Measuring objectively real differences in files with identical checksums

[Teresa]

I just want to state for the record that I expect the music files I've created to not change sound. Thus, I'm praying that no one finds any differences in files with identical checksums objectively or otherwise measured. 

[jabbr]

6 minutes ago, Teresa said:

I just want to state for the record that I expect the music files I've created to not change sound. Thus, I'm praying that no one finds any differences in files with identical checksums objectively or otherwise measured. 

Think of your music files like digital photographs. They might appear different in different screens, or when printed out differently, but all of the information is perfectly preserved. 

[alfe]

On 2/20/2020 at 3:00 PM, jabbr said:

Think of your music files like digital photographs. They might appear different in different screens, or when printed out differently, but all of the information is perfectly preserved. 

Or let's get back to the usual analogy of cars laps time, it's not because you do the same time that you brake and accelerate exactly at the same point.

[John Dyson]

On 2/20/2020 at 8:51 AM, Teresa said:

I just want to state for the record that I expect the music files I've created to not change sound. Thus, I'm praying that no one finds any differences in files with identical checksums objectively or otherwise measured. 

The only time that there is a difference in actual sound is when the audio data in the files is different and/or the reproducing system/environment is different.   There can be a difference in perception outside of those conditions above, but that is all in the brain, hearing system, etc.   That is, the human is a major part of the perception, but some people just do not admit that their perception is NOT 100% consistent...   There might be some people with very consistent and reliable hearing characteristics, but they must also be into meditation, controlling their blood pressure, totally stable moods, etc.

 

There are very extreme cases where file format differences with the same data, even assuming lossless compression, can make some technical difference in the sound, but hopefully that doesn't happen very often, because that system represents a design tradeoff.  (I am sure that there are cases where the RF/conducted interference patterns are different given different file formats.  Whether or not that interferences causes differences in the audio is a matter of the specifc equipment/configuration used.)

 

I regularly PERCEIVE differences on different presentations of the same material with the same equipment, but that perception differences  is ALL IN MY HEAD, and I readily admit/realize it.   I know that my body, like everyone elses, is not a perfect absolute measure of detailed audio characteristics.  This is from someone who used to have truly "perfect" (accurate)  pitch when younger.

 

John

[jabbr]

34 minutes ago, John Dyson said:

The only time that there is a difference in actual sound is when the audio data in the files is different and/or the reproducing system/environment is different.

What is the basis for this dogma? 
 

A CD has pits on a surface. The pits have different characteristics. 
 

What you are calling “audio data” is a mapping of one set of pits having certain shapes into a pile labeled “0” and the other pits into a pile labeled “1”

 

Audio files are convenient digital abstractions but the CD surface us analog.

 

You are assuming that two CDs each mapping into the same piles of 0 and 1 will be reproduced the same. What is the basis for your assumption? Do you believe that every CD audio playback system is perfect?

[kirkmc]

The person was talking about files, not CDs. CDs can certainly have problems in their plastic and pressing, though these would not affect the "sound" but rather introduce artifacts. 

[jabbr]

4 minutes ago, kirkmc said:

The person was talking about files, not CDs. CDs can certainly have problems in their plastic and pressing, though these would not affect the "sound" but rather introduce artifacts. 

Do what you can “artifacts” change the sound? Clicks, pops are sound right?

 

Files are an abstraction. Let’s get real.

 

*** as I’ve said many many times, my own system is not subject to these issues of analog differences between files. My own files are mirrored at block level and scrubbed which also helps with checksum errors over time from cosmic rays etc ...

[kirkmc]

Clicks and pops are disturbances, they don't affect the overall sound quality. I've rarely had CDs with more than one or two. 

 

Files are not an abstraction, especially if you buy files by download. If anything, CDs are an abstraction. Music is recorded either using analogue or digital techniques, and ends up digital at some point if it's going to be put on a CD. So the final existence of music before it is reproduced is a digital file. 

