The funny thing is that many so-called HDCDs (from my experience even most of them) don't have either of those features enabled, yet they are labeled as HDCD. But it only means that they were mastered on a HDCD capable device - they contain nothing that needs to be decoded.They only contain the control bits that will cause the HDCD light on an HDCD capable player to lighten up, so people in audiophile internet forums can rant about how much better this or that album sounds on a HDCD player, not knowing that there's really no difference at all.Now you should relax and make peace with your wife
Why is the PM-2 considered to be that great exactly?
* I have a couple of CDs and at least one isolated track -- which report minimum gain = 0, maximum gain = 0, PE enabled and transient filter "Intermittent". Now I have no idea what the latter means, but is it so that the plugin does nothing about the transient filter at all (apart from detecting)? Various internet sources claim that no software decoder does.
* Quite a few of the HDCD tracks (including entire discs) have minimum gain = maximum gain = 0 dB, and both features reported to be "Disabled". These are the ones which have only been touched by a HDCD-enabled device, but not used any of the features? I.e., these should be regarded a "non-HDCD with extra noise in the LSB"? (And the plugin does so, when I have checked volume reduction to be applied only when P.E. enabled, right?)
* I found a CD where all tracks except one report gain 0, T.F. disabled, P.E. enabled -- and the exceptional track #2 reporting peak extension as "Intermittent". I can only guess that the entire CD is supposed to be treated equal? Or? How does the plugin react?
* There seems to be some difference between the component and the hdcd.exe utility, as one track which was skipped by the latter, was identified by foo_hdcd -- this on a CD which both of them agree, has both HDCD and non-HDCD tracks.
* There are a few which peak at .5 or slightly below. Is this due to hdcd having halved volume on discs with no peak extension? Would these have been better off not coded over to 24 bits?
* Are there other values to look out for? (There are quite a few close to the square root of .5). Other signs so? For example, while peak scan shows that one end of the bitword is unused, is there any easy scan which will reveal it if the other end is unused?
. So while there may be perceptibly more signal room due to dithering, there is no actual information below the 16 bits of the input signal unless either of those two features is used.
(1) hdcd.exe: Does not detect HDCD on any of the tracks. Maybe due to not scanning the entire track?
(4) I am not sure how to use your component to encode a HDCD -- I suspect I have to set bit-depth manually?
First, WAV and AIFF are almost exactly the same thing. They are both containers for RAW PCM audio. The only difference is that WAV originally was a Microsoft / IBM standard and AIFF originaly was an Apple standard. They both will preserve the resolution of the CD PCM audio perfectly and untouched. The downside is that the size of the files will be exactly the same as it was on the CD and thus very large (about 10MB per minute of CD audio).
I am interested if there are any HDCD encoders that can be added in foobar. If I am playing a 24bit, 88.2kHz (native) file, to be able to convert it to a HDCD 16bit, 44.1kHz with both extensions enabled (to play at 20bit level in my universal DVD player or in my receiver - the WDTV Live that I use doesn't pass over SPDIF the 88.2k files, just the 96k ones).
OK, but what is done with the signal at all when there's no peak extension enabled? I examined some tracks with foobar and the bit comparison did show some minor differences (~0.01% of the samples), though the peak values were unchanged and I didn't ever bother to try to ABX the tracks.
There should be no difference at all if the volume isn't being halved and there's no peak extension. The signal is otherwise passed through unmodified, except for the conversion from and back to float.
So the differences are most likeley due to rounding error, I guess. But what about this low level extention stuff?
"bit shift" is 6,020599913...dB, not just 6 dB.
The main volume change in this decoder is a bit shift, while the low level extension is in half decibel increments.
Convert -> Processing -> untick Additional decoding checkbox.
This seems more and more like a bug -- versions 1.12 and 1.14 differ in behaviour:In a posting above, I described a case where "Scan for HDCD" yields peak extension- "Enabled" on file.flac- "Intermittent" on file.wavand for the record, they verify as identical using foo_bitcompare.I just checked this with another computer which used version 1.12, and the issue was not present. Upgraded to 1.14 and applied changes to (= restarted) fb2k, and problem reproduced on that other computer. So, problem seems to be introduced in 1.13 or 1.14. (Both fb2ks were current version 1.10.)Also, to clarify:Quote from: Porcus on 18 December, 2011, 04:45:47 PM(1) hdcd.exe: Does not detect HDCD on any of the tracks. Maybe due to not scanning the entire track?It returns "HDCD not detected".