Sign in   |  Join   |  Help
Untitled Page

ARCHIVED FORUM -- March 2012 to February 2022
READ ONLY FORUM

This is the second Archived Forum which was active between 1st March 2012 and 23rd February 2022

 

24 bit / 192 kHz and why it’s silly

rated by 0 users
This post has 22 Replies | 1 Follower

seethroughyou
Top 100 Contributor
UK
Posts 999
OFFLINE
Bronze Member
seethroughyou Posted: Wed, Nov 22 2017 9:03 PM
One of the best articles on why 24/192 makes no sense. Very helpful for me. I’m going you stick to CDs and stop thinking ‘what if...’

https://xiph.org/~xiphmont/demo/neil-young.html

.

.

 

 

Present: BL90, Core, BL6000, CD7000, Beogram 7000, Essence Remote.

Past: BL1, BL2, BL8000, BS9000, BL5, BC2, BS5, BV5, BV4-50, Beosystem 3, BL3, DVD1, Beoremote 4, Moment.

.

Mr 10Percent
Top 150 Contributor
So West, Its Now East
Posts 591
OFFLINE
Bronze Member

......and no-doubt by now listening to your two fat ladies you will have discovered that a good deal of recordings are cr@p or mediocre at best no matter what the sampling quality. 

 

Mr 10Percent
Top 150 Contributor
So West, Its Now East
Posts 591
OFFLINE
Bronze Member

Mr 10Percent:

......and no-doubt by now listening to your two fat ladies you will have discovered that a good deal of recordings are cr@p or mediocre at best no matter what the sampling quality. 

 

 

Hummm....speed reading......"Better Masters" paragraph probably covers off on my statement above.

 

Millemissen
Top 10 Contributor
Flensborg, Denmark
Posts 14,680
OFFLINE
Gold Member

Exactly....and better gear!

The 24 bits are okay - room for processing the sound.

But the 192 kHz - which (B&O) speaker can handle that...

....and which human ear?

Start with better masters!

MM

There is a tv - and there is a BV

Jeff
Top 25 Contributor
USA
Posts 3,793
OFFLINE
Silver Member
Jeff replied on Wed, Nov 22 2017 11:48 PM

seethroughyou:
One of the best articles on why 24/192 makes no sense. Very helpful for me. I’m going you stick to CDs and stop thinking ‘what if...’

 

 

https://xiph.org/~xiphmont/demo/neil-young.html

 

 

.

An excellent article that very clearly and simply describes the pertinent technical details! Thanks for posting this, a very worthwhile read.

The biggest disadvantage of sticking with CD quality is the music makers can't sell you yet another copy of The White Album or Dark Side Of The Moon!

Jeff

I'm afraid I'm recovering from the BeoVirus. Sad

trackbeo
Top 150 Contributor
Posts 644
OFFLINE
Bronze Member
trackbeo replied on Thu, Nov 23 2017 5:17 AM

Richard Murison, writing for PS Audio -- yes, who happen to make higher-res DACs -- takes exception to certain assumptions in Monty's article.  See "Copper" magazine, issue 2, on page 14: http://www.psaudio.com/wp-content/uploads/2016/03/Copper-Issue-2.pdf

#2 of his quibbles is not relevant because he's talking about the recording chain, where 96KHz sampling is necessary to avoid brick-wall filtering in the analog domain prior to sampling.  (When recording, although the Nyquist frequency is already at or beyond the limit of human hearing, the brick-wall filter necessary to allow 20KHz sounds to pass on to the ADC and yet reject spurious 22KHz and higher-frequency signals (which may be present whether we can hear them or not) that would mess up 44.1KHz sampling, itself has audible effects.  So to allow less steep filters, and to start filtering at higher frequencies, you use a higher sampling frequency.  But that is only to ensure a decent recording; once you have it digital you can manipulate the signal and produce a 16/44.1 stream sufficient for playback.)

But what about his quibbles #1 & #3?  I'll have to read the other issues of "Copper"...

Millemissen
Top 10 Contributor
Flensborg, Denmark
Posts 14,680
OFFLINE
Gold Member

Jeff:

The biggest disadvantage of sticking with CD quality is the music makers can't sell you yet another copy of The White Album or Dark Side Of The Moon!

Jep - and companies like PS Audio can’t sell new DAC’s.

MM

There is a tv - and there is a BV

Millemissen
Top 10 Contributor
Flensborg, Denmark
Posts 14,680
OFFLINE
Gold Member

.....and fancy audiophile magazines would have nothing to write about.

MM

There is a tv - and there is a BV

Carolpa
Top 75 Contributor
Posts 1,700
OFFLINE
Silver Member
Carolpa replied on Thu, Nov 23 2017 9:38 AM

seethroughyou:
One of the best articles on why 24/192 makes no sense. Very helpful for me. I’m going you stick to CDs and stop thinking ‘what if...’ https://xiph.org/~xiphmont/demo/neil-young.html

I find it semi- scientifically and on points debatable!

