Forget storage comparison. Even the same storage can sound different with different file copying methods

[sandyk]

Whatever you lot are hearing, it's caused by something else.

 

At least, unlike some of your C.A. buddies, you are no longer insisting we are all delusional !

[mansr]

At least, unlike some of your C.A. buddies, you are no longer insisting we are all delusional !

 

I'm not discounting the possibility.

[plissken]

. This is a playback media discussion. Please cease and desist with off topic troll comments

 

I would normally agree but I didn't bring up MD5 hashes. Plus these are all played off the same medium so it's also about what utility was used to make the copy.

 

Please use the reporting tool for any flagrant posts.

[plissken]

You couldn't organise a fair, unbiased and scientifically meaningful test if your life depended on it !

 

I designed one where you get to perform it in your own rig fully sighted with no test administrator. It's a beautiful thing.

[jabbr]

You didn't like my link about buffer size, host drive bus, SD bus, host controller synchronising SDCLK and duty cycle distortion.

You prefer the volume control of your link:)

BTW do not extrapolate to HDD and SSD different beasts.

 

This is perhaps the crux of the argument. Two files could be bitewise identical but have very different PHYSical layouts which affect the playback circuitry.

 

The point is however that such differences do not survive file copy.

 

I think if we could agree on two things we would avoid some long repetitive flame wars:

 

1) bitwise identical files are not necessarily physically identical

 

2) physical differences between bitwise identical files are not preserved in file copy or transformation operations

[jabbr]

This is nonsense. I have hard drives with data that was written in the '90s, and they're still perfectly fine. Magnetic hard drives do not simply fade. High temperatures or strong external magnetic fields might of course cause data corruption, but that doesn't appear to be what this is about...

 

No common storage devices require any kind of maintenance to retain data when used normally.

 

Entirely agree with the essence of your argument, although blocks can become corrupted over time -- I scrub my online discs (ZFS) every few months even if they haven't had much activity. I also have some in offline storage

[mansr]

Entirely agree with the essence of your argument, although blocks can become corrupted over time -- I scrub my online discs (ZFS) every few months even if they haven't had much activity. I also have some in offline storage

 

Yes, periodic checking will catch a failing drive while most of the data is still intact. It won't prevent the drive going bad in the first place.

[davide256]

This is perhaps the crux of the argument. Two files could be bitewise identical but have very different PHYSical layouts which affect the playback circuitry.

 

The point is however that such differences do not survive file copy.

 

I think if we could agree on two things we would avoid some long repetitive flame wars:

 

1) bitwise identical files are not necessarily physically identical

 

2) physical differences between bitwise identical files are not preserved in file copy or transformation operations

 

its all about whats happening inside the hardware as firmware/software instruction at the CPU tries to execute data transfer from disk read to USB output. The "transmission" inside has complexity for timing integrity and most machines appear designed to "Yugo" tolerances, not "BMW".

[Paul R]

This is perhaps the crux of the argument. Two files could be bitewise identical but have very different PHYSical layouts which affect the playback circuitry.

 

The point is however that such differences do not survive file copy.

 

I think if we could agree on two things we would avoid some long repetitive flame wars:

 

1) bitwise identical files are not necessarily physically identical

 

2) physical differences between bitwise identical files are not preserved in file copy or transformation operations

 

Those are good points, so long as we all realize they are "in theory" and not hard proof. There is some compelling evidence for *and* against these ideas.

 

For example, a file is never played directly from storage media, it is *always* read into RAM and then copied numerous times before it ever hits the DAC D2A stage. By numerous, I mean anywhere from 11-12 times, to over a hundred in some cases. And not only has it been copied, it has transited various forms of transmission media.

 

If the storage media is, in any way, causing a difference, it should show up on playback of the same file on the same media as well as in bit for bit identical files on the same media. I personally have never seen this demonstrated under controlled conditions save with heroic measures to (intentionally) fragment a file on a disk.

 

Oh, I have seen electrical noise from disk drives, even SSDs, and other absolutely proven and reproducible issues, but not to the point where they cause audible differences on my equipment. Given that the equipment is all operating properly and within it's design parameters. My system is quite good enough to show up the differences between some speaker cables, any two DACs, and other less controversial equipment. At least to my ears. Should be good enough to show up this difference, if it is there.