 

Unless you're just trolling... 

[jabbr]

7 minutes ago, kirkmc said:

Unless you're just trolling... 

no this is the thread title, I’m talking about measurable differences between files with identical checksums ... in order to do that you need to look at an actual real world files stored on a media ... obviously the numerical abstractions are the same. Surely you recognize that a file is an abstraction, no?
 

Let me restate this for you:

 

“Files with identical checksums should not sound different in a well engineered digital audio system“

[kirkmc]

So why mention CDs?

[jabbr]

4 minutes ago, kirkmc said:

So why mention CDs?

For @alfe 😂

[alfe]

2 hours ago, jabbr said:

For @alfe 😂

 

Thunder will also create disturbances if you use the cloud to store your files.

[John Dyson]

4 hours ago, jabbr said:

What is the basis for this dogma? 
 

A CD has pits on a surface. The pits have different characteristics. 
 

What you are calling “audio data” is a mapping of one set of pits having certain shapes into a pile labeled “0” and the other pits into a pile labeled “1”

 

Audio files are convenient digital abstractions but the CD surface us analog.

 

You are assuming that two CDs each mapping into the same piles of 0 and 1 will be reproduced the same. What is the basis for your assumption? Do you believe that every CD audio playback system is perfect?

If a file structure/storage medium has errors in the digital data, then all bets are off if the ECC doesnt' function.  PLL and stuff like that doesn't count unless it is a truly hamhanded design.

 

I am speaking data and data files on a storage medium, and a CD stores audio data similar in some ways as files.  IT IS dogma when you start talking about jitter propagation, unless you are talking about true data errors.

 

Anyone who actually knows how the errors and data propagation work -- jitter is all factored out of the mess, unless it causes errrors...  I know, as I do that kind of thing at the lowest levels -- no problem.

 

There are snake oil people who espouse religion, much like an illegitimate preacher -- avoid those kinds of people -- they just want your money or your trust for some strange emotional reasons.

 

PS:  please don't think that CD data is read anything like laserdisk data.  Small errors DO propagate on video laserdisks as the signal is pure FM modulated similar to a 1" video tape deck.   CDS are different beasts.

 

 

John

 

[John Dyson]

3 hours ago, jabbr said:

no this is the thread title, I’m talking about measurable differences between files with identical checksums ... in order to do that you need to look at an actual real world files stored on a media ... obviously the numerical abstractions are the same. Surely you recognize that a file is an abstraction, no?
 

Let me restate this for you:

 

“Files with identical checksums should not sound different in a well engineered digital audio system“

If you do two reads of the same CD, and the data is the same in the two reads, then the results will be the same.  There is no secret pathway for jitter noise to propagate through the data path, well unless you are talking about bad grounding, weird conducted noise problems, or in the worst case RF radiation.  (That is, really bad design -- jitter fixes are in the realm of better grounding noise control -- most circuits don't have neough sampling window issues to cause errors.)

 

If the data is read from a CD differently in two reads, and ECC doesnt' catch it, or the ECC scheme is hamhanded with poor buffering, then you can get error propagation, but not jitter.   This is either cr*ppy equipment or a poor quality CD that reads differently very often (or a dirty one at that.)  There can be sporadic errors anywhere, but I hope we aren't talking about that kind of thing as an important behavior.

 

 

John

 

[jabbr]

1 minute ago, John Dyson said:

If you do two reads of the same CD, and the data is the same in the two reads, then the results will be the same.  There is no secret pathway for jitter noise to propagate through the data path, well unless you are talking about bad grounding, weird conducted noise problems, or in the worst case RF radiation.  (That is, really bad design -- jitter fixes are in the realm of better grounding noise control -- most circuits don't have neough sampling window issues to cause errors.)

 

If the data is read from a CD differently in two reads, and ECC doesnt' catch it, or the ECC scheme is hamhanded with poor buffering, then you can get error propagation, but not jitter.   This is either cr*ppy equipment or a poor quality CD that reads differently very often (or a dirty one at that.)  There can be sporadic errors anywhere, but I hope we aren't talking about that kind of thing as an important behavior.

