HDCD: an investigation

Same thing here. I tried with Mark Knopfler's Sailing To Philadelphia which definitely has peak extension.

 

Because bit depth isn't just about dynamic range. It's also about the resolution at a given volume level. For example, if you have a sample that's at -48db on a 16 bit recording then only 8 bits are used to represent that sample.

So greater bit depth equals more resolution, especially at lower levels.

 

This is what I found on my own when I ripped "Testify To Love" by CCM group Avalon using the greatest hits compilation. The original HDCD decoded file although "Detected HDCD" on hdcd.exe didn't show up the decoded file is only a 17 bit file, but it is just as clipped and compressed as the 16 bit playback, so when I ran clip restoration, it gave me a 32-bit float file with a few more clips to restore and the remastering sounds better than the original file it came from.

 

I agree, actually. Far better would be if they just mastered the albums wiht full dynamic range for 16-bit delivery in the first place....without the need for 'HDCD' processing. Just using the HDCD A/D converter would be fine. In cases where all that appears to happen is that the levesl are backed off ~6 dB, that's probably the case. However, so far, these tend to be albums that have been dynamically smashed (e.g. Beck). The King Crimson series may be an exception. For HDCDs where the logo lights up only because the HDCD A/D converter was used, playing them over a regular CD player is arguably BETTER, because you don't get the -6 dB level drop. Unless the undecoded file has lots of peaking near 0 dB, in which case you might want some level cut.... <headache>

So, I'm only going through this because to get the 'well mastered' version of some of these CDs -- the ones with peak extension -- I have to decode them.

 

Btw, when I use hdcd.exe and WMP plugin to decode the same ripped HDCD track, the results are virtually identical - they null out down to something like -88 dB. So I'm 99% convinced that it doesn't matter which app is used...and HDCD.exe outputs its files in a more player-friendly 24-bit .wav format.

 

Now, take that 16-bit version, but this time made with noise-shaped dither....et voila!

 

I'm very happy to be able to decode and burn to a new cd, but I find it maddening that Hdcd is really a big boondoggle.

Judging from the discs in my collection, very few take advantage of what decoding brings. Jimi Hendrix Ultimate Experience, some Joni Mitchells, Freakwater are the only hdcd encoded discs that seem to have peak extention.

Most others - whether labeled or not - don't appear to benifit at all. I've got a couple of discs that don't see an increase in dynamic range, but seem to take advantage of the extra bits when the signal calls for it, but the rest don't see any changes except for lower gain. Even one of the Joni Mitchell albums - For the Roses - doesn't see a bit change.

 

So these ones stay at 17-bit when decoded?

 

No, you lose 1 bit.

 

Source for this.

 

Only some of the time though!

 

Read any book on how PCM works - the more bits you use to quantize the amplitude at each sample point then the more accurately you are approximating the "true" amplitude.

 

Still, it IS worse to playback a HDCD on a non HDCD player. That's a fact.
HDCD is, according to me, a joke. DVD-A is a MUCH MUCH better system.

 

Yeah, on paper it's not the best idea I ever saw. Still, Steve seems to think he can make CDs using it without compromising playback on a non HDCD compatible player so I guess there is more to the implemention of the HDCD options than meets the eye:


http://www.stevehoffman.tv/forums/showpost.php?p=834299&postcount=86

http://www.stevehoffman.tv/forums/showpost.php?p=834742&postcount=94

 

No, it's not. Not always. Some "HDCDs" only trigger the HDCD indicator because the Keith Johnson/Euphonix A/D converter was used. There's nothing to be 'decoded' in this case, it's just a way of advertising that that particular high-quality A/D stage was used in the mastering.

 

I'll bet it depends on the cd player. A Shanling, Arcam or Cayin
HDCD player would be pretty serious machines to play back cd's on.

 

yup

 

I don't see how that is. With many, they don't incorporate peak extention or increase the bit depth. I don't see how in these cases hdcd decoded adds a thing other than being a very good digital filter.

 

I noticed the one 2001 two-fer Beach Boys track ("The Little Girl I Once Knew") I've decoded so far does the same thing - stays at 17-bit throughout

The Pet Sounds Sessions tracks go up to 20-bit on the loudest sections so I guess there is more benefit in decoding them.

Still, "The Little Girl I Once Knew" definitely sounded better decoded so maybe there is still more than just the -6dB going on?!

 

My question is in these cases, does dithering down to 16bit (ala uv22) work properly? Does the sliding bit range screw things up?

 

So, is this what Steve did on the gold AF discs - just used the Pacific Microsonics A/D conversion?.......

Steve?........

 

My guess is that he didn't use any of the features that require decoding. I'm embarrissed to say, though I have many of his hdcd discs, I haven't run them through the computer yet.

 

Great minds think alike!

 

I've no idea - however my guess is it will work fine. These are 24-bit wav files after all - even if all the bits aren't used all the time.

 

I eagerly await your test results.......