Do people believe Hi Res is better than standard cd?

EJS posted:

Agree with Max’ point that hires files don’t make up for a deficient playback rig. Those of us with a 555 shouldn’t worry they are missing out on much!

Actually, that was not my exact meaning: but yes, it was a joke about this idée fixe of HiRes. Taking the matter a little more seriously, having heard (it's on the internet) what mp3 compression removes from a 16/44.1 and having heard that it's largely parts of ambient resonances, echoes, 'air' between the sounds (which is obviously made of sound, not of silence), I'd be tempted to infer that HiRes can 'take in' more of the same, that is a 24/96 or 192 recording could have better overall rendition of the sound of the recording venue and of the 'breathing' of sound in space. After all, listening to mp3 only gives us more sharp edged sounds and less spatial details, which is why perhaps most youngsters are ok with it: given the music they listen to.

The big problem is that HiRes recordings are as manipulated and not more reliable than any other one, so we often risk to mistake a more prominent violin or greater dynamics for more music, while it is not necessarily even greater resolution. As for myself, I only trust two types of silver discs: AADs and DDDs of acoustic music played, produced and recorded with true understanding of what music is, digital is and what it can do. The rest is often just bad noise. 

Simon-in-Suffolk posted:

It’s also worth noting if the master recordings are sample rated at 96kHz it’s never going to sound optimum for 44.1kHz PCM... for no or minimum digital sampling noise errors the recording master should be a whole multiple of the target rate. Therefore CD should ideally be recording sample rate mastered at 44.1. / 88.2 / 176.4 kHz so as to produce least distortion. Likewise a master at 44.1 kHz is going to introduce distortion and artefacts if upsampled to 96 kHz.

I’m not certain this is true for non-naive / non-simple re-sampling schemes; integer multiples and sub-multiples are certainly easier to implement and imagine in your mind’s eye, but I think most re-sampling (in either direction) now invokes at least some level of native reconstruction first. Same thing applies in both time and amplitude domain, afaik. Recording industry equipment generally runs at the 192 kHz and multiples, but you have lots more control for changes when you downsample than just, for a wild simplification, discard every other point (or average, or whatever). I’d be interested to know if you have some insight on these processes, Simon, especially given your knowledge of the multiple internal conversions employed in the Naim architectures... and indeed living behind their “digital active” philosophy with the analogue inputs on the new Uniti line.

Regards alan

They deleted my post because of a link I pasted  about Peter Qvortrup‘s statement about highres & co. You can find the interview with him under parttime audiophile, than The Occacional Magazin (pdf for free). Frankly, this interview changed my view about hifi and the direction it goes now. Highly recommended!

Here one of his statements:

Digital Audio is the latest ex- ample of how the high- delity industry has distorted the con- cept of research, and improve- ment. Since the introduction of the Compact Disc in late 1982, the technology has entered into the usual numbers game. 20-bit is better than 16-bit, 96kHz is better than 44.1 kHz, 24-bit is better still, as is a 192kHz sam- pling rate, and so on. In a...

ralf

Alan, I was going to point you to some mathematical references... but perhaps not helpful...so I will try in English instead.. essentially where oversampling is involved ie where a discrete integer multiple is used there are no arithmetic  rounding errors, however where a non integer multiple is used when upsampling then rounding errors can occur... and rounding errors equals digital noise and artefacts.  The same holds true for decimation...(reducing the sampling rate)

As far as as Naim internal systems as far as I aware they usually use discrete multiple oversampling so as to avoid arithmetic rounding errors.

Simon-in-Suffolk posted:

essentially where oversampling is involved ie where a discrete integer multiple is used there are no arithmetic  rounding errors, however where a non integer multiple is used when upsampling then rounding errors can occur... and rounding errors equals digital noise and artefacts.  The same holds true for decimation...(reducing the sampling rate)

Thanks Simon.