Jitter???

 

I'm not talking about any of this as an "important behavior"

3 hours ago, jabbr said:

“Files with identical checksums should not sound different in a well engineered digital audio system“

 

The title of this thread is "Measuring objectively real differences ..." and I am saying that give me two copies of a file stored on any media, and I can measure (given proper equipment) "objectively real" differences (at the analog level) between any two file copies.

 

Y'all are missing the point.

 

These differences do not have a practical effect on the output of a well engineered digital audio system.

[John Dyson]

On 2/19/2020 at 9:18 AM, jabbr said:

In all seriousness files stored on media could certainly, no, will have different levels of readout noise. That could be audible. However, the readout noise is not preserved with file copy. In the case of ZFS, the data is also mirrored, and read into RAM cache. In my case, from RAM cache, it is transmitted across network so whatever is going on the NAS is entirely electrically isolated from the streaming endpoint. My network is 10G fiber and assuming my switch meets the compliance testing standard, there is extremely little noise in the bits. Of course there is zero common mode noise  

 

Absolute best or worst stored in my magnetic media, as long as the bits are readable, the bits sent to my DAC do not reflect this noise with virtual certainty (we have no reason to invoke entanglement 😂) and this issue is moot for me.

'Readout' noise doesn't propagate through the data, but is the result of poor grounding or other bad EE design.  Don't mistake various side-noise sources as having anything to do with the 'data' per-se.   If you have system noise, then chase down the problem, but the file reading method or where it comes from has NOTHING to do with noise or jitter, unless it is a really attrocious design where the diskdrive radiates energy or causes so much conducted noise through a common ground that you can see it.  Note that a CPU can have 100-200 amp spikes, but the noise from them are also well controlled also.  A 10ma glitch (changing into a voltage, or coupled to the ether as radiated) shouldn't be a problem, unless it is a wiring, bad circuit board or other similar problem.

 

John

 

[jabbr]

18 minutes ago, John Dyson said:

I am speaking data and data files on a storage medium, and a CD stores audio data similar in some ways as files.  IT IS dogma when you start talking about jitter propagation, unless you are talking about true data errors.

 

Anyone who actually knows how the errors and data propagation work -- jitter is all factored out of the mess, unless it causes errrors...  I know, as I do that kind of thing at the lowest levels -- no problem.

 

 

Have I mentioned jitter propagation??? You must be reading something into what I've wrote. My own system, and I've said this many times, uses 10Gbe optical Ethernet, and the specified compliance testing explicitly measures for propagated jitter -- and ensures this doesn't exist -- this is called "stressed eye pattern" ...

[John Dyson]

3 minutes ago, jabbr said:

Jitter???

 

I'm not talking about any of this as an "important behavior"

 

The title of this thread is "Measuring objectively real differences ..." and I am saying that give me two copies of a file stored on any media, and I can measure (given proper equipment) "objectively real" differences (at the analog level) between any two file copies.

 

Y'all are missing the point.

 

These differences do not have a practical effect on the output of a well engineered digital audio system.

Reading the same file multiple times will result in the same data, unless the HW or medium is borked.  One time, a long time ago, CDs might have had more errors (mostly buffered and ECC corrected), or the CDs in the hands of your kids with peanut butter all over them, they will also have errors.

So, given the same medium, same equipment and same data, they all should be equal -- except for the human experience.  That is the BIG variable.

 

John

 

[jabbr]

3 minutes ago, John Dyson said:

'Readout' noise doesn't propagate through the data, but is the result of poor grounding or other bad EE design.  Don't mistake various side-noise sources as having anything to do with the 'data' per-se.   If you have system noise, then chase down the problem, but the file reading method or where it comes from has NOTHING to do with noise or jitter, unless it is a really attrocious design where the diskdrive radiates energy or causes so much conducted noise through a common ground that you can see it.  Note that a CPU can have 100-200 amp spikes, but the noise from them are also well controlled also.  A 10ma glitch (changing into a voltage, or coupled to the ether as radiated) shouldn't be a problem, unless it is a wiring, bad circuit board or other similar problem.