So after listening to music in different resolutions I came to the conclusion: for me 16bit 44.1kHz is minimum quality.  

 

 

 

 

 

 

vikinger
Top 25 Contributor
Vestri Kirkjubyr, UK
Posts 5,422
OFFLINE
Gold Member
vikinger replied on Thu, Nov 23 2017 9:52 AM

Excellent article. Conclusion.... stick to simple stereo configurations and CD quality?

Graham

Jeff
Top 25 Contributor
USA
Posts 3,793
OFFLINE
Silver Member
Jeff replied on Thu, Nov 23 2017 2:02 PM

Carolpa:

seethroughyou:
One of the best articles on why 24/192 makes no sense. Very helpful for me. I’m going you stick to CDs and stop thinking ‘what if...’ https://xiph.org/~xiphmont/demo/neil-young.html

I find it semi- scientifically and on points debatable!

So after listening to music in different resolutions I came to the conclusion: for me 16bit 44.1kHz is minimum quality.  

Missed the whole thing about controlled listening tests apparently. That's the other excellent part of the article. Having done and participated in such tests, I have learned full well how easily the ear and mind are tricked by unrelated things, and perceive things that aren't really there.

There was an old saw that "It's very hard to find a black cat in a dark room, especially if he isn't there." The corollary is, "But if you really, really want him to be there, it's easy to convince yourself you hear him purr, even if he's not there."

Jeff

I'm afraid I'm recovering from the BeoVirus. Sad

Carolpa
Top 75 Contributor
Posts 1,700
OFFLINE
Silver Member
Carolpa replied on Thu, Nov 23 2017 3:37 PM

Jeff:
Missed the whole thing about controlled listening tests apparently. That's the other excellent part of the article. Having done and participated in such tests, I have learned full well how easily the ear and mind are tricked by unrelated things, and perceive things that aren't really there
Assumption!

I know and have experienced the effect!

still my point: it is semi- scientifically written and on points debatable stands.

For example the simple point of "reproducing" a sinus, if you sample it 192.000 times per second or 44.100 times per second and every sample is valued in the more accurate 24bit value (+/- 1.000.000) to a 16 bit value (+/- 65.000), is in the case of 24bit 192kHz more accurate, independent if it would finally lead to a better auditable sound or not.

 

.   

Jeff
Top 25 Contributor
USA
Posts 3,793
OFFLINE
Silver Member
Jeff replied on Thu, Nov 23 2017 4:21 PM

Carolpa:

Jeff:
Missed the whole thing about controlled listening tests apparently. That's the other excellent part of the article. Having done and participated in such tests, I have learned full well how easily the ear and mind are tricked by unrelated things, and perceive things that aren't really there
Assumption!

I know and have experienced the effect!

still my point: it is semi- scientifically written and on points debatable stands.

For example the simple point of "reproducing" a sinus, if you sample it 192.000 times per second or 44.100 times per second and every sample is valued in the more accurate 24bit value (+/- 1.000.000) to a 16 bit value (+/- 65.000), is in the case of 24bit 192kHz more accurate, independent if it would finally lead to a better auditable sound or not.

 

.   

Your description of sampling theory shows you don't really understand it, at all I fear, what you're saying is the same, false misunderstandings the high end audio community has been clinging to like glue regardless of the science and engineering. It's not more accurate, but your point about if it's audible or not is very well taken, it's possible to measure things that are in no way audible. The author very accurately points out the fallacies of the whole 16/44 isn't adequate for reproduction. 24 bits, as pointed out, has numerous benefits in recording, the penalty for exceeding 0 dB with digital is much more severe than just a slight roll off of high frequencies and increase in distortion. Exceeding 0 dB or the most significant bit of a digital recording, clips the signal, creating a lot of high frequency hash as it rounds off peaks to square waves.

Not trying to be too prying, but have you ever participated in a properly run, level matched, double blind test? I can say, it's an eye opening and humbling experience, to realize that all the things you've thought you could hear and all the audiophile approved wisdom are just flat not true. I'm quite happy I did these, because they eventually led me to B&O. I had thought the BS9000 was beautiful, but was believing the usual nonsense about DACs and thought B&O could not make a proper high end CD player. Upon realizing that my testing of various players and outboard DACs produced no discernible difference, I thought, well, why not buy the Beosound 9000, it's absolutely stunning and will sound just like every other properly designed CD player, and after that I stopped obsessing over things and enjoyed the looks and listening to the music. It does tend to remove obsession with equipment.