I still think that this “rounding error” perspective is not mathematically correct except for what we might have to agree are “poor” (perhaps of practical necessity) re-sampling schemes. If enough points in any given sampled sequence are taken to reduce the error with respect to the Nyquist-limited source signal below the numerical resolution of the new sampling frequency (and/or bit-depth), then there is no effective rounding, nor any corresponding rounding error (at least with respect to the original function).

A brute force method using all the sampled points and twice the zero padding for the deconvolution, as an extreme and limit-case example, would have no rounding error, not just rounding error below the new sampling limits. This is the natural corollary to the Nyquist theorem, albeit even Rob Watts’ million tap filter doesn’t go anywhere this far!

In any event, thanks for taking the time to think and reply on this interesting topic. Your comment, which reflects many others’, that Red Book digitizations have a lot more to offer than the early or less optimized implementations reveal, is close to my own thinking. But if technology had been just a little more advanced than it was at the time, going with a higher resolution spec in both amplitude and frequency would have been “even better”. Whether it would have been the dominant limiting factor for commercial implications is another question - one which was refreshingly addressed by Benchmark Audio saying that all the spec improvements in their latest DAC -3 are nice on paper but don’t really lead to much, or any (!), audible difference relative to the DAC-2. 

I hope 2018 is good to you and yours!

Regards alan

French Rooster posted:

 

44.1 kHZ is a predominant sampling frequency. Higher res is also used - 96 kHZ being the next obvious step.

To prepare a CD - format musi a recording has to be converted to 16 bit - usually done after mastering, when a CD-specific file is being made.

I know you are qualified to answer, you are working in music and studio recording. You say that 24 bit is practically default setting and also that 44,1 khz is predominant sampling frequency.  So you say that most studio recordings in the world are recording albums in a more than 16/44 format ?  so minimum 24/ 44 format ?   i am surprised by that information because the industry is still producing cds and the digital format to buy albums is mp3 or 16/44 in the vast majority of commercial offerings.    I have also read that hdtracks or qobuz are selling in majority hirez that come from upsampled 16/44 files.

In essence yes - studios record / mix more information than is being ‘packed’ onto a CD.

24 bit depth is a norm (32 floating can also be used). It’s the bit rate that contains the musical information. So preparing a CD version of a digitial recording will necessarily take the bit rate down to 16.

Sampling frequencies used can be: 44.1 kHz, 48 kHz, 96 kHZ and higher. The higher the sampling rates the more demands this places on the entire recording / mixing equipment - all must be up to the same standard.

Mastering is a way of ‘preparing’ a mix for different  media - dynamic range is ‘squashed’ for a vinyl - otherwise a stylus would jump out of a groove; bit-depth is reducted for a CD to make the data physically fit onto a 700Mb storage space; mastering for iTunes (MP3 / AAC) tries to compress the music but still make it sound ‘reasonable’ etc etc.

So as you can see CD re-play is naturally ‘compromised’ to a certain degree, but it doesn’t make it inferior by default.

Personally I believe that what ‘makes or breaks’ a record is the mixing stage - even the best-recorded material can be really badly messed up 

French Rooster posted:

i know that some sites propose true dsd recordings :  these albums are originally recorded in dsd, directly. Some find that the sound is one of the best with that specific recording. But the music is vastly classical.

The vast majority of hd we buy, on qobuz or hdtracks, are analog recordings at their beginning or 16/44 and after they are changed in high rez.

Why modern music is not recorded directly in high rez?   not all music, like commercial tunes, but some albums for jazz or pop, rock....The sound would be even better....

