v24.04

What's Changed

New features:

• Datasets (DatasetSynthetic and DatasetMIRACLE): include a new feature targetmap_analytic and targetmap_estimated, which is a sparse mapping of the analytic / estimated squared sound pressure distribution. "Estimated" means from a limited number of snapshots (e.g. via Welch's method) (feature by @adku1173 in #34)

Bugfixes

• Datasets: sample the squared RMS value as at source strength instead of the pure RMS @adku1173 in #35

Others

• uses an updated version of the MIRACLE dataset
• AcouPipe now requires at least Acoular 24.03

Full Changelog: v23.11...v24.04

v23.11.

What's Changed

The new AcouPipe 23.11. Release includes several fundamental changes and new features.

New datasets submodule

A new submodule named acoupipe.datasets is now included, which contains two default dataset classes that can be instantiated to generate microphone array data for machine learning. They can be used to generate training data on the fly without writing data to file. See the docs for an example

Improved parallel computation speed

The acoupipe.DistributedPipeline now makes use of the Actor principle available in Ray. In each parallel task, random sampling and feature extraction is now performed by a dedicated acoupipe.pipeline.SamplerActor. This architectural change drastically improves performance.

Multiple Docker images

Different Docker images are now available on DockerHub.

Minor changes

• improvement of code documentation
• update package structure
• extended tests

Important Pull Requests

• Docker CI by @adku1173 in #24
• Dev by @adku1173 in #31
• Pre release work by @adku1173 in #32

Full Changelog: v21.08...v23.11

v21.08

First official release.

Includes source code to create a large-scale microphone array data set for machine learning.