Head-Fi.org › Forums › Equipment Forums › Dedicated Source Components › Upsampling vs. HDCD
New Posts  All Forums:Forum Nav:

Upsampling vs. HDCD

post #1 of 11
Thread Starter 
Tonight, I was listening to Yes' Magnification CD, which is HDCD encoded. My Cary player does both HDCD and upsampling (upsampling can be engaged or disengaged); note that HDCD decoding only works when the upsampling is disengaged as I believe different digital filters are used.

Here's the thing see - just out of curiosity, in the middle of the CD, I turned on the upsampling (thereby disengaging the HDCD decoding). I actually preferred the upsampling to the HDCD playback. The HDCD was a little more dynamic and upfront, and there was perhaps a shade more detail, but the soundstaging seemed to clump around the speakers, and I lost the spaciousness that upsampling creates. With upsampling, the music was more cohesive and there was more depth (especially to the drums which were placed considerably farther back in the soundstage). Jon Anderson's vocals had more body and seemed more real. The differences were not super subtle. I kept switching back and forth once in a while during the rest of the CD and consistently preferred upsampling. Smoother, less digital, more musical. Disclaimer: I only tried this on this disc; maybe with other HDCD discs I would feel differently.

Am I crazy to prefer the fabricated, "interpolated" facsimile of extra information provided by upsampling to the actual extra information provided by HDCD encoding?

Has anyone else (Jude?) tried this experiment? I'd be curious to know what others thought of this.
post #2 of 11
well, I've read more than one review of the Musical Fidelity upsampling DAC, so I don't think you're crazy.

keep posting, though.
post #3 of 11
recephasan mentions the MF DAC...their Nu-Vista CD player also does upsampling and got a rave report in Stereophile, for making CDs sound close to vinyl. The guy who designed it apparently was originally planning to design an SACD player but got fed up with the lack of software, so instead he tried to create a CD player that would maximize CD sound quality, with apparently good results (I've never heard one myself).
post #4 of 11
John,

Yes, I have compared upsampling DACs versus HDCD-capable player with HDCD-encoded CDs. The Music Hall MMF CD-25 does give up -- even with HDCDs -- the big soundstage that both upsampling DACs I have on loan provide. However, the overall detail retrieval has so far been better on the Music Hall with HDCDs than either upsampling DAC with the HDCDs I've used for comparison (both Alana Davis albums, Tool's Lateralus, and Dave Brubeck Quartet's Time Out). Though less spacious sounding, more information seems to be coming out of the Music Hall with HDCDs than either upsampling DAC.

Given the choice with HDCD-encoded CDs, I so far prefer the Music Hall MMF CD-25 to either upsampling DAC. Admittedly, though, I don't have too many HDCD-encoded CDs.
post #5 of 11
Thread Starter 
Thanks, Jude!

I've been playing with my HDCD discs a little more and have reached much the same conclusions. With upsampling, there is more spaciousness present in regards to soundstaging, more smoothness and perhaps more palpability/realism to the vocalist and/or instruments. However, with the HDCD running, there is more detail, more crispness and a more dynamic presentation.

A matter of taste, it sounds like. I'll have to keep switching back and forth and see which I prefer ( I have about 20 or so HDCD encoded discs).
post #6 of 11
As I understand it, HDCD encodes the music with 20 bits of data as compared to the standard 16 bits. HDCD claims to be able to encode an additional 4 bits of data using their system. (see http://www.hdcd.com/about/index.html). Using HDCD supposedly provides 20 bits of real data compared to dithering, interpolating 16 bits to 24 bits.

Hence, it would not be surprising that HDCD CDs played back through a HDCD DAC provides more detail and more information
post #7 of 11
Yes, upsampling will provide you more "return on your investment" than HDCD decoding, if for no other reason, than because so few CD's are encoded with HDCD.

However, I prefer HDCD decoded. I end up just plain not liking HDCD undecoded and at this point, refuse to listen to them that way.

Usually (99% of the time), they are mutually exclusive, and what I've been telling most people is that upsampling is better, because it will give you the greater return on all your CD's, not just a small portion of them.

But I'm one of those freaks who needs that 100%, so I got the same thing kwkarth has, a Kenwood RD-VH7, which is one of the few devices I know of that provides HDCD decoding into the digital realm (the only other that I know of is the Pacific Microsonics HDCD Processors One and Two -- and they are ridiculously expensive). So in this rare occasion, it is doable.
post #8 of 11
Quote:
Originally posted by ddriveman
As I understand it, HDCD encodes the music with 20 bits of data as compared to the standard 16 bits. HDCD claims to be able to encode an additional 4 bits of data using their system. (see http://www.hdcd.com/about/index.html). Using HDCD supposedly provides 20 bits of real data compared to dithering, interpolating 16 bits to 24 bits.

Hence, it would not be surprising that HDCD CDs played back through a HDCD DAC provides more detail and more information
The following is from one of the FAQ's on the HDCD website:

Q: In simple terms, how does the HDCD Decoder operate with the HDCD code data?
A: The HDCD Decoder examines the LSB of the audio data for each channel and determines if there is the HDCD hidden code present. If so, it resets the code detect timers for each channel and decodes the commands. An HDCD code detect signal is made valid, thus identifying the audio as having been encoded with the HDCD process. Following code extraction, those parameters that affect dynamic range are restored. This includes the expansion of the instantaneous soft limiting peak levels (PE) performed by the encoder if that option was enabled and expansion of the low level gain compression(LLE) based on average signal levels. These exactly complementary gain command instructions are timed to the audio signal so that gain changes are transparent. After the dynamic range of the audio signal has been restored, the HDCD Decoder then performs gain scaling and filter selection. The final step in the decode process involves interpolating the audio signal to twice the sampling frequency using a filter
complementary to the encoders anti-alias filter. This audio signal is available as an output to the decode process, or within the HDCD Decoder, this audio signal can be interpolated to four or
eight times the sampling frequency to drive common A-D converters.

So the system does not actually record 20 bits (otherwise the CD would not be Redbook compatible). It sacrifices the Least Significant Bit (LSB) to provide instructions to the decoder. The decoder is instructed which filter to apply to the audio being played, or whether an increase in dynamic range is necessary. They claim an "effective" 20 bit result. I'd say that depends on the music (i.e. for recorded white noise you should get only 15 bit resolution instead of 20; for music anywhere from 15 bits and up).
post #9 of 11
I basically agree -- decoding HDCD is not going to provide any information that isn't already there. But since a HDCD signal is already encoded, I feel that decoding it will make it more accurate to the original signal, so I still recommend decoding it.
post #10 of 11
I am not implying that HDCD does not offer advantage over redbook CD. It does.

However, HDCD seemingly claim that they have 4 more bits stashed somewhere. That is inaccurate. They should have said "the resultant audio quality is equivalent to a 20 bit sample".

BTW - I don't see a technical reason not to link an upsampling filter in series with an HDCD filter (of course the HDCD has to come first as it needs the LSB control data).
post #11 of 11
Quote:
Originally posted by AssafL
BTW - I don't see a technical reason not to link an upsampling filter in series with an HDCD filter (of course the HDCD has to come first as it needs the LSB control data).
Yeah, I don't either. It's stupid. But, having done it (Kenwood RD-VH7->MSB Tech Link DAC III), I can say that just about no-one does it. I was really holding out for hope against hope that the upsampling upgrade for the Camelot Uther v something Mk other was going to do it post-HDCD, but that fizzled...
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Dedicated Source Components
Head-Fi.org › Forums › Equipment Forums › Dedicated Source Components › Upsampling vs. HDCD