SACD vs HDCD

Are you sure that you are not mixing-up SACDs with HDCDs? HDCDs should play on literally anything. But SACDs can do what you describe if you have your player connected incorrectly.

 

Let me say up front that I am not an expert on HDCD features. But I have some educated guesses at what the answers might be based upon my admittedly limited understanding of the technology.

I find it to be very odd that any player could have an HDCD logo and yet not support the peak extend feature. This makes me wonder if that player only claimed HDCD compatibility, but did not decode them at all. In that regard it would be no different than any other redbook CD player, and it would mere represent some deceitful marketing.

If it genuinely has an HDCD logo, then I wonder if this comment was only in reference to its sp/dif outputs? Because it is theoretically impossible to fully represent the peak-extended signal in only 16 bits without lowering the volume by 6dB or without converting the signal to either 20 or 24 bit resolution, which is incompatible with many older DACs.

Perhaps you should try using software to decode the peak extend feature. Free software decoders are available if you have the ability to play wave files through your system.

That is one point of view. Like many others have stated, those Pacific Microsonics ADCs were really superb sounding ADCs, so even using them and not taking advantage of any of their HDCD features which required decoding could be beneficial. Some of the variable filtering algorithms were which the HDCD supported did not require decoding to provide an audible benefit, or so at least it has been claimed.

Engaging the peak extension feature is effectively a means for an engineer to effectively release two slightly different mastering's of a given recording at once. Undecoded would have more compression, decoded would have less. So this is a way to giving audiophiles a bit more kick, if they listen with an HDCD capable player as compared to the general population who would probably prefer more compression anyways.

But if a mastering engineer either didn't want for a given mastering to sound different depending upon how it was played back, then they would not have opted to use peak extension.

As I stated above, as I understand it, some advanced variable (presumably anti-aliasing) filter algorithms were used inside of these ADCs which were designed-to sound better than conventional fixed filters could with all signals. So these might have been sonically beneficial, even if nothing special was done on the play-back side of things to decode HDCDs.

 

The Pacific Microsonics ADCs were very highly regarded when they first came out in the mid-1990s and at that time there wasn’t really anything that could touch them.

While they still produce good results by today’s standards they have, as you point out, been out of production for 25 years, during which time many more good sounding ADCs have appeared. So they are perhaps not the be-all-and-end-all they once were.

IIRC, when Abbey Road were preparing the latest Beatles remasters they chose a Prism ADC as the best sounding in a shootout between - among others - the PM units. And that was a decade ago - there are likely even better units on the market these days.

I believe this is only true of the early HDCD-capable players that used the Pacific Microsonics PMD-100 and -200 decoder chips, which did indeed contain a very highly regarded digital filter section.

But Microsoft ended hardware support for HDCD when they bought PM around the turn of the millennium and those chips have been unavailable since the mid-2000s.

If you wanted to offer HDCD capability on your player after that, MS would licence you the decoder code if you could find a chip that was capable of running it (there were several, but the last one ceased production in 2017 - hence why Oppo and others dropped support on their machines). AFAIK, those chips used their own digital filter section rather than the HDCD one on the earlier, bespoke chips.

 

Perhaps the engineer was using the peak extend feature to effectively provide two masterings in one release: One would be loud for the masses (when undecoded), and the other would still sound good for those who had the ability to decode HDCDs.

Frankly this approach sounds preferable to todays approach of compressing the living daylights out of everything.

 

I guess it would be preferable if the more dynamic version was available in a generally playable format rather than something like HDCD.

 

Never owned a SACD, and never heard of a HDCD??????

 

That was a major part of the marketing blurb by Pacific Microsonics for the HDCD format. Unfortunately, quite a few HDCD releases do not use the Peak Extend feature nor the engineers really cared to use that feature much. I believe in previous discussions on this forum that the licensing fees are quite high as well.

 

Unfortunately that idea goes against the modern trend which is the loudness wars.

 

Thanks to TarnishedEars and Redeyedandblue for responding. Could be deceitful marketing by Panasonic in regards to my player. I guess the same could be said about some CD's labeled as HDCD that didn't fully utilize those features and why some labels chose not to stamp their discs or packaging even though those discs will trigger the HDCD indicator on players that are equipped. I think the idea of two masterings on one disc is spot on a part of the original intent with HDCD giving some the advantage of better audio quality while providing the more compressed louder version for the remaining 99%. Too bad this was not standard issue on all CD's.

Another useful quote I found regarding the volume with Peak Extend.

"As for volume level differences, that comes with the territory. If an HDCD employs the peak extend function, it will possess greater dynamic range, hence lower average volume level when decoded. A +6 dB level shift never should be utilized with peak extend – because that will result in digital clipping on some tracks. So, turn up an analog volume control downstream, adjust to listening at a lower volume, or disable HDCD decoding. And in that last case, HDCDs that do not employ peak extend or low level enhancement do not require any decoding in the first place. Those discs are HDCDs only because they were converted with a Pacific Microsonics ADC."

