Why does lengthofinputsamples/hop_length not equal to the stft's output 2nd dimension. Instead, I noticed that the stft's output was equal to ceil(lengthofinputsamples/hop_length). Why is this?
In my instance:
sgram = np.abs(librosa.stft(samples,n_fft=2048,hop_length=512,center=True))
returns 21189 when the input samples length is equal to 10848704