 

John

 

 

 

I can't tell if you are trying to agree or disagree with what I wrote...

[John Dyson]

6 minutes ago, jabbr said:

Jitter???

 

I'm not talking about any of this as an "important behavior"

 

The title of this thread is "Measuring objectively real differences ..." and I am saying that give me two copies of a file stored on any media, and I can measure (given proper equipment) "objectively real" differences (at the analog level) between any two file copies.

 

Y'all are missing the point.

 

These differences do not have a practical effect on the output of a well engineered digital audio system.

Even my simple, pretty much single ground audio test system -- no differences.

 

Where you get differences (like betwen flac and .wav files) is where noise sources propagate into unexpected places.   That doesn't happen on my equipment (wouldn't be correct to call my project environment a 'system'), but could happen on any assembled system.

 

This thing isn't about 'jitter' either, even though that had been one of the big snake oil sources in the past... Audio is really very good nowadays, and if 'audiophiles' really knew the op-amps or discrete circuits that their signals pass through, many would just give it up now, find a nice grave, and die 🙂  (Not intended literally. :-)).

John

 

[John Dyson]

Just now, jabbr said:

 

 

I can't tell if you are trying to agree or disagree with what I wrote...

Heee heee...   I usually don't agree or disagree...  (I try to avoid it), but really try to state the facts, and then take them or leave them.  I cannot convince anyone of anything, but only offer what I can contribute.

 

John

 

[John Dyson]

I aplogize if I misread what anyone wrote.  I feel really bad when claims become metaphysical about how noise can propagate.   I get a little heavy when try to make sure that things stay 'real' and based on the world -- not an alternative idea of reality.   Again, I TRULY apologize if I came across as rude -- I do care that people get off onto tangents, and I thought that I saw a 'tangent'.   No way to fix problems once they start metastsizing.

 

So, if we weren't going in the direction of multiple reads results in different 'sound' (not talking about perception here), then I owe you an apology.

 

John

 

[alfe]

http://www.ecma-international.org/publications/files/ECMA-ST/Ecma-130.pdf

 

[miguelito]

On 2/19/2020 at 7:54 AM, yamamoto2002 said:

About 20 years ago of issue of local computer magazine 月刊アスキー, there was an article about sound difference of identical DAW project file stored on different hard drive.

One drive is IDE HDD and another drive is SCSI HDD, and a music composer says SCSI HDD sounds better.
On the next issue, the cause of sound difference is found, it is caused by sound stutter due to slow drive transfer speed of IDE HDD

So this is not necessarily bogus if one of the disk types was, for example, putting back more electrical noise into the computer and the DAC was also on that same computer. Much harder to believe if this was not the case.

 

If, on the other hand, they were recorded to different drive types, then copied and played from the same type, then I would venture there should not be a difference.

[jabbr]

2 hours ago, John Dyson said:

 

I aplogize if I misread what anyone wrote.  I feel really bad when claims become metaphysical about how noise can propagate.   I get a little heavy when try to make sure that things stay 'real' and based on the world -- not an alternative idea of reality. 

 

No problem. I am trying to be very clear about what is real, and if I speculate, state that clearly. 
 

It’s abundantly evident that there is a serious rift between so-called objectivists and subjectivists here. I read people on both sides stating dogmas as facts without examination. 
 

The whole issue of files and file formats having “sound” is one of the fault lines. I’m trying to shed light not heat on these arguments. For example my file system may compress data on disc and decompress to memory — files really are a useful abstraction so we shouldn’t get caught up on these details and just use a well designed network. No noise on my NAS makes it to my DAC!