Closed MattCarrickPL closed 3 weeks ago
Quantize() transform currently uses 16, 24, 32, 40, 48, 56, 64 levels which are all represented by 4-6 bits.
def __init__( self, num_levels: IntParameter = UniformDiscreteRD( np.asarray([16, 24, 32, 40, 48, 56, 64], dtype=int) ), round_type: List[str] = (["floor", "middle", "ceiling"]), ) -> None:
https://github.com/TorchDSP/torchsig/blob/v0.4.2/torchsig/transforms/transforms.py#L2961
Would like to increase the quantization range to include from 7 to 16 bits.
Old issue, closing out
Quantize() transform currently uses 16, 24, 32, 40, 48, 56, 64 levels which are all represented by 4-6 bits.
https://github.com/TorchDSP/torchsig/blob/v0.4.2/torchsig/transforms/transforms.py#L2961
Would like to increase the quantization range to include from 7 to 16 bits.