audio - Giving large no of samples to KissFFT -

i wanted find 4096 point dft of audio signal of duration 10 seconds sampling rate 44100 hz. hence there 441000 input samples. kissfft takes 4096 input size. how go finding fft of such large signal?

the power spectrum of real-world audio signals (speech, music, etc) time-varying, typically calculate series of short-term ffts using overlapping windows, produce sequence of power spectra, aka spectrogram.

i suggest starting 50% overlap, first fft samples 0..4095, second 2048..6143, etc.