from ..base import SeparationBase
[docs]class MixAsEstimate(SeparationBase):
"""
This algorithm does nothing but scale the mix by the number of sources. This can
be used to compute the improvement metrics (e.g. improvement in SDR over using the
mixture as the estimate).
Args:
input_audio_signal (AudioSignal): Signal to separate.
num_sources (int): How many sources to return.
"""
def __init__(self, input_audio_signal, num_sources):
self.num_sources = num_sources
super().__init__(input_audio_signal=input_audio_signal)
def run(self):
pass
def make_audio_signals(self):
scalar = 1
estimates = [
self.audio_signal * scalar for i in range(self.num_sources)
]
return estimates