[alfe]

Entirely agree with the essence of your argument, although blocks can become corrupted over time -- I scrub my online discs (ZFS) every few months even if they haven't had much activity. I also have some in offline storage

 

A good quality sputtering of the platters, a high coercivity of the material and you can keep your data safe for ages.

[davide256]

Those are good points, so long as we all realize they are "in theory" and not hard proof. There is some compelling evidence for *and* against these ideas.

 

For example, a file is never played directly from storage media, it is *always* read into RAM and then copied numerous times before it ever hits the DAC D2A stage. By numerous, I mean anywhere from 11-12 times, to over a hundred in some cases. And not only has it been copied, it has transited various forms of transmission media.

 

If the storage media is, in any way, causing a difference, it should show up on playback of the same file on the same media as well as in bit for bit identical files on the same media. I personally have never seen this demonstrated under controlled conditions save with heroic measures to (intentionally) fragment a file on a disk.

 

Oh, I have seen electrical noise from disk drives, even SSDs, and other absolutely proven and reproducible issues, but not to the point where they cause audible differences on my equipment. Given that the equipment is all operating properly and within it's design parameters. My system is quite good enough to show up the differences between some speaker cables, any two DACs, and other less controversial equipment. At least to my ears. Should be good enough to show up this difference, if it is there.

 

Paul... more than your ears in play here... the individual limitations of your system components also affect what you can hear. Hence why trade shows and dealer visits are important in understanding whats possible vs mistaking our home system limits for universal truths.

[Sal1950]

I still trust the maths behind FLAC as well as the fact that an encode/decode cycle demonstrably doesn't alter the sample values. Whatever you lot are hearing, it's caused by something else.

 

No, trust Cookie's ears.

 

Now here's the thing, I do trust her ears to produce great-sounding recordings. I just don't trust her explanations for why they sound great.

 

Ah, a voice of reason.

Cookie does some wonderful recordings, many of which are done to analog tape. And there are many other boutique labels out there also doing wonderful work that don't take/disagree with her chosen technical paths.

As to any other audible or technological claims coming from Cookie they are strictly her opinions. Maybe I missed something but my search revealed nothing in the way of a EE degree or anything approaching it.

 

At least, unlike some of your C.A. buddies, you are no longer insisting we are all delusional !

Not all, maybe just you?

[Paul R]

Paul... more than your ears in play here... the individual limitations of your system components also affect what you can hear. Hence why trade shows and dealer visits are important in understanding whats possible vs mistaking our home system limits for universal truths.

 

Yes of course systems are limited. But trade shows and dealer visits have their own, really serious limitations. Dealer rooms and dealer rooms at shows are usually not the best place to hear how a component will sound in your home. in fact, there may be more limits there than at your home.

 

Secondly, the idea that if you can't hear something in your system, the system must be inferior is very much open to dispute. The best possible system would sound the same all the time, and sound better to *everyone* than any other system. We of course will never get to that point, so everyone "tunes" their system to their listening environment and personal preferences.

 

One person may thing wall walloping bass is the signature of a perfect system, while another person might consider a system where the bass is perfectly blended and does not draw all that much attention to itself the perfect system. A musician might look for different things in a system than an accountant. (shrug)

 

The point is, assuming a system is "better" because it shows a flaw is erroneous thinking. It might be showing that flaw because it is not handling the audio correctly, or because the owner happens to like that particular flaw, whatever it is.

 

So to answer your unasked question, no - the idea that one cannot hear the difference between a FLAC file and a WAV file because one has too poor a system is an utterly bogus idea in most cases. No truth to it at all, and certainly no truth to it in my personal systems. Which are of course, tuned as neutral for my ears!

 

-Paul

[plissken]

This is perhaps the crux of the argument. Two files could be bitewise identical but have very different PHYSical layouts which affect the playback circuitry.

 

The point is however that such differences do not survive file copy.

 

I think if we could agree on two things we would avoid some long repetitive flame wars:

 

1) bitwise identical files are not necessarily physically identical

 

2) physical differences between bitwise identical files are not preserved in file copy or transformation operations

 

 

For pete's sake. Information is read off of medium into cache (RAM) set aside by the player application via an API call to the OS. And it takes multiple copies. No such thing as a zero copy stack.

