batdetect2/docs/source/explanation/preprocessing-consistency.md
2026-03-28 19:42:09 +00:00

1010 B

Preprocessing consistency

Preprocessing consistency is one of the biggest factors behind stable model performance.

Why consistency matters

The detector is trained on spectrograms produced by a specific preprocessing pipeline. If inference uses different settings, the model can see a shifted input distribution and performance may drop.

Typical mismatch sources:

  • sample-rate differences,
  • changed frequency crop,
  • changed STFT window/hop,
  • changed spectrogram transforms.

Practical implication

When possible, keep preprocessing settings aligned between:

  • training,
  • evaluation,
  • deployment inference.

If you intentionally change preprocessing, treat this as a new experiment and re-validate on reviewed local data.

  • Configure audio preprocessing: {doc}../how_to/configure-audio-preprocessing
  • Configure spectrogram preprocessing: {doc}../how_to/configure-spectrogram-preprocessing
  • Preprocessing config reference: {doc}../reference/preprocessing-config