Clarification regarding spectogram shapes about cnns-speech-music-discrimination HOT 1 OPEN

from pyAudioAnalysis import audioBasicIO as io, audioFeatureExtraction as aF, audioSegmentation
[Fs, x] = io.readAudioFile("test_data/1992_10_28_06_10.wav") #read in a sample file
signal = io.stereo2mono(x)

mtWin = 2.4 #also WIDTH_SEC
mtStep = 1.0 
mtWin2 = int(mtWin * Fs)
mtStep2 = int(mtStep * Fs)
stWin = 0.020
stStep = 0.015
curPos = 0
N1 = curPos
N2 = curPos + mtWin2 + stStep*Fs
N = len(signal)
if N2 > N:    
    N2 = N

xtemp = signal[int(N1):int(N2)] #let is just look at the first 2.4 seconds

specgram, TimeAxis, FreqAxis = aF.stSpectogram(xtemp, Fs, 882.0, 662.0, False)
print(specgram.shape)
print(Fs)
print(N)

>>> print(specgram.shape)
(160, 441)
>>> print(Fs)
44100
>>> print(N)
216970240