or Connect
Head-Fi.org › Forums › Equipment Forums › Sound Science › Why would 24 bit / 192 khz flac sound any better than 16 bit / 44.1 khz flac if both are lossless (if at all)?
New Posts  All Forums:Forum Nav:

# Why would 24 bit / 192 khz flac sound any better than 16 bit / 44.1 khz flac if both are lossless (if at all)? - Page 20

Perhaps simplest way to address this issue is just to ask exactly what part of the Nyquist–Shannon sampling theorem is it that one disagrees with and to please provide mathematical proof of the error in the original theorem. Everything else is just pointless back and forth with no chance of resolution.

So for example, if the minimum sampling rate, which Nyquist states needs to be a minimum of twice the highest frequency, i.e. 44.1kHz sample rate for frequencies up to 22kHz, needs to be higher than 2X then state the minimum required sample rate and provide mathematical proof.

Those who refuse to do blind listening tests don't often seem convinced by appeals to signal processing theory.

Quote:
Originally Posted by ralphp@optonline

Perhaps simplest way to address this issue is just to ask exactly what part of the Nyquist–Shannon sampling theorem is it that one disagrees with and to please provide mathematical proof of the error in the original theorem.

Seems like a good idea in theory but this approach still has problems:

1. Many don't understand what a theorem actually is, or how this particular one applies to digital audio. Many theories exist to explain some observed natural phenomena and may or may not be correct or entirely correct. This doesn't apply to digital audio though, because digital audio is a technology invented from the theory, not something which existed naturally and then a theory was invented to try to explain it. A fairly subtle but important difference which many audiophiles can't or don't want to grasp.

2. Physical audio equipment (ADCs and DACs) cannot perfectly apply the Nyquist-Shannon sampling theorem, there are real world engineering constraints which make it effectively impossible, a fact which the better educated audiophiles and audiophile equipment marketing departments are happy to exploit. In practise, real world engineering constraints still allow us to get very close to a perfect implementation of Nyquist-Shannon and any imperfections which do exist are (or should be) many times below audibility.

3. You assume, incorrectly, that Nyquist-Shannon automatically applies to all digital audio but this isn't always the case in the audiophile world. There was a trend at one point (I don't know if it still exists) by a few audiophile DAC manufacturers to market "filterless" DACs. This deliberately breaks the Nyquist-Shannon theorem, which specifically requires a band-limited signal (a signal which does not exceed half the sampling rate).

4. To some/many audiophiles their perception is sacrosanct, hearing is believing and they "trust" their hearing. Because this belief is sacrosanct and therefore unquestionable, anything which does question (or worse, disagrees with) it, MUST be incorrect. No matter how proven Nyquist-Shannon is, or how impossible to disprove, it just doesn't make any difference, it MUST somehow be wrong or it (and any other part of accepted science) MUST have missed something. Unquestionable belief in a perception is an irrational position to take and even more frustratingly, rational arguments are typically ineffective against irrational people. In fact, the more rational/proven/accepted the facts presented, the more irrational/ludicrous the responses have to become in order to preserve the sanctity of their belief, and this inevitably results in accusations of a complete lack of basic education or of actual insanity. As personal attacks are forbidden here, and this site largely exists to cater to audiophiles, the only option is to effectively ban science and rationality from any discussion. We end up with a hobby/passion entirely based on science but where the discussion of that science is effectively forbidden (except in this sub-forum), a thoroughly bizarre state of affairs!!

I'm usually all for letting sleeping dogs lie and allowing people to accept any old nonsense they want to support whatever beliefs they want. The problem is that many of these dogs aren't sleeping! Many/Most who come to this site come here for information and immediately become targets for indoctrination by these (very awake) dogs and the manufacturers/retailers who make a living from them (and their shills). All of which is incredibly frustrating for those of us who actually love audio. [/rant]! :)

G

maybe this topic should end, or be renamed for what it has turned into? as the flac question was really answered on first page.

I was tempted to just move posts to some cd vs highres topic, but it hard to find one that wasn't locked for turning into a troll fest or a boxing match.

Quote:
Originally Posted by gregorio

Seems like a good idea in theory but this approach still has problems:

1. Many don't understand what a theorem actually is, or how this particular one applies to digital audio. Many theories exist to explain some observed natural phenomena and may or may not be correct or entirely correct. This doesn't apply to digital audio though, because digital audio is a technology invented from the theory, not something which existed naturally and then a theory was invented to try to explain it. A fairly subtle but important difference which many audiophiles can't or don't want to grasp.

2. Physical audio equipment (ADCs and DACs) cannot perfectly apply the Nyquist-Shannon sampling theorem, there are real world engineering constraints which make it effectively impossible, a fact which the better educated audiophiles and audiophile equipment marketing departments are happy to exploit. In practise, real world engineering constraints still allow us to get very close to a perfect implementation of Nyquist-Shannon and any imperfections which do exist are (or should be) many times below audibility.

