Skip to content

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.43.1] - 2025-09-13

Fixed

  • Fix a bug introduced in 0.43.0 where the noise added by AddBackgroundNoise had wrong offset if the noise was longer than the given signal.

0.43.0 - 2025-09-09

Changed

  • Make LoudnessNormalization 400% faster by switching from pyloudnorm to the new and faster alternative loudness
  • Improve AddBackgroundNoise performance: Only load/decode the part of the noise file that is needed, instead of the entire file.
  • Bump min Python version to 3.10
  • Switch to a more modern type hint style

Deprecated

  • Deprecate "pydub" backend in Mp3Compression, because pydub isn't maintained anymore, and depends on audioop, which got removed in Python 3.13. The recommended alternative is "fast-mp3-augment"

Removed

  • Remove upper Python version limit to avoid hindering early adopters
  • Remove LRU cache in AddBackgroundNoise

0.42.0 - 2025-07-04

Added

  • Add support for Python 3.13
  • Add support for librosa 0.11.0

Changed

  • Make Mp3Compression 25-300% faster (depending on hardware, audio properties like duration and number of channels and various params, like bitrate) with the new backend="fast-mp3-augment" (now default). The extra dependency for this is fast-mp3-augment , which pulls a few useful tricks for faster execution.
  • Make Limiter 30% faster and easier to install (extra dependency is now numpy-audio-limiter instead of cylimiter ). The Limiter behavior has not changed, although there are minor numerical differences.

Fixed

  • Handle non-contiguous audio ndarray input to PitchShift and TimeStretch properly

0.41.0 - 2025-05-05

Added

  • Add support for NumPy 2.x
  • Add weights parameter to OneOf. This lets you guide the probability of each transform being chosen.

Changed

  • Improve type hints

⚠ The TimeMask transform has been changed significantly:

  • Breaking change: Remove fade parameter. fade_duration=0.0 now denotes disabled fading.
  • Enable fading by default
  • Apply a smooth fade curve instead of a linear one
  • Add mask_location parameter
  • Change the default value of min_band_part from 0.0 to 0.01
  • Change the default value of max_band_part from 0.5 to 0.2
  • ~50% faster

The following examples show how you can adapt your code when upgrading from <=v0.40.0 to >=v0.41.0:

<= 0.40.0 >= 0.41.0
TimeMask(min_band_part=0.1, max_band_part=0.15, fade=True) TimeMask(min_band_part=0.1, max_band_part=0.15, fade_duration=0.01)
TimeMask() TimeMask(min_band_part=0.0, max_band_part=0.5, fade_duration=0.0)

Removed

  • SpecCompose, SpecChannelShuffle and SpecFrequencyMask have been removed. You can read more about this here: #391

0.40.0 - 2025-03-20

Added

  • Add support for scipy>=1.13

Changed

  • Lay the groundwork for NumPy 2.x support (version constraint update coming in the next release)
  • Speed up LoudnessNormalization by ~20%
  • Improve test coverage and documentation
  • Bump min python-stretch version and remove the limitation on the number of channels in PitchShift
  • Bump min numpy version to 1.22
  • Bump min pyroomacoustics version to 0.7.4

Fixed

  • Fix a bug where TimeMask could raise an exception if the fade length became 0
  • Disallow min_cutoff_freq <= 0 in HighPassFilter
  • Make AdjustDuration picklable (useful for multiprocessing)

Removed

  • Remove support for Python 3.8

0.39.0 - 2025-02-12

Changed

  • Place an upper distance limit of 2500 meters in AirAbsorption in order to avoid numerical issues
  • Expand the allowed shift range in PitchShift from [-12, 12] to [-24, 24]
  • Switch to a higher quality method, "signalsmith_stretch", in PitchShift and TimeStretch. It sounds significantly better (e.g. less smearing) and is 50-100% faster than "librosa_phase_vocoder"

If you want to keep using the old method, "librosa_phase_vocoder", it can be done like this:

PitchShift(method="librosa_phase_vocoder")
TimeStretch(method="librosa_phase_vocoder")

Fixed

  • Fix a bug where AddShortNoises(include_silence_in_noise_rms_estimation=False) sometimes raised a ValueError due to digital silence in a portion of a short noise. This bug was introduced in v0.36.1.

