Red Book vs Hi Res

Posted by: AMA on 14 December 2009

I have a feeling (not strong enough to make it a statement) that high quality Red Book records (like XRCD24) can be indistinguishable from hi-res master tape.

I didn't make direct tests on the same records but for example Super Sound collection from FIM Studio in XRCD24 produced a very big impression on me.
When I first listened them on a very expensive tube setup it really lifted up the level of the 16/44.1 intrinsic capabilities in my book.
I was much more impressed than with the best SACD and hi-res PCM replays I had in the past.

Of course I agree that many hi-res records do sound better than Red Book -- but all of these cases are mostly related to the simple fact that hi-res tracks have been usually
recorded or ripped ... properly

Whenever you contest top-quality hi-res record with top-quality Red Book the difference is not that evident -- if any.

I'm waiting for Naim DAC to continue my home experiments -- trying to understand is there something essential in hi-res format which can unambiguously expose Red Book limitation.
I have a feeling that extending to 18/20 bit can increase dynamic range (when paired with top-quality amps and very resolving speakers) -- we can see this on HDCD records which do
a small trick to recover 20-bit bitstream from 16 bit data file.
But increasing a sample rate is definitely not that evident.

Any experience in this area?

Posted on: 14 December 2009 by js

I'm not with you on this one, AMA. I'm not a fan of HDCD in general and have heard original commercial pressing that I have preferred to their XRD equivilents. Early on in the Naim label process, HDCD encoding/decoding was compared to standard Redbook using an analog master and the redbook was much more like the original. Not as cool sounding perhaps but clearly more honest to the music.

As for XRD, this is a lot of converting back and forth to achieve a 'sound' and remastering is always an iffy proposition.

"The XRCD process begins at mastering. The original digital master is converted to analog using JVC's new K2 24 bit Digital-to-Analog Converter. The analog signal is then mastered using a custom mastering console and digitized via JVC's newly designed K2 24 bit Analog-to-Digital Converter .The 24 bit digital word then passes through JVC's Digital K2 ,which regenerates a pure 24 bit digital signal that is recorded to a SONY PCM-9000 magneto-optical disk.The XRCD process takes advantage of the stability of the magneto-optical disk,as well as its 24 bit capacity,by using it as the audio storage medium for delivery to manufacturing."

Seems to me that using your favorite DAC on an original dub regardless of the bit rate would be more honest to what the original studio work intended. I've also found the best modern digital mastering programs to be less intrusive than analog consoles. I can unequivically guarantee you that you can hear the difference between a 24/96 master and a very correctly downsampled one played from the same source. Really, all you need to do is download both the CD res and Hi Def versions of one of Ken's Naim label recordings to hear it if the setup is up to it. These are the same files with one downsampled the best way he could find and we tried lots of dither settings and types in the process to get it right.

Nothing wrong with XRD if you like the result but this is more a target function approach to a sound then an accurate transcription of the master.

Posted on: 14 December 2009 by Aleg

quote:
Originally posted by AMA:
I have a feeling (not strong enough to make it a statement) that high quality Red Book records (like XRCD24) can be indistinguishable from hi-res master tape.

I didn't make direct tests on the same records but for example Super Sound collection from FIM Studio in XRCD24 produced a very big impression on me.
When I first listened them on a very expensive tube setup it really lifted up the level of the 16/44.1 intrinsic capabilities in my book.
I was much more impressed than with the best SACD and hi-res PCM replays I had in the past.

Of course I agree that many hi-res records do sound better than Red Book -- but all of these cases are mostly related to the simple fact that hi-res tracks have been usually
recorded or ripped ... properly
Whenever you contest top-quality hi-res record with top-quality Red Book the difference is not that evident -- if any.

I'm waiting for Naim DAC to continue my home experiments -- trying to understand is there something essential in hi-res format which can unambiguously expose Red Book limitation.
I have a feeling that extending to 18/20 bit can increase dynamic range (when paired with top-quality amps and very resolving speakers) -- we can see this on HDCD records which do
a small trick to recover 20-bit bitstream from 16 bit data file.
But increasing a sample rate is definitely not that evident.

