What determines the time axis in spectrogram using normalized frequencies?

I'm attempting to make a spectrogram using normalized frequencies, and the resulting spectrogram looks as expected with the exception of the time axis. Specifically, while the true duration of the data is ~8 minutes, when calculating the spectrogram the time axis ranges to ~21.5 hours!

I've tried manipulating each of the input arguments, and the only thing that seems to affect the time axis is the length of the input data vector - which obviously doesn't make the time axis any more accurate. Any help would be greatly appreciated.

PS: I calculate the normalized frequencies thusly:

FreqsInHz = [0.5:0.1:50]; %The range of frequencies I'm interested in.
SamplingRate = 1000Hz;
normFreqs = (2*pi).*FreqsInHz./SamplingRate; Normalized frequencies in rad/sample.

Answer 1

[Sean de Wolski]

Time is a function of the sampling frequency, fs.

Compare

spectrogram(rand(1,100000),64,0,64,8000)
spectrogram(rand(1,100000),64,0,64,44100)

For how to use Fs

>> doc spectrogram

[Jostein Holmgren]

This only works for cyclical frequencies.

From the spectrocram documentation:

[s,w,t] = spectrogram(x,window,noverlap,w) returns the spectrogram at the normalized frequencies specified in w.

[s,f,t] = spectrogram(x,window,noverlap,f,fs) returns the spectrogram at the cyclical frequencies specified in f.

If I add fs, the frequencies are interpreted to be cyclical, not normalized.