Jeff

I'm afraid I'm recovering from the BeoVirus. Sad

Dillen
Top 10 Contributor
Copenhagen / Denmark
Posts 13,191
OFFLINE
Founder
Moderator
Dillen replied on Thu, Nov 23 2017 4:53 PM

In my opinion, it's not all just a matter of bits and numbers.
It's very much a matter of the way we humans experience and receive the results.
The (first B&O) CD player Beogram CDX is generally rated as one of the best sounding CD players ever
produced, and everytime I have one on the bench I am amazed by its sound quality and its warmth.
Put in short, it's pleasant to listen to.
It simply sounds good.
- And it's a 14-bit machine.

All theory and math screams no, but the vast majority of humans like it and can listen to it for
hours without wearing their ears out.

Are we striving for nice numbers on paper or pleasant sound?

Martin

Jeff
Top 25 Contributor
USA
Posts 3,793
OFFLINE
Silver Member
Jeff replied on Thu, Nov 23 2017 5:16 PM

Martin, you bring up a good point. You can show, with controlled listening tests, that no difference exists. But where a difference does exist, the test will not tell you which is preferred, that's highly dependent not only on how the ear works but how the individual perceives sound and what they like.

A prime example was and is the continuing war over LP vs. CD. CD is accurate, period. Accuracy has a definition when applied to a transfer mechanism that's used to deliver sound from studio to end user...is the transfer function transparent. Accurate is described as, after you apply the transfer function of the medium, is the end result the same as the input. Many times LPs may sound better to many people, but the CD is more accurate to the original master, it's just that the master source doesn't sound that good. LP has many inaccuracies, it introduces phase, frequency response, and channel separation errors, but to the human ear many of these errors are euphonic, they sound good even if they aren't accurate.

Bob Carver made CD players for a while with what he called a "Digital Time Lens" which was a fairly simple circuit that introduced some of these euphonic colorations into the analog output signal of the player. It was pretty effective at mimicking LP sound, and was especiallly helpful on poorly mastered early CD releases.

But before you can do preference testing you need to test to see if any real, repeatable audible difference exists at all, otherwise you're measuring delusions.

I have the Magnavox/Philips version of that player, need to get it serviced, but my memory of it was it sounded remarkably good despite being "only" 14 bits.

Jeff

I'm afraid I'm recovering from the BeoVirus. Sad

Carolpa
Top 75 Contributor
Posts 1,700
OFFLINE
Silver Member
Carolpa replied on Thu, Nov 23 2017 5:22 PM

Jeff:
you don't really understand it, at all I fear,
yeah, yeah,

this made me stop reading it further.......... after first an assumption....... you're completely right no discussion to it ........

seethroughyou
Top 100 Contributor
UK
Posts 999
OFFLINE
Bronze Member
Mr 10Percent:

......and no-doubt by now listening to your two fat ladies you will have discovered that a good deal of recordings are cr@p or mediocre at best no matter what the sampling quality.

After listening to my two fat ladies, and to some of the most regarded audiophile speakers out there in last 3 years on my research field trips to dealers around the country, I can confidently say that the BL90 are some of the most revealing and impressive speakers out there. I’ve also had a chance to compare 24/192 with 16/44.1 recordings (genuine ‘high res’ Linn and Chandos recordings), I can hand on heart say I can’t hear a difference. I’ve uncovered probably 20-40% more detail on 16/44.1 like I’ve never heard before. It’s at this point on an audiophile forum I’d be accused or not having revealing speakers/gear or...being deaf or lacking golden ears.

.

 

 

Present: BL90, Core, BL6000, CD7000, Beogram 7000, Essence Remote.

Past: BL1, BL2, BL8000, BS9000, BL5, BC2, BS5, BV5, BV4-50, Beosystem 3, BL3, DVD1, Beoremote 4, Moment.

.

Jeff
Top 25 Contributor
USA
Posts 3,793
OFFLINE
Silver Member
Jeff replied on Thu, Nov 23 2017 6:28 PM

Carolpa:

Jeff:
you don't really understand it, at all I fear,
yeah, yeah,

this made me stop reading it further.......... after first an assumption....... you're completely right no discussion to it ........

Perhaps you'd like to enlighten me on your listening tests, and what controls you used to avoid the obvious issues? II'd honestly like to hear how you performed them, I might learn a trick or two. It's not my fault that your post is completely technically in error, but apparently you don't like or are unwilling to read and absorb why. But your attitude is extremely common I fear. If you read my post, instead of having an emotional reaction and skipping it, and if you've read the original article linked to, the fact you claim essentially that 2 plus 2 is not equal to 4, that you don't understand sampling theory/Nyquist/etc., well, there's nothing I can do to change that. As the saying goes, you can lead a horse to water, but if you can get him to float on his back you've really got something! I have done a lot of work related to sampling theory in areas a lot more difficult than audio, I have no reason to doubt the mathematics is accurate, audiophile company hand waving not withstanding.