0.38.0 - 2024-12-06

Added

  • Add/improve parameter validation in AddGaussianSNR, GainTransition, LoudnessNormalization and AddShortNoises
  • Add/update type hints for consistency
  • Add human-readable string representation of audiomentations class instances

Changed

  • Improve documentation with respect to consistency, clarity and grammar
  • Adjust Python version compatibility range, so all patches of Python 3.12 are supported

Removed

For example:

Old (deprecated since v0.31.0) New
Gain(min_gain_in_db=-12.0) Gain(min_gain_db=-12.0)

Fixed

  • Fix a bug where AirAbsorption often chose the wrong humidity bucket
  • Fix wrong logic in validation check of relation between crossfade_duration and min_part_duration in RepeatPart
  • Fix default value of max_absolute_rms_db in AddBackgroundNoises. It was incorrectly set to -45.0, but is now -15.0. This bug was introduced in v0.31.0.
  • Fix various errors in the documentation of AddShortNoises and AirAbsorption
  • Fix a bug where AddShortNoises sometimes raised a ValueError because of an empty array. This bug was introduced in v0.36.1.

0.37.0 - 2024-09-03

Changed

  • Leverage the SIMD-accelerated numpy-minmax package for speed improvements. These transforms are faster now: Limiter, Mp3Compression and Normalize. Unfortunately, this change removes support for macOS running on Intel. Intel Mac users have the following options: A) use audiomentations 0.36.1, B) Create a fork of audiomentations, C) submit a patch to numpy-minmax, D) run Linux or Windows.
  • Limit numpy dependency to >=1.21,<2 for now, since numpy v2 is not officially supported yet.

0.36.1 - 2024-08-20

Changed

  • Leverage the SIMD-accelerated numpy-rms package for significant speed improvements. These transforms are faster now: AddBackgroundNoise, AddColorNoise, AddGaussianSNR, AddShortNoises, Mp3Compression and TanhDistortion. Unfortunately, this change removes support for Windows running on ARM.

0.36.0 - 2024-06-10

Added

  • Add support for multichannel impulse responses in ApplyImpulseResponse

Changed

  • ⚠ Limiter no longer introduces delay. This is a backwards-incompatible change.
  • Make RoomSimulator faster by avoiding unneeded calculations when the transform is not going to be applied (p<1)
  • Limit scipy dependency to <1.13 because 1.13 is not compatible for now.

0.35.0 - 2024-03-15

Added

  • Add new transforms: AddColorNoise, Aliasing and BitCrush

0.34.1 - 2023-11-24

Changed

  • Bump min numpy version from 1.18 to 1.21
  • Use numpy.typing in type hints
  • Optimize max abs calculations in terms of memory and speed. This makes Normalize, Mp3Compression and Limiter slightly faster.

0.33.0 - 2023-08-30

Changed

  • Bump min numpy version from 1.16 to 1.18
  • Bump min scipy version from 1.3 to 1.4
  • Bump min python version from 3.7 to 3.8, because 3.7 is beyond end-of-life already
  • Change some AssertionError exceptions to ValueError

⚠ The Shift transform has been changed:

  • Removed fade parameter. fade_duration=0.0 now denotes disabled fading.
  • Rename min_fraction to min_shift and max_fraction to max_shift
  • Add shift_unit parameter
  • Fading is enabled by default
  • Smoother fade curve

These are breaking changes. The following examples show how you can adapt your code when upgrading from <=v0.32.0 to >=v0.33.0:

<= 0.32.0 >= 0.33.0
Shift(min_fraction=-0.5, max_fraction=0.5, fade=True, fade_duration=0.01) Shift(min_shift=-0.5, max_shift=0.5, shift_unit="fraction", fade_duration=0.01)
Shift() Shift(fade_duration=0.0)

Fixed

  • Correct some wrong type hints

0.32.0 - 2023-08-15

Added

  • Add new RepeatPart transform

Changed

  • Bump min version of numpy dependency from 1.13 to 1.16
  • If a transform is in "frozen parameters" mode, but has no parameters yet, the transform will randomize/set parameters when it gets called for the first time
  • Increase the threshold for raising WrongMultichannelAudioShape. This allows some rare use cases where the number of channels slightly exceeds the number of samples.

Fixed

  • Fix some type hints that were np.array instead of np.ndarray

0.31.0 - 2023-06-21

Changed

  • Raise exception instead of warning when the given multichannel ndarray has wrong shape
  • Add support for the latest librosa 0.10 version
  • Switch to a faster default resampler internally in PitchShift, leading to much faster execution. This requires soxr.
  • Bump min scipy requirement from 1.0 to 1.3
  • Rename "_in_db" to "_db" in args and parameters. Passing args with the old names still works, but is deprecated and will stop working in a future version.

