Recently I noticed when I checked (with TLH) fingerprints in newer hi-res 24bit flac releases from HDTracks/Qobuz that it returned either 'file format is not supported' or 'fingerprint stored is zero'? Older hi-res flacs don't ave this problem although I have no clue exactly when the transition happened. Has anyone else reflected over this? Is it a problem with the FLAC format or does it have something to do with the ffp?
I don't know TLH specifically, nor HDTracks/Qobuz which I don't have access to, but elsewhere I've seen .ffp generated by metaflac from the FLAC file's audio-md5.
I.e. it is common to read the md5 that has already been calculated and stored.
There are applications that write files without the md5, and that could be an explanation.
Here is what to do to sort this out:
* Re-encode your Qobuz.flac as a new file audiocopy.flac. Should be the same audio signal.
* Verify that they are the same audio signal using e.g. foobar2000 with foo_bitcompare
* Test both files with various tools; TLH, AudioTester, foo_verifier, ... - do the two files give you the same result?
ffp files exist, I believe, because the bootleg trading community holds on to the practices that were established when they used Shorten (which had no error checking) or .wav/.aiff (ditto - and I think the later ALAC in .m4a is just as bad).