordAttrs, CorrError, Alias, with_units
 from corrscope.spectrum import SpectrumConfig, DummySpectrum, LogFreqSpectrum
 from corrscope.util import find, obj_name, iround
-from corrscope.utils.trigger_util import get_period, normalize_buffer, lerp
+from corrscope.utils.trigger_util import (
+    get_period,
+    normalize_buffer,
+    lerp,
+    MIN_AMPLITUDE,
+)
 from corrscope.utils.windows import leftpad, midpad, rightpad
 from corrscope.wave import FLOAT
 
 if TYPE_CHECKING:
     from corrscope.wave import Wave
+    from corrscope.renderer import RendererFrontend
+
+
+def abs_max(data, offset=0):
+    return np.amax(np.abs(data)) + offset
 
 
 # Abstract classes
@@ -194,7 +204,6 @@ class PerFrameCache:
     # NOTE: period is a *non-subsampled* period.
     # The period of subsampled data must be multiplied by stride.
     period: Optional[int] = None
-    sum: Optional[float] = None  # Only used in branch 'rewrite-area-trigger'
     mean: Optional[float] = None
 
 
@@ -238,9 +247,8 @@ class CorrelationTriggerConfig(
     " buffer_falloff ",
 ):
     # get_trigger()
-    mean_responsiveness: float = 0.05
-
     # Edge/area finding
+    sign_strength: float = 0
     edge_strength: float
 
     # Slope detection
@@ -280,7 +288,6 @@ class CorrelationTriggerConfig(
         validate_param(self, "slope_width", 0, 0.5)
 
         validate_param(self, "responsiveness", 0, 1)
-        validate_param(self, "mean_responsiveness", 0, 1)
         # TODO trigger_falloff >= 0
         validate_param(self, "buffer_falloff", 0, np.inf)
 
@@ -339,7 +346,6 @@ class CorrelationTrigger(MainTrigger):
         self._prev_window: Optional[np.ndarray] = None
         self._prev_slope_finder: Optional[np.ndarray] = None
 
-        self._prev_mean: float = 0.0
         self._prev_trigger: int = 0
 
         # (mutable) Log-scaled spectrum
@@ -443,14 +449,12 @@ class CorrelationTrigger(MainTrigger):
         # Get data (1D, downmixed to mono)
         stride = self._stride
         data = self._wave.get_around(index, N, stride)
-        cache.sum = np.add.reduce(data)
 
         # Update data-mean estimate
-        raw_mean = cache.sum / N
+        if cfg.sign_strength != 0:
-        self._prev_mean = cache.mean = lerp(
+            signs = sign_times_peak(data)
-            self._prev_mean, raw_mean, cfg.mean_responsiveness
+            data += cfg.sign_strength * signs
-        )
+            data -= np.add.reduce(data) / N
-        data -= cache.mean
 
         # Window data
         period = get_period(data)
@@ -498,19 +502,28 @@ class CorrelationTrigger(MainTrigger):
         else:
             radius = None
 
-        peak_offset = self.correlate_offset(data, prev_buffer, radius)
+        score = correlate_offset(data, prev_buffer, radius)
+        peak_offset = score.peak
         trigger = index + (stride * peak_offset)
 
-        # Apply post trigger (before updating correlation buffer)
+        del data
+
         if self.post:
+            new_data = self._wave.get_around(trigger, N, stride)
+            cache.mean = np.add.reduce(new_data) / N
+
+            # Apply post trigger (before updating correlation buffer)
             trigger = self.post.get_trigger(trigger, cache)
 
         # Avoid time traveling backwards.
         self._prev_trigger = trigger = max(trigger, self._prev_trigger)
 
         # Update correlation buffer (distinct from visible area)
-        aligned = self._wave.get_around(trigger, self._buffer_nsamp, stride)
+        aligned = self._wave.get_around(trigger, N, stride)
+        if cache.mean is None:
+            cache.mean = np.add.reduce(aligned) / N
         self._update_buffer(aligned, cache)
+
         self.frames_since_spectrum += 1
 
         return trigger
@@ -548,13 +561,13 @@ class CorrelationTrigger(MainTrigger):
             boost_y = 1.0
 
         # If we want to double pitch...
-        resample_notes = self.correlate_offset(
+        resample_notes = correlate_offset(
             spectrum,
             prev_spectrum,
             scfg.max_notes_to_resample,
             boost_x=boost_x,
             boost_y=boost_y,
-        )
+        ).peak
         if resample_notes != 0:
             # we must divide sampling rate by 2.
             new_len = iround(N / 2 ** (resample_notes / scfg.notes_per_octave))
@@ -566,49 +579,6 @@ class CorrelationTrigger(MainTrigger):
             # assert len(self._buffer) == new_len
             self._buffer = midpad(self._buffer, N)
 
