albumentations.augmentations.spectrogram.transform
Transforms for spectrogram augmentation. This module provides transforms specifically designed for augmenting spectrograms in audio processing tasks. Includes time reversal, time masking, and frequency masking transforms commonly used in audio machine learning applications.
Members
- classFrequencyMasking
- classTimeMasking
- classTimeReverse
FrequencyMaskingclass
FrequencyMasking(
freq_mask_param: int = 30,
p: float = 0.5
)Apply masking to a spectrogram in the frequency domain. This transform masks random segments along the frequency axis of a spectrogram, implementing the frequency masking technique proposed in the SpecAugment paper. Frequency masking helps in training models to be robust against frequency variations and missing spectral information in audio signals. This is a specialized version of XYMasking configured for frequency masking only. For more advanced use cases (e.g., multiple masks, time masking, or custom fill values), consider using XYMasking directly.
Parameters
| Name | Type | Default | Description |
|---|---|---|---|
| freq_mask_param | int | 30 | Maximum possible length of the mask in the frequency domain. Must be a positive integer. Length of the mask is uniformly sampled from (0, freq_mask_param). |
| p | float | 0.5 | probability of applying the transform. Default: 0.5. |
References
- SpecAugment paper: https://arxiv.org/abs/1904.08779
- Original implementation: https://pytorch.org/audio/stable/transforms.html#freqmask
TimeMaskingclass
TimeMasking(
time_mask_param: int = 40,
p: float = 0.5
)Apply masking to a spectrogram in the time domain. This transform masks random segments along the time axis of a spectrogram, implementing the time masking technique proposed in the SpecAugment paper. Time masking helps in training models to be robust against temporal variations and missing information in audio signals. This is a specialized version of XYMasking configured for time masking only. For more advanced use cases (e.g., multiple masks, frequency masking, or custom fill values), consider using XYMasking directly.
Parameters
| Name | Type | Default | Description |
|---|---|---|---|
| time_mask_param | int | 40 | Maximum possible length of the mask in the time domain. Must be a positive integer. Length of the mask is uniformly sampled from (0, time_mask_param). |
| p | float | 0.5 | probability of applying the transform. Default: 0.5. |
References
- SpecAugment paper: https://arxiv.org/abs/1904.08779
- Original implementation: https://pytorch.org/audio/stable/transforms.html#timemask
TimeReverseclass
TimeReverse(
p: float = 0.5
)Reverse the time axis of a spectrogram image, also known as time inversion. Time inversion of a spectrogram is analogous to the random flip of an image, an augmentation technique widely used in the visual domain. This can be relevant in the context of audio classification tasks when working with spectrograms. The technique was successfully applied in the AudioCLIP paper, which extended CLIP to handle image, text, and audio inputs. This transform is implemented as a subclass of HorizontalFlip since reversing time in a spectrogram is equivalent to flipping the image horizontally.
Parameters
| Name | Type | Default | Description |
|---|---|---|---|
| p | float | 0.5 | probability of applying the transform. Default: 0.5. |
Notes
This transform is functionally identical to HorizontalFlip but provides a more semantically meaningful name when working with spectrograms and other time-series visualizations.
References
- AudioCLIP paper: https://arxiv.org/abs/2106.13043
- Audiomentations: https://iver56.github.io/audiomentations/waveform_transforms/reverse/