Any experience in this area?

XRCD is 'only' a type of mastering, in the end it is still downsampled to redbook specs. So I would say this qualifies for your 'properly mastered' remark.

I have 1 or 2 albums in Redbook spec and 88.2/24 master. I haven't truely AB-ed extensively but I definitely prefer the HiRes.

Furthermore I have heard very nice HDCD's and also not so nice ones. Also I heard bad XRCDs and nice XRCDs, so I'm not convinced the XRCD-technology in itself will make a better listening experience.

I haven't heard a bad HiRes yet, but I don't own that many yet.
Though I must say that 96/24 LP-rips aren't always so much better than Redbooks.

-
aleg

Posted on: 14 December 2009 by Joe Bibb

quote:
Originally posted by Aleg:

I haven't heard a bad HiRes yet, but I don't own that many yet.
Though I must say that 96/24 LP-rips aren't always so much better than Redbooks.

-
aleg

Properly mastered Redbook is superb. In fact, if it was all mastered as well as it can be, there would be no great demand for other formats, imo. The preference for vinyl has much to do with the fact that it was usually mastered well, compared to today's loudness victims - very well.

Joe

Posted on: 14 December 2009 by AMA

Hi-res is in the beginning of it's way.
That's why I try to avoid strong statements at this stage.
I have a LOT of hi-res recording including Red Book downsamples. I confirm that many hi-res sound better than Red Books. What makes me worry is that in my case hi-res reproduction sounds better on NEW records and also hi-res sounds better comparing to Red Books which in my case replayed on lower class source. So my tests are not fair. On the other hand I have a lot of examples when 16/44 sounds indistinguishable from the same hi-res record on the same streamer (Logitech Transporter).

To through in more considerations let me refer to the recent thread where likesmusic (thanks a lot!!!) has brought a link to the 1984 audio session audio session when Sony introduced first digital device PCM-F1 which was converting audio from analog source to 16/44.1 and then immediately converts it back to analog. The devices was placed into the chain between LP12 and naim amps (NAC32/NAP250) and controlled with switch. In the series of blind tests the listeners (one of them was major digital oppositionist Linn's head Ivor Tiefenbrun) were unable to distinguish when PCM-F1 was breaking the analog chain

At the very least this suggests that Red Book is intrinsically good enough to rival vinyl replay. If we rely on vinyl quality and can extend that PCM-F1 is inaudible in any analogue source (for example master tape recorder) than this proves that hi-res is excessive and all it's benefits are somehow related to the proper recording/coding/pressing implementation.

Posted on: 14 December 2009 by ferenc

quote:
Originally posted by AMA:
Hi-res is in the beginning of it's way.
That's why I try to avoid strong statements at this stage.
I have a LOT of hi-res recording including Red Book downsamples. I confirm that many hi-res sound better than Red Books. What makes me worry is that in my case hi-res reproduction sounds better on NEW records and also hi-res sounds better comparing to Red Books which in my case replayed on lower class source. So my tests are not fair. On the other hand I have a lot of examples when 16/44 sounds indistinguishable from the same hi-res record on the same streamer (Logitech Transporter).

To through in more considerations let me refer to the recent thread where likesmusic (thanks a lot!!!) has brought a link to the 1984 audio session audio session when Sony introduced first digital device PCM-F1 which was converting audio from analog source to 16/44.1 and then immediately converts it back to analog. The devices was placed into the chain between LP12 and naim amps (NAC32/NAP250) and controlled with switch. In the series of blind tests the listeners (one of them was major digital oppositionist Linn's head Ivor Tiefenbrun) were unable to distinguish when PCM-F1 was breaking the analog chain At the very least this suggests that Red Book is intrinsically good enough to rival vinyl replay. If we rely on vinyl quality and can extend that PCM-F1 is inaudible in any analogue source (for example master tape recorder) than this proves that hi-res is excessive and all it's benefits are somehow related to the proper recording/coding/pressing implementation.

