TensorflowInputVGGish

standard mode | Spectral category

Inputs

  • frame (vector_real) - the audio frame

Outputs

  • bands (vector_real) - the log compressed mel bands

Description

This algorithm computes mel-bands specific to the input of VGGish-based models.

References:

[1] Gemmeke, J. et. al., AudioSet: An ontology and human-labelled dataset for audio events, ICASSP 2017

[2] Hershey, S. et. al., CNN Architectures for Large-Scale Audio Classification, ICASSP 2017

[3] Supported models at https://essentia.upf.edu/models/

Source code

See also

Scale (standard) Scale (streaming) TensorflowInputVGGish (streaming)