-    @staticmethod
-    def correlate_offset(
-        data: np.ndarray,
-        prev_buffer: np.ndarray,
-        radius: Optional[int],
-        boost_x: int = 0,
-        boost_y: float = 1.0,
-    ) -> int:
-        """
-        This is confusing.
-
-        If data index < optimal, data will be too far to the right,
-        and we need to `index += positive`.
-        - The peak will appear near the right of `data`.
-
-        Either we must slide prev_buffer to the right,
-        or we must slide data to the left (by sliding index to the right):
-        - correlate(data, prev_buffer)
-        - trigger = index + peak_offset
-        """
-        N = len(data)
-        corr = signal.correlate(data, prev_buffer)  # returns double, not single/FLOAT
-        Ncorr = 2 * N - 1
-        assert len(corr) == Ncorr
-
-        # Find optimal offset
-        mid = N - 1
-
-        if radius is not None:
-            left = max(mid - radius, 0)
-            right = min(mid + radius + 1, Ncorr)
-
-            corr = corr[left:right]
-            mid = mid - left
-
-        # Prioritize part of it.
-        corr[mid + boost_x : mid + boost_x + 1] *= boost_y
-
-        # argmax(corr) == mid + peak_offset == (data >> peak_offset)
-        # peak_offset == argmax(corr) - mid
-        peak_offset = np.argmax(corr) - mid  # type: int
-        return peak_offset
-
     def _is_window_invalid(self, period: int) -> Union[bool, float]:
         """ Returns number of semitones,
         if pitch has changed more than `recalc_semitones`. """
@@ -657,6 +627,75 @@ class CorrelationTrigger(MainTrigger):
         self._buffer = lerp(self._buffer, data, responsiveness)
 
 
+@attr.dataclass
+class CorrelationResult:
+    peak: int
+    corr: np.ndarray
+
+
+def correlate_offset(
+    data: np.ndarray,
+    prev_buffer: np.ndarray,
+    radius: Optional[int],
+    boost_x: int = 0,
+    boost_y: float = 1.0,
+) -> CorrelationResult:
+    """
+    This is confusing.
+
+    If data index < optimal, data will be too far to the right,
+    and we need to `index += positive`.
+    - The peak will appear near the right of `data`.
+
+    Either we must slide prev_buffer to the right,
+    or we must slide data to the left (by sliding index to the right):
+    - correlate(data, prev_buffer)
+    - trigger = index + peak_offset
+    """
+    N = len(data)
+    corr = signal.correlate(data, prev_buffer)  # returns double, not single/FLOAT
+    Ncorr = 2 * N - 1
+    assert len(corr) == Ncorr
+
+    # Find optimal offset
+    mid = N - 1
+
+    if radius is not None:
+        left = max(mid - radius, 0)
+        right = min(mid + radius + 1, Ncorr)
+
+        corr = corr[left:right]
+        mid = mid - left
+
+    # Prioritize part of it.
+    corr[mid + boost_x : mid + boost_x + 1] *= boost_y
+
+    # argmax(corr) == mid + peak_offset == (data >> peak_offset)
+    # peak_offset == argmax(corr) - mid
+    peak_offset = np.argmax(corr) - mid  # type: int
+    return CorrelationResult(peak_offset, corr)
+
+
+SIGN_AMPLIFICATION = 1000
+
+
+def sign_times_peak(data: np.ndarray) -> np.ndarray:
+    """
+    Computes peak = max(abs(data)).
+    Returns `peak` for positive parts of data, and `-peak` for negative parts,
+    and heavily amplifies parts of the wave near zero.
+    """
+    data = data.copy()
+
+    peak = abs_max(data)
+    data *= SIGN_AMPLIFICATION / (peak + MIN_AMPLITUDE)
+
+    sign_data = np.tanh(data)
+    sign_data *= peak
+
+    return sign_data
+
+
 #### Post-processing triggers
 # ZeroCrossingTrigger
 

tests/test_trigger.py

@@ -13,6 +13,7 @@ from corrscope.triggers import (
     PerFrameCache,
     ZeroCrossingTriggerConfig,
     SpectrumConfig,
+    correlate_offset,
 )
 from corrscope.wave import Wave
 
@@ -220,12 +221,11 @@ def test_correlate_offset():
     """
 
     np.random.seed(31337)
-    correlate_offset = CorrelationTrigger.correlate_offset
 
     # Ensure autocorrelation on random data returns peak at 0.
     N = 100
     spectrum = np.random.random(N)
-    assert correlate_offset(spectrum, spectrum, 12) == 0
+    assert correlate_offset(spectrum, spectrum, 12).peak == 0
 
     # Ensure cross-correlation of time-shifted impulses works.
     # Assume wave where y=[i==99].
@@ -236,15 +236,17 @@ def test_correlate_offset():
 
     # We need to slide `left` to the right by 10 samples, and vice versa.
     for radius in [None, 12]:
-        assert correlate_offset(data=left, prev_buffer=right, radius=radius) == 10
+        assert correlate_offset(data=left, prev_buffer=right, radius=radius).peak == 10
-        assert correlate_offset(data=right, prev_buffer=left, radius=radius) == -10
+        assert correlate_offset(data=right, prev_buffer=left, radius=radius).peak == -10
 
     # The correlation peak at zero-offset is small enough for boost_x to be returned.
     boost_y = 1.5
     ones = np.ones(N)
     for boost_x in [6, -6]:
         assert (
-            correlate_offset(ones, ones, radius=9, boost_x=boost_x, boost_y=boost_y)
+            correlate_offset(
+                ones, ones, radius=9, boost_x=boost_x, boost_y=boost_y
+            ).peak
             == boost_x
         )