Be careful, with a comparison like this. To make a difference between sources and sound you (your perception) needs to be "educated". I think we can make a difference between a phonograph and real live voice nowadays, but in 1915 the folks could not

"Recitals were conducted to prove the merit of Edison's Phonograph. Edison recording artists would sing along with a disc recording of their voices, daring the audience to be able to tell the difference. In late 1915, Edison began its famous Tone Tests, which featured artists alternating their live performance on a darkened stage with that on the disc in front of large audiences, challenging them to detect a difference. Reaction was positive to these tests, and reinforced the Edison motto that the discs were "re-creations" of performances, not merely recordings of them."

The History of the Edison Disc Phonograph

Posted on: 14 December 2009 by ferenc

quote:
Originally posted by js:
I'm not with you on this one, AMA. I'm not a fan of HDCD in general and have heard original commercial pressing that I have preferred to their XRD equivilents. Early on in the Naim label process, HDCD encoding/decoding was compared to standard Redbook using an analog master and the redbook was much more like the original. Not as cool sounding perhaps but clearly more honest to the music.

As for XRD, this is a lot of converting back and forth to achieve a 'sound' and remastering is always an iffy proposition.

"The XRCD process begins at mastering. The original digital master is converted to analog using JVC's new K2 24 bit Digital-to-Analog Converter. The analog signal is then mastered using a custom mastering console and digitized via JVC's newly designed K2 24 bit Analog-to-Digital Converter .The 24 bit digital word then passes through JVC's Digital K2 ,which regenerates a pure 24 bit digital signal that is recorded to a SONY PCM-9000 magneto-optical disk.The XRCD process takes advantage of the stability of the magneto-optical disk,as well as its 24 bit capacity,by using it as the audio storage medium for delivery to manufacturing."

Seems to me that using your favorite DAC on an original dub regardless of the bit rate would be more honest to what the original studio work intended. I've also found the best modern digital mastering programs to be less intrusive than analog consoles. I can unequivically guarantee you that you can hear the difference between a 24/96 master and a very correctly downsampled one played from the same source. Really, all you need to do is download both the CD res and Hi Def versions of one of Ken's Naim label recordings to hear it if the setup is up to it. These are the same files with one downsampled the best way he could find and we tried lots of dither settings and types in the process to get it right.

Nothing wrong with XRD if you like the result but this is more a target function approach to a sound then an accurate transcription of the master.

I am with you

Sometimes I got a feeling with these XRCDs that they are too colorful, polished, enhanced. These recordings sometimes try to impress you, not let you to be buried in the music. Like a too much saturated, artificially contrast enhanced film after post-production, finishing and color correction. It can be tyring. I have lots of XRCDs and original CDs as well, and in lots of cases I just prefer the original now, years ago it was different. It is only my view, so do not take it too seriously

Posted on: 14 December 2009 by js

quote:
Originally posted by Joe Bibb:
quote:
Originally posted by Aleg:

I haven't heard a bad HiRes yet, but I don't own that many yet.
Though I must say that 96/24 LP-rips aren't always so much better than Redbooks.

-
aleg

Properly mastered Redbook is superb. In fact, if it was all mastered as well as it can be, there would be no great demand for other formats, imo. The preference for vinyl has much to do with the fact that it was usually mastered well, compared to today's loudness victims - very well.

Joe

No one said Redbook couldn't be good and of course, I've heard HiRes that was pretty awful too but apples to apples it's a noticable step.

Posted on: 14 December 2009 by pcstockton

Yes, there is a marked difference between the two assuming the same recording/mastering.

That being said, I would rather listen to a Red Book if done well, over a subpar Hires, any day of the week.

It is very much like HDTV. The Hires/def factor does not overcome shitty content.

I think this all goes without saying of course, and it is certainly not limited to 16/44 vs. 24/96.
I would rather listen to Barry Diament's original Zeppelin Atlantic masters on a Sony Discman, than the Page Remasters on a CD555.

Very inexpensive upgrade

-Patrick

Posted on: 15 December 2009 by Joe Bibb

quote:
Originally posted by js:
No one said Redbook couldn't be good and of course, I've heard HiRes that was pretty awful too but apples to apples it's a noticeable step.