3. You assume, incorrectly, that Nyquist-Shannon automatically applies to all digital audio but this isn't always the case in the audiophile world. There was a trend at one point (I don't know if it still exists) by a few audiophile DAC manufacturers to market "filterless" DACs. This deliberately breaks the Nyquist-Shannon theorem, which specifically requires a band-limited signal (a signal which does not exceed half the sampling rate).

4. To some/many audiophiles their perception is sacrosanct, hearing is believing and they "trust" their hearing. Because this belief is sacrosanct and therefore unquestionable, anything which does question (or worse, disagrees with) it, MUST be incorrect. No matter how proven Nyquist-Shannon is, or how impossible to disprove, it just doesn't make any difference, it MUST somehow be wrong or it (and any other part of accepted science) MUST have missed something. Unquestionable belief in a perception is an irrational position to take and even more frustratingly, rational arguments are typically ineffective against irrational people. In fact, the more rational/proven/accepted the facts presented, the more irrational/ludicrous the responses have to become in order to preserve the sanctity of their belief, and this inevitably results in accusations of a complete lack of basic education or of actual insanity. As personal attacks are forbidden here, and this site largely exists to cater to audiophiles, the only option is to effectively ban science and rationality from any discussion. We end up with a hobby/passion entirely based on science but where the discussion of that science is effectively forbidden (except in this sub-forum), a thoroughly bizarre state of affairs!!

I'm usually all for letting sleeping dogs lie and allowing people to accept any old nonsense they want to support whatever beliefs they want. The problem is that many of these dogs aren't sleeping! Many/Most who come to this site come here for information and immediately become targets for indoctrination by these (very awake) dogs and the manufacturers/retailers who make a living from them (and their shills). All of which is incredibly frustrating for those of us who actually love audio. [/rant]! :)

G

So before it gets locked, as I still am quite confused by all this, is there a benefit to setting my PC to 24-bit/192 kHz as opposed to 16-bit/44.1 kHz?

Playing two different original files (192/24 vs 44.1/16, HiRes vs CD) or choosing between playing an original file at 44.1/16 or upsampling it in the computer  to 192/24? If it is the latter you should be upsampling to a multiple, in this case 176.4. If it should be done in the computer/software or in the DAC/firmware depends on which does a better job, or if it should be done at all... To add a bit to the confusion in delta-sigma demodulators there is also oversampling that take the signal to the MHz region. Google a bit about binary-weighted DACs (and R-2R) and delta-sigma DACs, and search in other forums.

I don't upsample original files above 88.2 or 96 kHz, in the computer or in the DAC (some DACs don't allow turning off upsampling or choosing the oversampling of the delta-sigma) and since the music player does a good job (upsamples before it starts transmitting and allocates enough resources so the pace through SPDIF is kept constant) I prefer to upsample ripped CDs in the computer (but when using iTunes not upsampling at all sounded better). But I guess it will depend a lot on the player/DAC combination, SPDIF/USB async., etc. Google, google, google... and do whatever sounds better to you (if you notice any difference at all).

Edited by axle_69 - 6/5/16 at 5:21am
Iv got question, if we could all hear the difference would a 24/96 track that is brick walled to something like DR5-6 does it make high resolution aspect of it pointless considering is cutting off the top and bottom?
Quote:
Originally Posted by 1c3d0g

So before it gets locked, as I still am quite confused by all this, is there a benefit to setting my PC to 24-bit/192 kHz as opposed to 16-bit/44.1 kHz?

No.

Nein.

Non.

Nee.

It's really not that confusing, is it? ;-) Just use 16 bit / 44.1 kHz, it's more than sufficient for our human ears.

Quote:
Originally Posted by Deftone

Iv got question, if we could all hear the difference would a 24/96 track that is brick walled to something like DR5-6 does it make high resolution aspect of it pointless considering is cutting off the top and bottom?

It would certainly make the 24-bits part pointless, but delivery at 24-bits is already pointless even for material with much higher DR ratings.

Quote:
Originally Posted by Deftone

Iv got question, if we could all hear the difference would a 24/96 track that is brick walled to something like DR5-6 does it make high resolution aspect of it pointless considering is cutting off the top and bottom?

Depends.  If we can all hear the difference maybe.  We can't of course.  Somewhere near 20 khz is the limit.  So called higher rez is merely higher bandwidth.  Much of it we can't hear.

The higher sample rate gives you higher bandwidth.  The extra bits don't do a better or finer job of encoding material below 20 khz.  That is the mind picture used to market higher rez.  But it is not so.

Hi,

I´m new here and sorry if the questions were asked in another post:

Is it better to download a track - 16 bit 44.1 - directly from let´s say Qobuz/HD Tracks etc. in FLAC / ALAC version OR is it better to buy a CD and then convert it with iTunes in lossless ALAC?

