Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: HDCD conundrum, yet agin (Read 2399 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

HDCD conundrum, yet agin

I'm archiving all my HDCD encoded disks to a format that will digitally apply the HDCD enhancements to the resulting FLAC files.  I'm using dbpoweramp to do this, applying its HDCD DSP stage, with its 6db boost enabled, to make 24 bit 44.1 khz flac files.  This seems to work well, but the results don't quite cut it when compared to the source disk played on my Denon DCM 390, which utilizes a pacific microsonics HDCD chip with a burr-brown dac.  I have a Cambridge Audio CXA80 integrated amp, which includes a built in DAC that I use to play the digital rips.  When I play non-HDCD disks on my denon, I use the player as a transport feeding that DAC for I find the disks sounds better that way.   As for the seemly 'different' quality of the HDCD decoded flac file playback, digging around the web brought up this nugget on the Reference Recordings web site.  Note that Reference Recordings worked hand-in-hand with the development of HDCD so I regard this info as absolutely authoritative.  Anyways, it seems that we're all screwed when it comes to fully recovering the "HDCD sound" via digital encoding.  I guess I'm now hoping that my denon will last me forever.  Any thoughts on any of this?  Thanks for reading (my first ever post here, though I've been lurking for decades).

Edit: BTW, I'm using a Reference Recording HDCD disk as the test case here, which foobar2000 assures me it's using all the HDCD bells and whistles.

Re: HDCD conundrum, yet agin

Reply #1
I'm using dbpoweramp to do this, applying its HDCD DSP stage
I also made that mistake. You cannot undo HDCD decoding, but you can apply it afterwards; ergo, don't convert.

HDCD is flagged in-signal, so if you have a HDCD-aware DAC, it will know even if is from (lossless) files. As long as you have not decoded already ... and worst is, sometimes even then IIRC; you are not guaranteed that decoding using hdcd.exe (which is what dBpoweramp uses) will kill the HDCD packets, so you could at worst apply it twice that way.

When I play non-HDCD disks on my denon, I use the player as a transport feeding that DAC for I find the disks sounds better that way.
... than? Such observations are typically volume differences or placebo.

Re: HDCD conundrum, yet agin

Reply #2
I made two flac files, one a straight through flac rip of the disk and the other that applied the DSP.  Foobar2000 signals that the clean rip is indeed HDCD when it's played back, and the DSP processed rip is not identified as such.  Of course, this is completely relying that foobar2000 'knows its stuff'.  I'm pretty sure that my integrated amp's DAC does not decode HDCD, and that's the motivation for doing the conversions.  I'll agree that playback comparisons always involve placebo, but when comparing the two DAC, the integrated amp's DAC obviously sounds better (at least subjectively) than the cd player's.  As the usual endorsement goes, even my wife notices the difference.


Re: HDCD conundrum, yet agin

Reply #4
I made two flac files, one a straight through flac rip of the disk and the other that applied the DSP.  Foobar2000 signals that the clean rip is indeed HDCD when it's played back,
OK, so you can then decode it on-the-fly. Not necessary to decode in advance. And you don't always want to do that, due to all sorts of quirks. Like, do you really want to attenuate it 6 dB when peak extension is never active and the next track doesn't have HDCD at all?

and the DSP processed rip is not identified as such.
Those issues I've seen have been on a very few files only.


Re: HDCD conundrum, yet agin

Reply #5
I want the HDCD DPS encoded files for uses other than with foobar2000.  I'm not really hung up on when/how the HDCD software decode is applied, only that it's all based on HDCD.exe which is most likely not doing all the sonic processing that the pacific microsonics chip does when working in close conjunction with a DAC.  The good news for me is that of all my HDCD disks, only about 20% really use the resolution enhancing features of HDCD.

Re: HDCD conundrum, yet agin

Reply #6
Yeah, if for example you do lossies, then that is a point. Keep a lossless undecoded original though (... wish I did), instead of decoding upon ripping.

If memory serves me right, hdcd.exe was created by reverse-engineering what Windows Media Player was doing;
Microsoft bought Pacific and by that the entire HDCD thing, but nothing says that WMP did more HDCD decoding than Microsoft found to be essential at the time, before they let HDCD fade out and die, so even the "original" software implementation might not be much exact science;
Pacific in itself didn't implement more than one of the "transient filters" it set out to use, it turns out there was a prior patent;
the public information about HDCD was intentionally vague (for one thing, Pacific didn't want to say out loud that they couldn't do everything in the white paper; and, they didn't want to go into detail on what part of the process was only done by mastering and had nothing to do with decoding and was neither better nor worse than any other downmix-and-dithering-to-CD process)

Re: HDCD conundrum, yet agin

Reply #7
It's sad that regardless of its intention, HDCD is one of those things that ended causing more grief than it ever did to bring joy.  Falls into the category of "just because you can, it doesn't mean that you should".  Too bad the MS s'canned the whole thing.  Things could have been different, at least for the remaining remnants.

Re: HDCD conundrum, yet agin

Reply #8
Quote
I'll agree that playback comparisons always involve placebo, but when comparing the two DAC, the integrated amp's DAC obviously sounds better (at least subjectively) than the cd player's.
A level-matched blind ABX test will remove the placebo bias.     It's not as easy with hardware as when comparing file formats (where the computer can do the random-blind switching) but if you and your wife are both interested in this stuff it can be done.

