Peak Extend is one of several functions available in the chain of HDCD processing which can be applied to a digital audio signal. It is an optional function - the mastering engineer may choose to turn on Peak Extend processing, or may choose not to use it. Whether to use Peak Extension processing is a decision made by the mastering engineer; so, yes, some discs use the Peak Extend processing and others do not. Your player "reads" the control information contained in HDCD-encoded audio, and from this knows if Peak Extend has been used and therefore if appropriate processing is required while decoding on playback. There has been an enormous amount of assumption, misinformation and speculative posting about this topic. So, for precision it may be better to ask: can anyone cite definitive information in which AF has confirmed specifically that none of their HDCD-encoded CDs utilize Peak Extend processing? You might wish to read a bit about HDCD amplitude processing (the AES paper is very good, see esp pdf pp12-13). Perhaps you don't realize that Peak Extend is soft limiting applied only on the highest levels of infrequent peaks? In addition to the resources noted a few posts earlier, another easy read is this Process Overview which explains the sequence of functions available in HDCD processing.
Another one with Peak Extend: Marc Ribot y Los Cubanos Postizos - The Prosthetic Cubans. Atlantic, 1998
So why does Kevin even bother with HDCD processing for the titles that he masters for Audio Fidelity? In other words, what benifit does the HDCD technology bring to the table for said Audio Fidelity discs?
Great question. It would seem that Steve and Kevin must differ in their opinions of the benefits provided by HDCD.
So, if you play an HDCD-encoded disc with peak extend on a quality HDCD-decoding player and then play it on a quality non-HDCD-decoding player (i.e., trying to take relative player quality out of the equation), do you hear a significant improvement on the HDCD-decoding player?
What confuses me is that for years, Steve talked down HDCD, and then suddenly a few years ago touted HDCD saying he had found a system to employ in his mastering that benefited from HDCD encoding. If he was not employing peak extend, then where was the benefit coming from? And it seems that he is no longer using HDCD (Kevin still is), so what changed Steve's mind again on HDCD?
Here's the thread where he mentions the decision to no longer use HDCD: I've decided not to run my masterings through an HDCD converter..
Whether an HDCD uses peak extension can be easily checked with hdcd.exe or a foobar plugin, so why would we need to ask AF? I've got about one third of AF's CDs and can confirm that none of those uses peak extension. Or are you saying that the tools I've mentioned can't reliably detect this feature after all?
Yeah, I remember that (long thread). Still, he seemed to really like HDCD at one time. I see why he no longer likes it, but what made him like it before? Was he using different equipment? It seems that Kevin Gray likes HDCD, or at the very least, feels it has no deleterious effect since he still uses it. Maybe it's the latter since he is not using peak extension.
Yes (Fragile, Close to the Edge, Yessongs, Tales of Topografic Ocean, Tormato) Japan Cd with mini lp sleeve (first and 2nd releases) are HDCD with peak extension. Most people says these are bad sounding becuase they look at the wav. These are very nice imo http://www.dr.loudness-war.info/index.php?search_artist=yes&search_album=hdcd
I'd say it would depend on how extended the peaks are. HDCD allows for a 6 dB gain in peak level, but many discs are well below that. I've seem around 2 dB on the Buffalo Springfield discs, and that's only on the *highest* peaks. Here's the description of the two features from the link above: HDCD decoding thus does 2 things: 1) Expand peaks that were mildly compressed by encoding. 2) Reduce the level for content under -45 dBfs. #2 is only going to be audible for fades and extremely quiet passages. #1 may happen often, but only for brief bursts. 10 samples here, 20 samples there...maybe 60 samples somewhere else. Most of the audio will essentially be untouched by the decoding process. By the way, the low level extend is why the hiss seems to get louder at the end of some fades (I've always noticed this on the Buffalo Springfield discs) - it *does* get louder, unless decoded.
To have some of the transients back doesn't mean that there's anything different with the overall dynamics.. Martin
Check the post above yours. Getting up to 6dB of dynamics back sounds fine to me. Here is the typical example of brickwalled CD encoded with HDCD Peak Extend feature BTW...
Regardless of what you've read as per further back in the thread Discs like Green Days nimrod (DR6 undecoded!) applies heavy (soft) limiting to every single drum beat. I'd still class heavy handed soft limiting as brickwalling. Even for infrequent peaks its all about making a louder master. However I have noted that the slogan that used to be on all their discs: 'The original dynamic range of this recording was not maximised, brickwalled, limited or compressed in anyway during remastering' Isn't on the newer discs, so I guess can't guarantie that peak extend isn't being used if audio fidelity are now using limiting. http://www.amazon.com/Deep-Purple-R...=sr_1_2?s=music&ie=UTF8&qid=1335506652&sr=1-2
I think one reason that blurp is gone is because "The original dynamic range of this recording was not [sic] maximized" is the opposite of what they meant to say.
I've never looked at the waveforms of the mini-lp HDCD Yessongs - I think it sounds bad because of the ****load of heavy-handed noise reduction, followed by over-bright eq settings. Give me the 1994 remaster any day (or even the original CD, tracking errors and all.) Just my two cents...
Add Rare Cuts & Oddities 1966 by the Grateful Dead to the list of HDCDs with peak extend (good luck finding it, though!)
Figured it out! If I put the CD in the drive and scan a track for HDCD using Foobar2000, it will give you a detailed analysis of the HDCD encoding. Takes about 10 second.
I believe that's only if there are features in the first few seconds of each track. If Peak Extend (or Low Level Range Extend) isn't used until later in the track it won't be detected.
I have an case of a CD, which is HDCD mastered, but only seems to be recognized by WMP and dbPoweramp, it's the soundtrack for Star Wars Revenge Of The Sith. It's not stated as HDCD mastered, there's no HDCD logo anywere, is not recognized as HDCD by my Oppo BDP-93, but there it is... Are this kind of discs/masterings very common?
I just decoded Pet Sounds tonight, and no track has peak extend enabled, according to my standalone decoder. I'm wondering what sonic differences are being heard? Wave forms look exactly the same for the encoded/decoded versions, once I normalized the levels of the decoded wav files. And I didn't hear any difference between the tracks encoded vs. decoded. -Steve
I wonders what features of HDCD used on Van Halen remasters from the Roth era, but they don't seem to sound good either decoded or undecoded.
