those hi-res files with 24 bits and samplerates > 90 kHz are a kind of challenge for FTF...
FTF expects frequencies according to the samplerate, in this case up to 48 kHz
As you can see, there ARE frequencies, but they get sketchy somehow. The algorithm interprets this as a cutoff around 32 kHz (wich is of course still very good)
Since it's hard to reconstruct the exact value of the VBR bitrate from the spectrum, FTF just reports the "320" (as the max. value for standard CBR files) to indicate: Something seems to be wrong.
You can chose to set the option "allow cutoffs above xxx Hz" and those files fill no longer be reported.
19600 -> 31000 : 14 ok
32000 : 1 bad / 13 ok
33000 -> 43000 : 9 bad / 5 ok
44000 : 12 bad / 2 ok
The cutoffs are clearly around 32 kHz rather than the expected 48 kHz, but that's a long way from the 22.4 kHz of 16-bit files.
32 kHz good?
"We could go down to 32 kHz because the highest frequencies will be shaved during MP3 compression. This would save a little storage space.
...However, with the rapid evolution of bandwidth and the amount of disk space currently available on even the smallest recent computer, the gain is derisory and does not go in the direction of improving sound."
Clearly, VBR cuts high-frequency treble
But
"It is measured in hertz (Hz), and indicates the vibrations per second that reach our ears. There, these vibrations are processed by the different parts of the hearing organ. On average, humans can hear frequencies between 20 and 20,000 hertz."
I doubt that anyone, especially at my age, can hear above 25,000 hertz.
On the other hand, by reducing the cutoffs, FtF seems to be able to calculate the bitrate, which is HD.
This explanation suits me and reassures me
In future I'll be able to refine the verification when the result at 48kHz isn't good.