0.30.0 - 2023-05-02

Added

  • Add new AdjustDuration transform

Fixed

  • Fix a bug where too loud inputs got wrap distortion when running them through Mp3Compression

0.29.0 - 2023-03-15

Added

  • Add apply_to parameter that can be set to "only_too_loud_sounds" in Normalize

Changed

  • Change default value of noise_rms from "relative" to "relative_to_whole_input" in AddShortNoises
  • Change default values of min_snr_in_db (from 0.0 to -6.0), max_snr_in_db (from 24.0 to 18.0), min_time_between_sounds (from 4.0 to 2.0) and max_time_between_sounds (from 16.0 to 8.0) in AddShortNoises

Fixed

  • Fix a bug where Limiter raised an exception when it got digital silence as input

0.28.0 - 2023-01-12

Added

  • Add/improve type hints
  • Add/improve documentation

Fixed

  • Fix a bug in RoomSimulator where the value of max_order was not respected

Removed

  • Remove FrequencyMask that had been deprecated since version 0.22.0. BandStopFilter is a good alternative.

0.27.0 - 2022-09-13

Changed

  • Speed up Limiter by ~8x
  • Fix/improve some docstrings and type hints
  • Change default values in Trim and ApplyImpulseResponse according to the warnings that were added in v0.23.0
  • Emit a FutureWarning when noise_rms in AddShortNoises is not specified - the default value will change from "relative" to "relative_to_whole_input" in a future version.

0.26.0 - 2022-08-19

Added

  • Add new transform Lambda. Thanks to Thanatoz-1.
  • Add new transform Limiter. Thanks to pzelasko.

Fixed

  • Fix incorrect type hints in RoomSimulator
  • Make Shift robust to different sample rate inputs when parameters are frozen

0.25.1 - 2022-06-15

Fixed

  • Fix a bug where RoomSimulator would treat an x value as if it was y, and vice versa

0.25.0 - 2022-05-30

Added

  • Add AirAbsorption transform
  • Add mp4 to the list of recognized audio filename extensions

Changed

  • Guard against invalid params in TimeMask
  • Emit FutureWarning instead of UserWarning in Trim and ApplyImpulseResponse
  • Allow specifying a file path, a folder path, a list of files or a list of folders to ApplyImpulseResponse, AddBackgroundNoise and AddShortNoises. Previously only a path to a folder was allowed.

Fixed

  • Fix a bug with noise_transform in AddBackgroundNoise where some SNR calculations were done before the noise_transform was applied. This has sometimes led to incorrect SNR in the output. This changes the behavior of AddBackgroundNoise (when noise_transform is used).

Removed

  • Remove support for Python 3.6, as it is past its end of life already. RIP.

0.24.0 - 2022-03-18

Added

  • Add SevenBandParametricEQ transform
  • Add optional noise_transform in AddShortNoises
  • Add .aac and .aif to the list of recognized audio filename endings

Changed

  • Show warning if top_db and/or p in Trim are not specified because their default values will change in a future version

Fixed

  • Fix filter instability bug related to center freq above Nyquist freq in LowShelfFilter and HighShelfFilter

0.23.0 - 2022-03-07

Added

  • Add Padding transform
  • Add RoomSimulator transform for simulating shoebox rooms using pyroomacoustics
  • Add parameter signal_gain_in_db_during_noise in AddShortNoises

Changed

  • Not specifying a value for leave_length_unchanged in AddImpulseResponse now emits a warning, as the default value will change from False to True in a future version.

Removed

  • Remove the deprecated AddImpulseResponse alias. Use ApplyImpulseResponse instead.
  • Remove support for the legacy parameters min_SNR and max_SNR in AddGaussianSNR
  • Remove useless default path value in AddBackgroundNoise, AddShortNoises and ApplyImpulseResponse

0.22.0 - 2022-02-18

Added

  • Implement GainTransition
  • Add support for librosa 0.9
  • Add support for stereo audio in Mp3Compression, Resample and Trim
  • Add "relative_to_whole_input" option for noise_rms parameter in AddShortNoises
  • Add optional noise_transform in AddBackgroundNoise

