eml-signal 0.1.0

Bounded signal-processing kernels and contract metadata for EML-style workflows.

eml-signal

eml-signal is a small pure-Python package of bounded signal-processing kernels and contract metadata for EML-style workflows.

The v0 package focuses on deterministic list-based kernels that are easy to test, inspect, and use in synthetic examples for Monowave, Field Kernel Lab, and EML course material.

It does not process copyrighted audio assets, make wellness or medical claims, or claim production DSP or real-time performance.

Kernels

• envelope_follower
• one_pole_lowpass
• band_energy_toy
• moving_average_filter
• zero_crossing_rate
• highpass_difference
• convolution_1d

Each kernel returns a SignalResult with values or features, an operation_count, and a compact contract dictionary.

CLI

eml-signal --help
eml-signal demo envelope
eml-signal demo lowpass
eml-signal demo band-energy
eml-signal validate examples/signal_contract.json

Python

from eml_signal import envelope_follower, one_pole_lowpass

signal = [0.0, 0.25, -0.5, 1.0, -0.25]
envelope = envelope_follower(signal, alpha=0.2)
smooth = one_pole_lowpass(signal, alpha=0.25)

print(envelope.operation_count)
print(smooth.values)

Monowave Bridge

The optional Monowave bridge maps bounded synthetic signals into the same broad feature names used by Monowave scenes:

• input_level
• bass_drive
• mid_drive
• treble_drive

from eml_signal import monowave_feature_summary

features = monowave_feature_summary([0.0, 0.25, -0.5, 0.75])
print(features["features"])

This bridge is for synthetic fixtures and contract checks. It is not a wellness, treatment, production DSP, or real-time performance claim.

Boundaries

• synthetic examples only in v0;
• no therapy, medical, healing, brainwave, entrainment, or consciousness claims;
• no copyrighted audio fixtures;
• no production DSP claim;
• no real-time/mobile performance claim.