Yes but I'm guessing most people's interest will be mainstream artists or labels. In which case comparing those apples will be very tricky. Unless there is a sudden (never previously apparent) interest by the main music owners in the end product - you will be left with kind of lobby muzak, minority interest dirge or free form jazz (nice) that usually infests audiophile records.

If I found myself searching out audiophile recordings just to explore the potential of higher res formats - I'd consider the tail was wagging the dog. My admittedly limited exposure to comparisons of the very best Redbook with higher res formats has left me more impressed by what Redbook can do, when done properly, than anything else.

Don't get me wrong, I have the ability to replay it - I just don't see too much being released in hi res, that will be of any interest. I hope I'm wrong.

Joe

Posted on: 15 December 2009 by js

Me too and it's worrisome. It's why I said try some of the Naim downloads to compare. There's some Grammy winners there along with heart felt playing. Mainstream is never driven by the top of a market but like 1080 in vid, I suspect that at some point HiDef may also become an audio necessity for this end of the market. People had to see 1080 to buy even if 720p was better than 1080i. We can hope but like cassettes of years past portability will always be the strongest portion of the audio software market. The future is impossible to predict. We could even eventually see things become more polarized with MP3 and HiDef becoming the favored formats and CD res least popular on new releases.

Posted on: 15 December 2009 by pcstockton

quote:
Originally posted by Joe Bibb:

Yes but I'm guessing most people's interest will be mainstream artists or labels. In which case comparing those apples will be very tricky. Unless there is a sudden (never previously apparent) interest by the main music owners in the end product - you will be left with kind of lobby muzak, minority interest dirge or free form jazz (nice) that usually infests audiophile records.
Joe

The key word above is "records". Especially if you dont own a nice TT. 24/96 vinyl rips from stellar releases are amazing and maybe the best of both worlds. The convenience of hires digital, legendary recordings, unremastered etc...

I agree that the in current mainstream, the quality is being sacrificed for convenience. I wouldn't expect hires audio to take the world by storm anytime soon.

Maybe if as culture(s) we appreciated music as much as NASCAR, Oprah, and Survivor, we would already be there.

Myself not excluded. I bought an HDTV before I ever heard of "Red Book" or 24 bit files. CDs were CDs then....

Posted on: 15 December 2009 by AMA

May I ask for the very particular question which I believe will help the discussion.

Did anyone compare the same record in 16/44.1 and 24/96 (or another hi-res) on the top-quality sources and found the hi-res to be superior?

If so -- can you mention your source, pls.

I suggest (for some ethic reasons) to disqualify the cases when SACD/CD combo demonstrates SACD superiority over CD

I can shrink the criteria further down: can someone report that CDS3 replay of 16/44.1 was trashed by some streamer or SACD playblack of the same record in hi-res???

Posted on: 15 December 2009 by Aleg

quote:
Originally posted by AMA:
May I ask for the very particular question which I believe will help the discussion.

Did anyone compare the same record in 16/44.1 and 24/96 (or another hi-res) on the top-quality sources and found the hi-res to be superior?

If so -- can you mention your source, pls.

I suggest (for some ethic reasons) to disqualify the cases when SACD/CD combo demonstrates SACD superiority over CD

I can shrink the criteria further down: can someone report that CDS3 replay of 16/44.1 was trashed by some streamer or SACD playblack of the same record in hi-res???

I played one album of which I have a 44.1/16 recording and a 88.2/24 recording from the same source and found the 88.2/24 recording superior esp. in clarity, detail and depth of space.

They were both played from my low grade source

Ubuntu 9.10 NAS-->PCH-->Beresford DAC --> NAC152XS(<--HiCap-2)-->NAP155XS-->Wilson Benesch Square Two.

Maybe on my Naim DAC (on order) it will even sound better, and maybe still even better from any future Naim streamer

, but even in this setup I preferred the Hi-Res version.

-
aleg

Posted on: 15 December 2009 by John R.

Yes, I compared 16/44,1 to 24/88,2 and to 24//176,4 with a HDX via USB stick and to my ears the higher resolution versions were way ahead: better dynamics (both macro dynamics and micro dynamics), more natural timbre with acoustic instruments and voices plus the feeling of simply more energy in the room meaning that a grand piano sounded more like a real grand piano.