Changed

  • Improve speed of PitchShift by 6-18% when the input audio is stereo

Deprecated

  • Deprecate FrequencyMask in favor of BandStopFilter

Removed

  • Remove support for librosa<=0.7.2

0.21.0 - 2022-02-10

Added

  • Add support for multichannel audio in ApplyImpulseResponse, BandPassFilter, HighPassFilter and LowPassFilter
  • Add BandStopFilter (similar to FrequencyMask, but with overhauled defaults and parameter randomization behavior), PeakingFilter, LowShelfFilter and HighShelfFilter
  • Add parameter add_all_noises_with_same_level in AddShortNoises

Changed

  • Change BandPassFilter, LowPassFilter, HighPassFilter, to use scipy's butterworth filters instead of pydub. Now they have parametrized roll-off. Filters are now steeper than before by default - set min_rolloff=6, max_rolloff=6 to get the old behavior. They also support zero-phase filtering now. And they're at least ~25x times faster than before!

Removed

  • Remove optional wavio dependency for audio loading

0.20.0 - 2021-11-18

Added

  • Implement OneOf and SomeOf for applying one of or some of many transforms. Transforms are randomly chosen every call. Inspired by augly. Thanks to Cangonin and iver56.
  • Add a new argument apply_to_children (bool) in randomize_parameters, freeze_parameters and unfreeze_parameters in Compose and SpecCompose.

Changed

  • Insert three new parameters in AddBackgroundNoise: noise_rms (defaults to "relative", which is the old behavior), min_absolute_rms_in_db and max_absolute_rms_in_db. This may be a breaking change if you used AddBackgroundNoise with positional arguments in earlier versions of audiomentations! Please use keyword arguments to be on the safe side - it should be backwards compatible then.

Fixed

  • Remove global pydub import which was accidentally introduced in v0.18.0. pydub is considered an optional dependency and is imported only on demand now.

0.19.0 - 2021-10-18

Added

  • Implement TanhDistortion. Thanks to atamazian and iver56.
  • Add a noise_rms parameter to AddShortNoises. It defaults to relative, which is the old behavior. absolute allows for adding loud noises to parts that are relatively silent in the input.

0.18.0 - 2021-08-05

Added

  • Implement BandPassFilter, HighPassFilter, LowPassFilter and Reverse. Thanks to atamazian.

0.17.0 - 2021-06-25

Added

  • Add a fade option in Shift for eliminating unwanted clicks
  • Add support for 32-bit int wav loading with scipy>=1.6
  • Add support for float64 wav files. However, the use of this format is discouraged, since float32 is more than enough for audio in most cases.
  • Implement Clip. Thanks to atamazian.
  • Add some parameter sanity checks in AddGaussianNoise
  • Officially support librosa 0.8.1

Changed

  • Rename AddImpulseResponse to ApplyImpulseResponse. The former will still work for now, but give a warning.
  • When looking for audio files in AddImpulseResponse, AddBackgroundNoise and AddShortNoises, follow symlinks by default.
  • When using the new parameters min_snr_in_db and max_snr_in_db in AddGaussianSNR, SNRs will be picked uniformly in the Decibel scale instead of in the linear amplitude ratio scale. The new behavior aligns more with human hearing, which is not linear.

Fixed

  • Avoid division by zero in AddImpulseResponse when input is digital silence (all zeros)
  • Fix inverse SNR characteristics in AddGaussianSNR. It will continue working as before unless you switch to the new parameters min_snr_in_db and max_snr_in_db. If you use the old parameters, you'll get a warning.

0.16.0 - 2021-02-11

Added

  • Implement SpecCompose for applying a pipeline of spectrogram transforms. Thanks to omerferhatt.

Fixed

  • Fix a bug in SpecChannelShuffle where it did not support more than 3 audio channels. Thanks to omerferhatt.
  • Limit scipy version range to >=1.0,<1.6 to avoid issues with loading 24-bit wav files. Support for scipy>=1.6 will be added later.

0.15.0 - 2020-12-10

Added

  • Add an option leave_length_unchanged to AddImpulseResponse

Fixed

  • Fix picklability of instances of AddImpulseResponse, AddBackgroundNoise and AddShortNoises

0.14.0 - 2020-12-06

Added

  • Implement LoudnessNormalization
  • Implement randomize_parameters in Compose. Thanks to SolomidHero.
  • Add multichannel support to AddGaussianNoise, AddGaussianSNR, ClippingDistortion, FrequencyMask, PitchShift, Shift, TimeMask and TimeStretch

0.13.0 - 2020-11-10

Added

  • Lay the foundation for spectrogram transforms. Implement SpecChannelShuffle and SpecFrequencyMask.
  • Configurable LRU cache for transforms that use external sound files. Thanks to alumae.
  • Officially add multichannel support to Normalize

Changed

  • Show a warning if a waveform had to be resampled after loading it. This is because resampling is slow. Ideally, files on disk should already have the desired sample rate.