I do think that classical music may benefit particularly from HR recording.  I don't think Andrew Everard has posted here for a while but I would be interested in his experience.  For myself, I have just finished listening to Fischer's new recording of Mahler 3 with the Budapest Symphony Orchestra on Channel Classics, downloaded from Qobuz in 24/96.  (It's also available in several 2 channel DSD versions.)  Mahler is hard to record well, partly because his music has huge fff climaxes as well as sections which are virtually chamber music.  This version is by some distance the best recorded Mahler I have heard (and I own many and have listened to many more recordings).  Indeed it is one of the very finest recordings of a large symphony orchestra and chorus I have encountered.  Somehow it manages to convey weight and force when appropriate and delicacy and intimacy in other sections.  Of course, the work of the recording engineers is paramount in this, but I cannot help feeling that hearing it in HR helps get their efforts across.  And no, I haven't bought the CD to compare it.  But this download has given me musical pleasure of a rare order and if HR is even a small part of that, I am very happy to continue buying HR recordings.

Roger

Peakman posted:

French Rooster posted:

i know that some sites propose true dsd recordings :  these albums are originally recorded in dsd, directly. Some find that the sound is one of the best with that specific recording. But the music is vastly classical.

The vast majority of hd we buy, on qobuz or hdtracks, are analog recordings at their beginning or 16/44 and after they are changed in high rez.

Why modern music is not recorded directly in high rez?   not all music, like commercial tunes, but some albums for jazz or pop, rock....The sound would be even better....

I do think that classical music may benefit particularly from HR recording.  I don't think Andrew Everard has posted here for a while but I would be interested in his experience.  For myself, I have just finished listening to Fischer's new recording of Mahler 3 with the Budapest Symphony Orchestra on Channel Classics, downloaded from Qobuz in 24/96.  (It's also available in several 2 channel DSD versions.)  Mahler is hard to record well, partly because his music has huge fff climaxes as well as sections which are virtually chamber music.  This version is by some distance the best recorded Mahler I have heard (and I own many and have listened to many more recordings).  Indeed it is one of the very finest recordings of a large symphony orchestra and chorus I have encountered.  Somehow it manages to convey weight and force when appropriate and delicacy and intimacy in other sections.  Of course, the work of the recording engineers is paramount in this, but I cannot help feeling that hearing it in HR helps get their efforts across.  And no, I haven't bought the CD to compare it.  But this download has given me musical pleasure of a rare order and if HR is even a small part of that, I am very happy to continue buying HR recordings.

Roger

you can just buy one track of 16/44 on qobuz and compare it to the 24/96 version you have.

If you like classical music, you have a site which proposes true native dsd files. NativeDsd Music, the name of the site.  Perhaps you will enjoy.

For myself i buy some hirez on qobuz and highrezaudio.com and am happy too to continue. Sometimes i am buying just one track or two, the ones i prefer.

Adam Zielinski posted:

French Rooster posted:

 

44.1 kHZ is a predominant sampling frequency. Higher res is also used - 96 kHZ being the next obvious step.

To prepare a CD - format musi a recording has to be converted to 16 bit - usually done after mastering, when a CD-specific file is being made.

I know you are qualified to answer, you are working in music and studio recording. You say that 24 bit is practically default setting and also that 44,1 khz is predominant sampling frequency.  So you say that most studio recordings in the world are recording albums in a more than 16/44 format ?  so minimum 24/ 44 format ?   i am surprised by that information because the industry is still producing cds and the digital format to buy albums is mp3 or 16/44 in the vast majority of commercial offerings.    I have also read that hdtracks or qobuz are selling in majority hirez that come from upsampled 16/44 files.

In essence yes - studios record / mix more information than is being ‘packed’ onto a CD.

24 bit depth is a norm (32 floating can also be used). It’s the bit rate that contains the musical information. So preparing a CD version of a digitial recording will necessarily take the bit rate down to 16.

Sampling frequencies used can be: 44.1 kHz, 48 kHz, 96 kHZ and higher. The higher the sampling rates the more demands this places on the entire recording / mixing equipment - all must be up to the same standard.

Mastering is a way of ‘preparing’ a mix for different  media - dynamic range is ‘squashed’ for a vinyl - otherwise a stylus would jump out of a groove; bit-depth is reducted for a CD to make the data physically fit onto a 700Mb storage space; mastering for iTunes (MP3 / AAC) tries to compress the music but still make it sound ‘reasonable’ etc etc.

