A lightweight yet powerful audio-to-MIDI converter with pitch bend detection
Python audiolightweightmachine-learningmidimusicpitch-detectionpolyphonicpythontranscriptiontypescript
last change prolly, rm magic numbers here too
Saltine0a551bf7c260
diff --git a/basic_pitch/layers/nnaudio.py b/basic_pitch/layers/nnaudio.pyindex 70273ea..3bc9bde 100644--- a/basic_pitch/layers/nnaudio.py+++ b/basic_pitch/layers/nnaudio.py@@ -26,12 +26,26 @@ from typing import Any, List, Optional, Tuple, Unionimport scipy.signal+DEFAULT_BAND_CENTER = 0.5+DEFAULT_KERNEL_LENGTH = 256+DEFAULT_TRANSITION_BANDWIDTH = 0.03+DEFAULT_DTYPE = tf.float32+DEFAULT_WINDOW_BANDWIDTH = 1.5+DEFAULT_CQT_HOP_LENGTH = 512+DEFAULT_CQT_FMIN = 32.70+DEFAULT_CQT_N_BINS = 84+DEFAULT_CQT_BINS_PER_OCTAVE = 12+DEFAULT_CQT_BASIS_NORM = 1+DEFAULT_CQT_WINDOW = "hann"+DEFAULT_CQT_PAD_MODE = "reflect"+DEFAULT_CQT_OUTPUT_FORMAT = "Magnitude"+DEFAULT_LOW_PASS_TRANSITION_BANDWIDTH = 0.001def create_lowpass_filter(- band_center: float = 0.5,- kernel_length: int = 256,- transition_bandwidth: float = 0.03,- dtype: tf.dtypes.DType = tf.float32,+ band_center: float = DEFAULT_BAND_CENTER,+ kernel_length: int = DEFAULT_KERNEL_LENGTH,+ transition_bandwidth: float = DEFAULT_TRANSITION_BANDWIDTH,+ dtype: tf.dtypes.DType = DEFAULT_DTYPE,) -> np.ndarray:"""Calculate the highest frequency we need to preserve and the lowest frequency we allow@@ -106,15 +120,15 @@ def get_early_downsample_params() -> Tuple[Union[float, int], int, float, np.array, bool]:"""Compute downsampling parameters used for early downsampling"""- window_bandwidth = 1.5 # for hann window+ window_bandwidth = DEFAULT_WINDOW_BANDWIDTH # for hann windowfilter_cutoff = fmax_t * (1 + 0.5 * window_bandwidth / Q)sr, hop_length, downsample_factor = early_downsample(sr, hop_length, n_octaves, sr // 2, filter_cutoff)if downsample_factor != 1:earlydownsample = Trueearly_downsample_filter = create_lowpass_filter(band_center=1 / downsample_factor,- kernel_length=256,- transition_bandwidth=0.03,+ kernel_length=DEFAULT_KERNEL_LENGTH,+ transition_bandwidth=DEFAULT_TRANSITION_BANDWIDTH,dtype=dtype,)else:@@ -455,19 +469,19 @@ class CQT2010v2(tf.keras.layers.Layer):def __init__(self,sr: int = 22050,- hop_length: int = 512,- fmin: float = 32.70,+ hop_length: int = DEFAULT_CQT_HOP_LENGTH,+ fmin: float = DEFAULT_CQT_FMIN,fmax: Optional[float] = None,- n_bins: int = 84,+ n_bins: int = DEFAULT_CQT_N_BINS,filter_scale: int = 1,- bins_per_octave: int = 12,+ bins_per_octave: int = DEFAULT_CQT_BINS_PER_OCTAVE,norm: bool = True,- basis_norm: int = 1,- window: str = "hann",- pad_mode: str = "reflect",+ basis_norm: int = DEFAULT_CQT_BASIS_NORM,+ window: str = DEFAULT_CQT_WINDOW,+ pad_mode: str = DEFAULT_CQT_PAD_MODE,earlydownsample: bool = True,trainable: bool = False,- output_format: str = "Magnitude",+ output_format: str = DEFAULT_CQT_OUTPUT_FORMAT,match_torch_exactly: bool = True,):super().__init__()@@ -516,7 +530,11 @@ class CQT2010v2(tf.keras.layers.Layer):# This will be used to calculate filter_cutoff and creating CQT kernelsQ = float(self.filter_scale) / (2 ** (1 / self.bins_per_octave) - 1)- self.lowpass_filter = create_lowpass_filter(band_center=0.5, kernel_length=256, transition_bandwidth=0.001)+ self.lowpass_filter = create_lowpass_filter(+ band_center=DEFAULT_BAND_CENTER,+ kernel_length=DEFAULT_KERNEL_LENGTH,+ transition_bandwidth=DEFAULT_LOW_PASS_TRANSITION_BANDWIDTH,+ )# Calculate num of filter requires for the kernel# n_octaves determines how many resampling requires for the CQT