functional__find_max_per_frame {torchaudio}

Find Max Per Frame (functional)

Description

For each frame, take the highest value of NCCF, apply centered median smoothing, and convert to frequency.

Usage

functional__find_max_per_frame(nccf, sample_rate, freq_high)

Arguments

nccf

(tensor): Usually a tensor returned by functional__compute_nccf

sample_rate

(int): sampling rate of the waveform, e.g. 44100 (Hz)

freq_high

(int): Highest frequency that can be detected (Hz)

Note: If the max among all the lags is very close to the first half of lags, then the latter is taken.

Value

tensor with indices