OR is it even better to download it in the original WAV version. Do you hear a difference between FLAC/ ALAC and WAV? Some different questions, I know. Just curious what you think/hear!

Thanks for any replies!

Quote:
Originally Posted by noplsestar

Hi,

I´m new here and sorry if the questions were asked in another post:

Is it better to download a track - 16 bit 44.1 - directly from let´s say Qobuz/HD Tracks etc. in FLAC / ALAC version OR is it better to buy a CD and then convert it with iTunes in lossless ALAC?

OR is it even better to download it in the original WAV version. Do you hear a difference between FLAC/ ALAC and WAV? Some different questions, I know. Just curious what you think/hear!

Thanks for any replies!

That is a hard question to answer. Sometimes Quobuz/HDtracks/any similar store might have different masters, sometimes they dont.

Case in point: The newest RHCP album from the dynamic range database

As you can see, they all appear to have identical DR. Now, I am sure it could be purely by coincidence, but I doubt it. That was likely one master, sampled 3 ways. They probably sound audibly identical. Take the cheapest of the 3 (which will probably be iTunes or CD, or go with 7Digital, Amazon, or Google, probably all identical)

However, Norah Jones Feels Like Home:

Those are different masters, without a doubt. And the DSD download fro Acoustic Sounds is also from that master.

But even the HD music stores aren't the same. Here is a quote from me in another post

Quote:
Michael Jackson Thriller provides an illuminating example. I currently own this, but I have the 30th Anniversary Edition, with an overall DR score of 09 (http://dr.loudness-war.info/album/view/110165) I uploaded that one, so I know its the score that my own copy got. The best example are the vinyls and the first pressing Japanese CD at 15. That one would be harder to find, presumably, but the DSD download from Acoustic Sounds is easily available to download (for \$25) just behind those masters with an overall score of 14. I assume 14 and 15 are pretty indistinguishable, and you in fact find that one track has more DR in the DSD version. I assume more is not always better (I am not an engineer so that is just a lay person's assumption) - and I pretty much consider anything on there in the green to be just fine.

So, go to Acoustic Sounds and get the DSD, but you could also go to HD Tracks and get the 24-bit/88.2KHz version which scores just behind the AS DSD, and is \$7 cheaper. You could also get a 24/96 version for the same price as the HDTracks version from Acoustic Sounds. But wait, it's the lowest scoring example (tied with some others). It's actually slightly worse than my MP3 download.

As a recap of that

• Original Japanese CD
• DR 15 (Price would vary, expensive on eBay from what I've seen)
• 25th Anniversary CD
• DR 9 (more tracks, though)
• Acoustic Sounds DSD
• DR 14 (\$25)
• HDtracks 24/88.2
• DR 13 (\$18)
• Acoustic Sounds 24/96
• DR 7 (\$18)

Quote:
Originally Posted by reginalb

That is a hard question to answer. Sometimes Quobuz/HDtracks/any similar store might have different masters, sometimes they dont.

Case in point: The newest RHCP album from the dynamic range database

As you can see, they all appear to have identical DR. Now, I am sure it could be purely by coincidence, but I doubt it. That was likely one master, sampled 3 ways. They probably sound audibly identical. Take the cheapest of the 3 (which will probably be iTunes or CD, or go with 7Digital, Amazon, or Google, probably all identical)

However, Norah Jones Feels Like Home:

Those are different masters, without a doubt. And the DSD download fro Acoustic Sounds is also from that master.

But even the HD music stores aren't the same. Here is a quote from me in another post

As a recap of that

• Original Japanese CD
• DR 15 (Price would vary, expensive on eBay from what I've seen)
• 25th Anniversary CD
• DR 9 (more tracks, though)
• Acoustic Sounds DSD
• DR 14 (\$25)
• HDtracks 24/88.2
• DR 13 (\$18)
• Acoustic Sounds 24/96
• DR 7 (\$18)

Thanks!

I already got the "thriller" in DSD from HD Tracks ... and now seeing that it has a very good DR score I´m kind of glad I made the right decision when buying it ;-)

What I didn´t get so far is the DR thingy. It means dynamic range, am I right? So when there´s more dynamic in the recording (master), it´s "better".

The strange thing is: I can hear that the new Red Hot Chili Peppers album is really compressed (like many rock albums), but despite this fact (and the bad DR data), I really like the mix of their new album (maybe because Nigel Godrich was the engineer, I don´t know). So does this mean, not to be tooo picky about the DR data?

And about my other questions: Do you know if it´s better to listen to FLAC, ALAC or WAVs? Or are they all the same (if I have a good DAC)?

As long as they are using the same master all lossless file are the same sound quality
New Posts  All Forums:Forum Nav:
Return Home
Back to Forum: Sound Science
• Why would 24 bit / 192 khz flac sound any better than 16 bit / 44.1 khz flac if both are lossless (if at all)?