I mean the FFT windowing of the impulse response. If I change the window to exclude low frequencies, and reload the spectrogram, the spectrogram low frequency cutoff will reflect the windowing change. I suppose that tricked me into thinking the IR window affected the spectrogram.