Skip to content

A lightweight yet powerful audio-to-MIDI converter with pitch bend detection

change minimum_note_len to 127.7ms (11 frames)

Doherty Guiranddc2a29aa2860
diff --git a/basic_pitch/inference.py b/basic_pitch/inference.py
index 3b0032d..2ec4d11 100644
--- a/basic_pitch/inference.py
+++ b/basic_pitch/inference.py
@@ -373,7 +373,7 @@ def predict_and_save(
model_path: Path to load the Keras saved model from. Can be local or on GCS.
onset_threshold: Minimum energy required for an onset to be considered present.
frame_threshold: Minimum energy requirement for a frame to be considered present.
- minimum_note_length: The minimum allowed note length in frames.
+ minimum_note_length: The minimum allowed note length in milliseconds.
minimum_freq: Minimum allowed output frequency, in Hz. If None, all frequencies are used.
maximum_freq: Maximum allowed output frequency, in Hz. If None, all frequencies are used.
multiple_pitch_bends: If True, allow overlapping notes in midi file to have pitch bends.
diff --git a/basic_pitch/predict.py b/basic_pitch/predict.py
index 075d141..76bb886 100644
--- a/basic_pitch/predict.py
+++ b/basic_pitch/predict.py
@@ -73,7 +73,7 @@ def main() -> None:
parser.add_argument(
"--minimum-note-length",
type=float,
- default=58,
+ default=127.70,
help="The minimum allowed note length, in miliseconds.",
)
parser.add_argument(