Fixed

  • Correctly find audio files with upper case filename extensions.
  • Fix a bug where AddBackgroundNoise crashed when trying to add digital silence to an input. Thanks to juheeuu.

0.12.1 - 2020-09-28

Changed

  • Speed up AddBackgroundNoise, AddShortNoises and AddImpulseResponse by loading wav files with scipy or wavio instead of librosa.

0.12.0 - 2020-09-23

Added

  • Implement Mp3Compression
  • Officially support multichannel audio in Gain and PolarityInversion
  • Add m4a and opus to the list of recognized audio filename extensions

Changed

  • Expand range of supported librosa versions

Removed

  • Python <= 3.5 is no longer officially supported, since Python 3.5 has reached end-of-life
  • Breaking change: Internal util functions are no longer exposed directly. If you were doing e.g. from audiomentations import calculate_rms, now you have to do from audiomentations.core.utils import calculate_rms

0.11.0 - 2020-08-27

Added

  • Implement Gain and PolarityInversion. Thanks to Spijkervet for the inspiration.

0.10.1 - 2020-07-27

Changed

  • Improve the performance of AddBackgroundNoise and AddShortNoises by optimizing the implementation of calculate_rms.

Fixed

  • Improve compatibility of output files written by the demo script. Thanks to xwJohn.
  • Fix division by zero bug in Normalize. Thanks to ZFTurbo.

0.10.0 - 2020-05-05

Added

  • AddImpulseResponse, AddBackgroundNoise and AddShortNoises now support aiff files in addition to flac, mp3, ogg and wav

Changed

  • Breaking change: AddImpulseResponse, AddBackgroundNoise and AddShortNoises now include subfolders when searching for files. This is useful when your sound files are organized in subfolders.

Fixed

  • Fix filter instability bug in FrequencyMask. Thanks to kvilouras.

0.9.0 - 2020-02-20

Added

  • Remember randomized/chosen effect parameters. This allows for freezing the parameters and applying the same effect to multiple sounds. Use transform.freeze_parameters() and transform.unfreeze_parameters() for this.
  • Implement transform.serialize_parameters(). Useful for when you want to store metadata on how a sound was perturbed.
  • Add a rollover parameter to Shift. This allows for introducing silence instead of a wrapped part of the sound.
  • Add support for flac in AddImpulseResponse
  • Implement AddBackgroundNoise transform. Useful for when you want to add background noise to all of your sound. You need to give it a folder of background noises to choose from.
  • Implement AddShortNoises. Useful for when you want to add (bursts of) short noise sounds to your input audio.

Changed

  • Disregard non-audio files when looking for impulse response files
  • Switch to a faster convolve implementation. This makes AddImpulseResponse significantly faster.
  • Expand supported range of librosa versions

Fixed

  • Fix a bug in ClippingDistortion where the min_percentile_threshold was not respected as expected.
  • Improve handling of empty input

0.8.0 - 2020-01-28

Added

  • Add shuffle parameter in Composer
  • Add Resample transformation
  • Add ClippingDistortion transformation
  • Add fade parameter to TimeMask

Thanks to askskro

0.7.0 - 2020-01-14

Added

  • AddGaussianSNR
  • AddImpulseResponse
  • FrequencyMask
  • TimeMask
  • Trim

Thanks to karpnv

0.6.0 - 2019-05-27

Added

  • Implement peak normalization

0.5.0 - 2019-02-23

Added

  • Implement Shift transform

Changed

  • Ensure p is within bounds

0.4.0 - 2019-02-19

Added

  • Implement PitchShift transform

Fixed

  • Fix output dtype of AddGaussianNoise

0.3.0 - 2019-02-19

Added

  • Implement leave_length_unchanged in TimeStretch

0.2.0 - 2019-02-18

Added

  • Add TimeStretch transform
  • Parametrize AddGaussianNoise

0.1.0 - 2019-02-15

Added

  • Initial release. Includes only one transform: AddGaussianNoise