Hyperbolic sweep support?

[DeepSOIC]

Hi! Log sweeps are nice, but the results it generates have excessive resolution in the high frequencies and poor resolution in the low end, forcing me to do long measurements to have a decent bass response accuracy in the measurement. Hyperbolic sweeps are better in that regard, they have a fixed number of oscillation cycles per octave, so should give a naturally even data density in the resulting frequency response plotted on log scale. I have generated a sweep file using scipy, in hopes of trying it out:

import scipy.signal as sig
import numpy as np
times = np.linspace(0,30, 30*44100)
wf = sig.chirp(times, 20000, 30, 20, method='hyperbolic')
import scipy.io.wavfile as wavfile
wavfile.write(r'drive:\path\to\logchirp-20-20k-30s.wav', 44100, wf)

but when i try to use it for a measurement, i get an error:

the file does not contain the data REW requires to generate the measurement, please use a measurement sweep saved from this REW version

So, apparently, i can't just trick REW into doing the hyperbolic sweep, it needs an in-depth support in the code... or does it?

using REW 5.20 on Windows10

 

[John Mulcahy]

Log sweeps are nice, but the results it generates have excessive resolution in the high frequencies and poor resolution in the low end

No, that isn't correct. A log sweep spends equal time in each octave, so per Hz it spends far more time at low frequencies than high, optimising the signal to noise ratio. Hyperbolic sweeps are typically used to spend less time at the ends of the frequency range and more in the middle. Frequency resolution is a property of the FFT and the window width rather than the sweep, longer impulse response windows give higher frequency resolution.

 

[DeepSOIC]

Well, that only goes to show i have little idea of what's going on under the hood in rew, i suppose =)
Thanks for your reply.