But to my ears there was no real difference when comparing 24/88,2 to 24/176,4 meaning that sample rates higher than 88,2kHz or 96kHz might not be useful - at least with my setup and my ears

Same experience in a 100% AYRE system using their asyncronous USB DAC (QB 9) and 16/44,1 and 24/96 files via MAC.

Posted on: 15 December 2009 by pcstockton

quote:
Originally posted by AMA:

Did anyone compare the same record in 16/44.1 and 24/96 (or another hi-res) on the top-quality sources and found the hi-res to be superior?

I have 24/96 needle drops which have also been dithered down to 16/44, so i could burn to a CD for a friend.

So the 16/44 and 24/96 are not only from the same release, but also the exact same transfer process, and the same time.

The hires file is much more exceptional in every case.

I had a few pieces I transferred originally in 16/44 then later in 24/96 when I learned of doing such. Of course I trashed the 16/44s so I cannot comment on any differences.

The source was a Music Hall turntable for most of the 24/96s, I believe. I dont recall what model it was. I think I remember it having stacked plinths underneath.

-p

Posted on: 15 December 2009 by AMA

pcstockton, I guess you have used Beresford DAC for replaying 16/44 and 24/96 with bitstream supplied form PC or Mac?

Posted on: 15 December 2009 by pcstockton

Yes indeed. Foobar > USB > M-Audio Transit > Toslink > Beresford > Chord RCA-to-DIN > 102

Posted on: 15 December 2009 by AMA

quote:
Nothing wrong with XRD if you like the result but this is more a target function approach to a sound then an accurate transcription of the master.

js, it's not actually about XRCD -- this was just an example. There is a good number of high quality Red Book pressings which sound flawlessly on CDS3.
I'm curious if the same records in hi-res will sound better through HDX or Naim DAC?

Posted on: 15 December 2009 by AMA

quote:
Yes indeed. Foobar > USB > M-Audio Transit > Toslink > Beresford > Chord RCA-to-DIN > 102

OK, and you are not alone who liked hi-res more through Beresford DAC. This may only mean that Beresford DAC is much better in handling hi-res than in 16/44

Honestly, let's test hi-res against the reputable Red Book veterans like CDS3 -- this will give more insight in the inherent potential of 16/44.

Posted on: 15 December 2009 by Aleg

quote:
Originally posted by AMA:
quote:
Yes indeed. Foobar > USB > M-Audio Transit > Toslink > Beresford > Chord RCA-to-DIN > 102

OK, and you are not alone who liked hi-res more through Beresford DAC. This may only mean that Beresford DAC is much better in handling hi-res than in 16/44

I have modded the output stage of the Beresford to best match the sound I prefer. I have tried several opamps before I found the one I liked best, so it isn't only about handling HiRes or Redbook.
And I can't by any measurement imagine a DAC being better at handling HiRes then Redbook. Can you bring any argument to table to substantiate such a thought

??????
I must say it almost sounds that if it doesn't match your findings then it must be something else. I am totaly flabbergasted!!!

quote:

Honestly, let's test hi-res against the reputable Red Book veterans like CDS3 -- this will give more insight in the inherent potential of 16/44.

But this won't let you test HiRes from the same source!!

-
aleg

Posted on: 15 December 2009 by AMA

quote:
And I can't by any measurement imagine a DAC being better at handling HiRes then Redbook. Can you bring any argument to table to substantiate such a thought Eek ??????

This is exactly what happening all around.
Most SACD/CD combos are so poor on Red Books that most of the high claims on SACD are mostly a consequence of serious technical challenges on integration of SACD and CDP with giving priority to the hi-res. One of the technical point for hi-res orientated DACs to dominate over 16/44 is that that hi-res is NATIVE for them while 16/44 is first upsampled (with a very primitive mathematical algorithms to allow run-time processing) and only then pushed to the hi-res converter.
At the same time many serious audiophiles design DACs which convert 16/44 in native format -- without upsampling. The other experienced designers specializing on Red Book replay design sophisticated upsampling methods in order to benefit from hi-res converter.
This is why you may find a lot of people who spend 30 K$ for Red Book source and value it's more than 10$K SACD.