 

I'm not sure I agree that there is no "inherent" advantage of an SACD converted from PCM. Whilst I agree it isn't the perfect conversion, SACD is a high rez format and if the PCM source (converted to SACD) is also high rez then there is a definite and distinct advantage over a standard CD or HDCD.

 

So, IS there a software solution to play these? For those who have ripped to hard drive, or don't want to use an ancient disc player?

 

Both Foobar2000 and dBPoweramp (FB with a plugin, dBP using a DSP option) can recognise HDCD recordings and decode them to 44.1/24bit files. The Foobar2000 plugin will also show you what HDCD features are being used. If neither of the peak extend or gain features are being used, my understanding is that there is little point in decoding to a 24 bit file. I don’t believe the transient filters that form part of HDCD can be mimicked in software.

The HDCD plugin of dBP can produce unexpected results in that occasionally CDs that are not labelled as HDCD show as HDCDs, albeit without using peak extend / gain. Whether this means a Pacific Microsonic ADC was used I have no idea.

Where the Foobar2000 plugin shows that an HDCD has used peak extend, I have ripped it as both a normal 16bit FLAC and a 24bit one. I haven’t had the time to compare the two side by side.

 

without a HDCD player... you don't have vantage of that kind of encoding...

 

This is why ripping and decoding HDCDs with peak extend can be a bit of a pain - add this greater dynamic range to the empty bits you'll get when you decode to a 24-bit file and the result will be *significantly* quieter than everything else in your digital library.

My method is to rip the whole CD as a single lossless file with cue sheet, decode the HDCD content using ffmpeg, then normalise the resulting 24-bit file to something close to 0dBFS using an audio editor. That way, you preserve the relative dynamics between the different tracks and can always split your rip into individual files for each track afterwards if that's how your library is organised.

Like I say, a bit of a hassle (especially when you decode the peak extend only to discover that the source material was brickwalled before it hit the ADCs anyway - found at least a couple of discs like this!) but IMHO worth it for the extra dynamics, given peak-extended HDCDs only account for around 1% of my collection.

The PM ADCs used switchable anti-aliasing filters that were automatically selected based on the amount of high frequency content in the source material - most of the time they used a filter that was rolled off in the high end but offered excellent time domain response; when they detected transients with a lot of high frequency content they switched to a filter that was effectively flat to 20+kHz but allowed some sampling artefacts.

The transient filter flag in the subcode on HDCD discs indicates when this switch took place and suggests there was a plan to use some kind of complementary reconstruction filter in the decoder chips, which *could* be mimicked with software. But - as far as anyone has been able to determine (and I'm willing to be corrected here) - this was either never implemented or yanked from the code before release. Vaguely recall reading somewhere that such a filter would have violated a rival company's patent and PM didn't want to pay to licence it.

 

I posted my comparisons here, comparing a track using HDCD Peak Extend and another without it: HDCD ripping properly to 24bit FLAC

Another interesting thing to note is that even a few albums may use Peak Extend during certain passages in the music, as I found out with the Ace Combat Zero: The Belkan War original video game soundtrack. If I have the time, I'll try to post some screenshots comparing the two with and without using PE.

 

I have yet to hear any HDCD's that blew me away. All the HDCD's in my collection are horribly mastered even with, even with decoding. The only really good one I've heard was the Best of Dire Straits, but that release has remastered versions and those pale in comparison with the original CD's. I've heard some great sounding SACD's, but most of that comes down to them just being really well done. At the end of the day, the quality of the format is secondary to the production value when it comes to sound quality, and HDCD doesn't have many great versions of the titles available. SACD is better in that regard, but because Redbook has been around so long and is so ubiquitous, it stands to reason that there's much better selection. The best sounding albums in my collection are mostly Redbook CD's anyway.

 

Thanks for sharing this. Can I ask which Audio Fidelity HDCDs you have heard?

 

There are quite a few HDCDs that are far from being "horribly mastered" (Chris Isaak, Mark Knopfler, Mike Oldfield, Neil Young, ...).

 

I have no idea which HDCDs you are talking about being so awful. Why not name some of these awful HDCD titles so we can understand what it is that you are talking about?

The majority of my HDCDs have sounded anywhere from good to great. And Reference Recordings made some jaw-droppingly good classical HDCDs IMO.

 

My Billy Joel 52nd Street HDCD mastered by Kevin Gray is really rather decent too!

 

The Independence Day soundtrack sounds amazing, even with or without Peak Extension enabled.

 

Indeed, but it was mastered by Steve Hoffman.

 

Flaming Lips - Soft Bulletin, Tool - Laturalus (Actually one of the better ones I own), Van Halen's old releases, Sigur Ros - ( ) *Probably the worst sounding one.

 

Haven't heard any of the Audio Fidelity ones. Only Audio Fidelity I've heard was an SACD of Evil Empire.

 

OK, thanks.