Second: Please allow nearest neighbor resizing of sounds. A maximum frequency sound becomes a square wave as frequency is raised. Perhaps a “Nearest neighbor (no aliasing)” that mirrors the same option for square wave, but make sure integer upsamplings are exact.
Please allow selecting a different algorithm for downsampling. A good simple one would be box sampling - pick a few samples and average them. This should somewhat remedy the issue of high frequency aliasing with downsampling using nearest neighbor and bilinear.
Please! We can’t choose 27.5! That’s important! Why make frequency choice very coarse for low frequencies?
In addition to nearest neighbor, also add bilinear and cosine interpolation for upsampling. Cosine is one of the interpolation options in Graphic mode of Equalization.
Audacity provides 4 resampling settings which can be selected in “Preferences >Quality > High-quality Sample Rate Converter”.
For detailed specifications of the resampling algorithms, please refer to SoX documentation (Audacity uses the soxr resampling library).
What does it matter? Why not just set the minimum frequency to 20, 25 or 27 Hz? That setting has no affect on the sound. It only affects the visual display in Spectrogram track view (Spectrogram View - Audacity Manual)
If I print-screen the logarithmic spectrogram with minimum frequency of 27Hz, and play it in Photosounder default (27.5Hz) it sounds high pitched. So there is a difference with logarithmic.
Photosounder is not our product and we do not offer support for it, but just as an idea, perhaps you could set the spectrogram settings in Audacity to an integer value, and tweak the screenshot in a graphics editing application (such as Gimp) before feeding it into Photosounder.
How I’m supposed to know where is 27.5? And I could as well adjust the minimum frequency in Photosounder to 27.0, but then vertical flip sounds low pitched.