Anyway when we talk about the technical limitations of 16/44 -- we have to sort out if these limitations are inherent to this format. We have to check this on the top Red Book performers.

Posted on: 15 December 2009 by AMA

quote:
quote:

Honestly, let's test hi-res against the reputable Red Book veterans like CDS3 -- this will give more insight in the inherent potential of 16/44.

But this won't let you test HiRes from the same source!!

It can't be the same source anyway -- 16/44 DAC and hi-res DAC are absoluetley two different physical devices even if you put them in the same box and call it a "universal player".

In fact we don't need the contenders to be played on the same source.
I only suggest to play 16/44 on the top-rank CDP and then play the same record in DSD or PCM24/96 on the other top-rank source.

Let's see for example if HDX/555PS will be worse than CDS3 on 16/44 and better than CDS3 on hi-res.
This will pretty much answer the very question.

I have finished my own exercises on various DACs (both non-reclocking and re-clocking) and now waiting for Naim DAC to repeat the tests again and post on findings.
Unfortunately I don't have a CDS3 at home for side-by-side contest but I believe Naim DAC is good enough for 16/44 replay to be a reference for the combat.

Posted on: 15 December 2009 by js

quote:
Originally posted by AMA:
quote:
Nothing wrong with XRD if you like the result but this is more a target function approach to a sound then an accurate transcription of the master.

js, it's not actually about XRCD -- this was just an example. There is a good number of high quality Red Book pressings which sound flawlessly on CDS3.
I'm curious if the same records in hi-res will sound better through HDX or Naim DAC?

I believe I gave other methods and examples of more similar types.

Posted on: 15 December 2009 by AMA

quote:
I believe I gave other methods and examples of more similar types.

js, not actually. I'm trying to attract your attention that accurate downsampling to 16/44 is only half of the job (important of course).

Next point is that many DACs today upsample 16/44 to 24/192 while the 24/192 records avoid this procedure and naturally benefit from this.
In case of run-time upsampling is not implemented properly you may find that 16/44 sounds clearly worse than 24/96 -- which proves nothing in terms of intrinsic 16/44 abilities.

I personally see a value of 24 bit resolution which increase dynamic range in quite an audible way.
I doubt that all 24 bits are involved in this pleasant effect -- I bet only 20 bits or even 18 bits actually do the job. Besides I have some serious concerns on the ability of modern amps and especially loudspeakers to resolve 24 bit dynamic range (145 dB!) even if it's properly resolved in digital domain (which is hardly to believe as low bits are usually still mistaken).

When it goes for the sampling rate -- this makes me even more uncomfortable as I actually ... don't hear an audible difference. I have found this one day when I was completely unable to resolve between 24/48, 24/96 and 24/192 files in a blind test! (while 16/44 vs 20/44 was clearly audible).
Possibly I got a problem with my DAC, or my file coding, or .. my ears

.

Now important observation -- the widely accepted efficiency of "upsampling" on 16/44 is mostly relates to the improving the jitter spectrum by shifting it to the higher frequency range. Sound becomes more spacious and airy and seemingly more clean and detailed.
Meanwhile the simple upsampling produces some negative effect on jitter-full bitstream adding more sibilence and pre-echo to the sound.

If using a proper re-clocking DAC these effects will diminish giving a natural sound clarity of the jitter-free bitstream. Now I would like to see if Naim DAC will still show 96 kHz is audibly better than 44 kHz ?!

Posted on: 15 December 2009 by ferenc

I think I linked these samples sometime a year ago here, but here they are again.

Live guitar recording by an internationally acclaimed guitarist. The original was recorded in 128 x DSD (5.6 MHz 1 bit sampling) with a Korg MR-1000 portable pro hard disk recorder. Mic was a DPA pair, with some tube Ear - Yoshino mic preamp, cables are XLR balanced Nirvana SL, including the mic cables. The samples are converted from the same 5.6 MHz 1 bit original DSD recording, using Korg Audiogate software and its conversion.

Live Guitar Conversion Samples

The microphone position was a kind of spontaneous as the artist moved it from the well balanced position, when started to play