So as you can see CD re-play is naturally ‘compromised’ to a certain degree, but it doesn’t make it inferior by default.

Personally I believe that what ‘makes or breaks’ a record is the mixing stage - even the best-recorded material can be really badly messed up 

ı wonder why a physical format for 24/96 did not succeed at all..... storage space is not a problem with silver discs as well... 

Emre posted:

ı wonder why a physical format for 24/96 did not succeed at all..... storage space is not a problem with silver discs as well... 

1) listen to the SQ of most modern music, by and large highly compressed,

2) consider that the mass market listens to MP3/4 files streamed via smart phones, these often "lifted" from YouTube rather than bought, then enjoyed through Beats HPs and the like,

3) bottom line - a huge bass presence is paramount to most modern listeners' musical enjoyment, and this requires little attention to meticulous mastering quality. In fact, mastering with extended dynamic range would likely play as a detriment, even sound foreign to most modern, mass market listeners,

4) it sucks to be an audiophile concerned with the SQ of modern music. Digital recording's potentials have been largely squandered due to marketing demands, and unfortunately the differently mastered vinyl options seem to have remained the superior SQ choice despite all its shortcomings and cumbersome physical requirements. Modern digital recordings set to vinyl are often the superior SQ choice for best quality replay, but not the news hi-res streamers want to hear.

What Joerand says plus commercial enterprise.

ie, this year we’ll sel 16 bit “masters” and next year we’ll release 24 bit “masters” etc

Peakman posted:

For myself, I have just finished listening to Fischer's new recording of Mahler 3 with the Budapest Symphony Orchestra on Channel Classics, downloaded from Qobuz in 24/96.  (It's also available in several 2 channel DSD versions.)  Mahler is hard to record well, partly because his music has huge fff climaxes as well as sections which are virtually chamber music.  This version is by some distance the best recorded Mahler I have heard (and I own many and have listened to many more recordings).  Indeed it is one of the very finest recordings of a large symphony orchestra and chorus I have encountered.

I strongly concur with those views. I have the 192KHZ version of that (downloaded direct from Channel Classics) and it is absolutely superb. I, too, take the view that it's often (although not invariably) in large scale symphonic works, well recorded, that the true benefits of high resolution music are most apparent.

If anyone else is interested Channel Classics also has a beautiful version of Beethoven's "Pastoral" symphony, recorded with the same artists, that also benefits from the same excellent recording technique and is available in various high resolution options.

Stephen (a hunter of excellent high resolution Mahler performances!)

Stephen_C posted

I strongly concur with those views. I have the 192KHZ version of that (downloaded direct from Channel Classics) and it is absolutely superb. I, too, take the view that it's often (although not invariably) in large scale symphonic works, well recorded, that the true benefits of high resolution music are most apparent.

If anyone else is interested Channel Classics also has a beautiful version of Beethoven's "Pastoral" symphony, recorded with the same artists, that also benefits from the same excellent recording technique and is available in various high resolution options.

Stephen (a hunter of excellent high resolution Mahler performances!)

I have the Tilson Thomas Mahler set in 24/96, obtained from Linn HD, (although I don't think they sell it anymore on their site), and it is superb. I also have a CD of the 6th symphony from the same cycle. There is no comparison. Even on my lowly S400's the 24/96 sounds light years "bigger" and has more oomph.

I also have a couple of Claire Martin albums in both hi-res and physical CD from Linn. The CD's are nice, either ripped or via a reasonable player, but the hi-res leave me with goosebumps.

And finally I have a CD of Phantom Limb and the hi-res version from Naim Records. And yes again no comparison. 

Could go on, (have many, many albums on CD, record, hi-res to compare), but most times the hi-res wins out. The better the mastering, the more complex the music, the more it works.