nvidia.dali.fn.to_decibels(__input, /, *, bytes_per_sample_hint=[0], cutoff_db=-200.0, multiplier=10.0, preserve=False, reference=0.0, seed=-1, device=None, name=None)#

Converts a magnitude (real, positive) to the decibel scale.

Conversion is done according to the following formula:

min_ratio = pow(10, cutoff_db / multiplier)
out[i] = multiplier * log10( max(min_ratio, input[i] / reference) )
Supported backends
  • ‘cpu’

  • ‘gpu’


__input (TensorList) – Input to the operator.

Keyword Arguments:
  • bytes_per_sample_hint (int or list of int, optional, default = [0]) –

    Output size hint, in bytes per sample.

    If specified, the operator’s outputs residing in GPU or page-locked host memory will be preallocated to accommodate a batch of samples of this size.

  • cutoff_db (float, optional, default = -200.0) –

    Minimum or cut-off ratio in dB.

    Any value below this value will saturate. For example, a value of cutoff_db=-80 corresponds to a minimum ratio of 1e-8.

  • multiplier (float, optional, default = 10.0) – Factor by which the logarithm is multiplied. The value is typically 10.0 or 20.0, which depends on whether the magnitude is squared.

  • preserve (bool, optional, default = False) – Prevents the operator from being removed from the graph even if its outputs are not used.

  • reference (float, optional, default = 0.0) –

    Reference magnitude.

    If a value is not provided, the maximum value for the input will be used as reference.


    The maximum of the input will be calculated on a per-sample basis.

  • seed (int, optional, default = -1) –

    Random seed.

    If not provided, it will be populated based on the global seed of the pipeline.