v1.2.0

• Added a random_state parameter to Scaper object, which allows all runs to be perfectly reproducible given the same audio and the same random seed.
• Switched from numpydoc to napoleon for generating the documentation. Also switched Sphinx to the most recent version.
• Added functions to Scaper object that allow one to reset the foreground and background event specifications independently. This allows users to reuse the same Scaper object and generate multiple soundscapes.
• Added a function to Scaper that allows the user to set the random state after creation.

v1.1.0

Added functionality which modifies a source_time distribution tuple according to the duration of the source and the duration of the event.

This release alters the behavior of Scaper compared to earlier versions.

v1.0.3

Fix bug where temp files might not be closed if an error is raised

v1.0.2

Minor bugfixes with respect to 1.0.0

• Store sample rate in output JAMS inside the scaper sandbox
• Fix bug where estimated duration of time stretched event is different to actual duration leading to incorrect silence padding and sometimes marginally incorrect soundscape duration (in audio samples).

v1.0.0

This is a major revision, supporting jams>=0.3 and switching from the sound_event to the scaper namespace. While the API remains compatible with previous versions, the change of underlying namespace breaks compatibility with jams files created using scaper for versions <1.0.0.

v1.0.0rc1

Second candidate release

Remove sound_event namespace entirely.

v1.0.0rc0

First release candidate for 1.0.0.

This is a major revision, supporting jams>=0.3 and switching from the sound_event to the scaper namespace. While the API remains compatible with previous versions, the change of underlying namespace breaks compatibility with jams files created using scaper for versions <1.0.0.

v0.2.1

Fixes bug related to creating temp files on Windows.

v0.2.0

This release includes several improvements to the loudness computation (i.e. LUFS), resulting in more accurate SNR of events in the synthesized soundscapes.

• LUFS is computed for foreground and background sound events after all other possible manipulations (trimming, pitch shifting, time stretching), resulting in a more accurate estimate.
• For audio files < 0.5 seconds in duration, the file is first self-concatenated until the minimum duration (500 ms) is reached, and the LUFS value is computed from the concatenated version. This is to circumvent ffmpeg behavior for files < 400 ms, for which it always returns a constant of -70.0 LUFS.

v0.1.2

Minor update to fix display of Markdown from README on new pypi