Ignoring this HDCD issue, differences in DACs usually go-away in a proper ABX test.  ;) 

Re: HDCD conundrum, yet agin

Reply #9
@Porcus @TheFamilyMan @DVDdoug
Hi
I own some hdcd cds too , may i know what software should I use to rip them correctly?
I own dbpoweramp last update and eac , i could try even EZ CD Audio Converter (in trial mode ) and cuetools

thanks

Re: HDCD conundrum, yet agin

Reply #10
may i know what software should I use to rip them correctly?
I own dbpoweramp last update and eac

(1) Rip them as normal, without applying the HDCD option in dBpoweramp.
Keep that rip with no decoding. The HDCD flags are in the audio, not in any subcode, so you can do all decoding later, and that is typically the best idea.

(2) To play, you can
* use foobar2000 with foo_hdcd
or
* use ffmpeg (folks, how is that easiest to implement?! Different codec/suffix and foobar --> ffmpeg?)
or
* make a decoded copy, using hdcd.exe (that is what dBpoweramp Music Converter uses) or ffmpeg

If you have a HDCD-aware DAC, it will detect the HDCD at least if you have bit-perfect output. You may have to disable volume and RG and ... and those DACs are getting rare anyway I think. But the problem about this is that those can treat it as HDCD for seconds after a track change.
Which also is one reason to keep an un-converted version. Suddenly you encounter a CD where one track has a HDCD flag and the other has not. And who knows if it changes precisely at track boundaries. Also a lot of "HDCDs" don't use any of the features.
You may have to flag some "HDCD" as "do not treat as HDCD".

Re: HDCD conundrum, yet agin

Reply #11
@Porcus
Hi
may I ask you a question?
Can EZ CD Audio Converter /CUERipper Cuetools / EAC or even fre:ac  rip HDCD ?
just for curiosisty and knowledge
thanks Porcus , appreciate your tutorial

Re: HDCD conundrum, yet agin

Reply #12
Everything that rips a CD can rip a HDCD.
If you want to decode the HDCD, you can do it afterwards, as long as you rip to lossless - because the HDCD flag is in the signal itself.

Re: HDCD conundrum, yet agin

Reply #13
Did an internet deep dive on hdcd.exe, and is there are LOTS of in-depth discussion on it especially around 2010-2012.  I can't say all that reading really helped me in discovering if hdcd.exe is doing enough of what's needed, but now I guess it's better than doing nothing (short of having a DAC that has a pacific microsonics HDCD chip).

Looked into the effect of using the 6db gain with dbpoweramp's HDCD DSP stage, and without any doubt, it does clip!  The below image shows smoking gun evidence of this.  The top waveform is of a direct 16 bit rip of the HDCD stereo track.  The second stereo track is the output of dbpoweramp using its HDCD stage with the 6db gain enable.  The cursor marks the position of a clipped signal, of which several exist throughout the disk's rip.  The third stereo track is also the output of dbpoweramp using its HDCD stage with the 6db gain disabled.  This shows that the volume of that rip's playback is substantially quieter than the two above tracks, along with the waveform of the unclipped signal.  Applying a 3.115db amplification normalizes this track without clipping.

What is interesting is that the 6db gain seems to correctly normalize most of the signal to 24 bits when compared the original 16 HDCD track, and the parts that are clipped appear to be peaks effected by the Peak Extend flag.  I can only think that such peaks are also effectively clipped in the 16 bit HDCD data, thus what hdcd.exe does is better than playing back the clipped 16 bit signal, even if hdcd.exe isn't doing all that's 'on the HDCD tin'.  Bottom line: I'm NOT going to be using the dbpoweramp HDCD 6db gain, though I will be applying a DSP 3db amplification after HDCD .  Note that the only disks that'll get this treatment will be the ones that use Peak Extension.  Fortunately I'm still only my testcase CD in my conversion project.  BTW, thanks for reading and enjoy!




Re: HDCD conundrum, yet agin

Reply #14
though I will be applying a DSP 3db amplification after HDCD
Peak extension creates peaks at up to +6dB relative to the original CD signal level, so your 3dB amplification may still cause clipping on other CDs with peak extension.

Re: HDCD conundrum, yet agin

Reply #15
@TheFamilyMan
Hi
I like a lot your test
have you tested other ripper softwares ?
about the first track the "The top waveform is of a direct 16 bit rip of the HDCD stereo track" , have you ripped with dbpoweramp without the HDCD Dsp ?

Quote
The top waveform is of a direct 16 bit rip of the HDCD stereo track.
Bottom line: I'm NOT going to be using the dbpoweramp HDCD 6db gain, though I will be applying a DSP 3db amplification after HDCD
may I ask you why don't keep the top waveform 16 bit rip of the HDCD stereo track ? And may I know why you have taken the decision to rip and after "applying a DSP 3db amplification after HDCD"

would be interesting asking to @spoon

thanks

Re: HDCD conundrum, yet agin

Reply #16
I used to use Exact Audio Copy, aka EAC, for ripping and it works very well, but it wasn't as fully featured as dbpoweramp so I ending up buying a dbpoweramp license, mainly to save me time and hassle ripping my decently sized CD collection.

I was a bit terse describing the waveforms in my above post.  The top stereo waveform is a direct rip of a CD that happens to be a HDCD disk.   The bottom waveform is the resultant output of hdcd.exe processing of that direct rip.  The middle waveform show the effect of applying a 6db gain to the output produced by hdcd.exe, which shows that such a gain will clip the waveform.

From your questions it seems to me that you are not grasping what HDCD is all about.  I'll leave it there as homework for you to figure out what's really going on here.