Jeff

I'm afraid I'm recovering from the BeoVirus. Sad

Puncher
Top 10 Contributor
Durham
Posts 11,729
OFFLINE
Bronze Member
Puncher replied on Thu, Nov 23 2017 7:16 PM

Simple enough really, 24 bit encoding has less quantisation noise than 16 bit giving the quoted SNR figures, its real and measurable. Can you hear beyond the 96db - I don't follow this type of thing with any real vigour so if someone can point me to a properly conducted double blind test that proves positively that it is audible I'd be very interested to read it.

Sampling at high rates has distinct advantages when it comes to anti-aliasing filter selection and design (many standard CD players use oversampling D/A converters for this very reason). Sampling the source at high rates can either again allow for a simplified input filter, or allow inaudible information above 20KHz to be captured. Again I'd be interested if someone could point me to a double blind test where this ultrsonic detail was reliably detected during a typical musical performance.

What sampling at high rates categorically does not do is capture a correctly bandwidth limited (I.e. no content above the Nyquist frequency) any more accurately than any other sampling frequency which meets the same requirement eg. A 20KHz bandlimited signal is captured by a 44.1KHz sampler and a 192KHz sampler will capture exactly zero additonal information/detail/accuracy. It may seem untrue looking a typical "stepped sine" diagram but true it is.

Ban boring signatures!

Puncher
Top 10 Contributor
Durham
Posts 11,729
OFFLINE
Bronze Member
Puncher replied on Sun, Nov 26 2017 3:51 PM

Here's the thing - I've just been listening to "Reflection" by Brian Eno streamed from my own ripped music collection stored in the cloud on Google Play Music at 320Kbps MP3 via a Chromecast Audio into my 80's BM5000, with a G&T in one hand (Martin Miller and FeverTree) and my eyes closed while I drifted to God only knows where. (It's over now coz  Mrs. Punch has returned from the shops with tales of attempted murder in the name of Santa and I've had to turn off the music and feign interest).

My (very long winded) point is that, before she returned, during that sublime, almost out of body, 54 minute experience - never did I ever think, "Gee - I wish this was streaming in 24bit, 192KHz format"!

 

p.s. - I'm in that "I always feel more enlightened after listening to Eno", and soothed beyond even the powers of Martin Miller!

Ban boring signatures!

Jeff
Top 25 Contributor
USA
Posts 3,793
OFFLINE
Silver Member
Jeff replied on Mon, Nov 27 2017 5:17 PM

Sir, I must insist, insist I say, that you turn in your audiophile credentials as you, sir, are no audiophile. Listening to the music instead of the recording and equipment!!! Shocking, shocking I say. Harumpf! SurpriseBig Smile

Jeff

I'm afraid I'm recovering from the BeoVirus. Sad

elephant
Top 10 Contributor
AU
Posts 8,219
OFFLINE
Founder
elephant replied on Tue, Nov 28 2017 5:00 AM
Puncher:

(It's over now coz  Mrs. Punch has returned from the shops with tales of attempted murder in the name of Santa and I've had to turn off the music and feign interest)

Ain’t that the sad truth

BeoNut since '75

trackbeo
Top 150 Contributor
Posts 644
OFFLINE
Bronze Member
trackbeo replied on Wed, Nov 29 2017 7:35 AM

Jeff:
Listening to the music instead of the recording and equipment!!! Shocking, shocking I say. 
What makes you think that Eno's "Reflection" is music? Laughing  (Seriously though, it's an app, isn't it, that just happens to generate sounds!  And then Eno recorded an hour or so to make a CD, right?)

Sure, I love "Music for Airports", but it's not certain that one would even notice if it were sampled to 160K MP3 -- assuming one had never heard it any other way just before that!  We have no idea (well, maybe the liner notes say) what made those sounds in the first place...  I didn't give Mr. Punch a hard time but if he always listens to Eno at 320K MP3, there's no reason to believe that whatever he is hearing isn't just as good, or even as "original" as any other encoding.  He would never miss 24/192 because it would just be a different sound, no better and no worse.  If I always make my jinto's with Sainsbury's instead, I'll be just as happy (or loopy!), but only if if I've never tasted Mr. Punch's punch in close proximity...  Maybe critical listening to a human voice, or an instrument you know well, but not Eno...  (OK, maybe not *any* human voice -- have you ever listened to Eno sing on "801 Live"?!?Erm..)

(Hey, now there's a genre to explore: Music that is intentionally manipulated in advance, so that a commonly-used MP3 encoder generates the same bits when decoded -- effectively making MP3 a lossless format!  I wonder if it's even theoretically possible...)

Page 1 of 1 (23 items) | RSS