 

Data is then streamed from RAM to the DAC. There is open air between the storage medium and the DAC. The DAC never actually sees the non-volatile storage medium with the data on it.

 

Bit wise files are indeed identical no matter if HDD, SDD, SD, mSD, USB Flash etc...

 

20 files on an SD card with the same MD5 hash are the same. I've got $1000 to some-ones $100 that I can send them an 8GB USB stick with two directories, 10 MD5 identical wavs each, and using UltraCopy on only one in each directory that they are going to fall flat on their face.

 

This is mass stupidity. Learn how data works folks.

[jabbr]

Those are good points, so long as we all realize they are "in theory" and not hard proof. There is some compelling evidence for *and* against these ideas.

 

1) I am not sure what you would need for hard proof, but it should be well known that the OS has no obligation to store two copies of a bitwise identical file in a physically identical fashion.

 

That said, the OS/filesystem may have the option of doing so, for example ZFS use of deduplication. But what I said is not theoretically true but demonstrably true.

 

2) At least the major filesystems, say for Windows, Mac, Linux do not mandate a particular physical manifestation of a file, in that blocks, for example, need not be contiguous etc. This is not just theoretically true, but demonstrably true.

 

For example, a file is never played directly from storage media, it is *always* read into RAM and then copied numerous times before it ever hits the DAC D2A stage. By numerous, I mean anywhere from 11-12 times, to over a hundred in some cases. And not only has it been copied, it has transited various forms of transmission media.

 

I never mentioned playback or sound or audible differences in my statements above. Perhaps you are extrapolating to a point you *think* I am trying to make? I am sure I've said the above myself.

 

If the storage media is, in any way, causing a difference, it should show up on playback of the same file on the same media as well as in bit for bit identical files on the same media. I personally have never seen this demonstrated under controlled conditions save with heroic measures to (intentionally) fragment a file on a disk.

 

Oh, I have seen electrical noise from disk drives, even SSDs, and other absolutely proven and reproducible issues, but not to the point where they cause audible differences on my equipment. Given that the equipment is all operating properly and within it's design parameters. My system is quite good enough to show up the differences between some speaker cables, any two DACs, and other less controversial equipment. At least to my ears. Should be good enough to show up this difference, if it is there.

 

Aside from the, what I hoped would be uncontroversial, points I made above, I expect that my own system would be immune to file format issues as well.

[jabbr]

For pete's sake. Information is read off of medium into cache (RAM) set aside by the player application via an API call to the OS. And it takes multiple copies. No such thing as a zero copy stack.

 

Data is then streamed from RAM to the DAC. There is open air between the storage medium and the DAC. The DAC never actually sees the non-volatile storage medium with the data on it.

 

Bit wise files are indeed identical no matter if HDD, SDD, SD, mSD, USB Flash etc...

 

20 files on an SD card with the same MD5 hash are the same. I've got $1000 to some-ones $100 that I can send them an 8GB USB stick with two directories, 10 MD5 identical wavs each, and using UltraCopy on only one in each directory that they are going to fall flat on their face.

 

This is mass stupidity. Learn how data works folks.

 

Listen Don Quixote, you are thrashing at the wrong windmill. Let me repeat:

 

1) bitwise identical files are not necessarily physically identical.

 

TRUE or FALSE? This statement has nothing to do with a DAC.

 

2) physical differences between bitwise identical files are not preserved in file copy or transmission operations

 

How does your proposal address my statement?

[Paul R]

1) I am not sure what you would need for hard proof, but it should be well known that the OS has no obligation to store two copies of a bitwise identical file in a physically identical fashion.

 

That said, the OS/filesystem may have the option of doing so, for example ZFS use of deduplication. But what I said is not theoretically true but demonstrably true.

 

The OS has the obligation to return the file bit for bit perfectly when it is read from the disk. That may involve any number of permutations, but when the driver releases the data it is exactly what was written to disk. (Ignoring failures of course.)

 

What is not conclusively proven is that storage media differences cause audible differences. I for one, tend to think that differences in storage media do - or at least *can* - cause an audible difference. However, as I said, there is compelling evidence for and against that idea.

 

2) At least the major filesystems, say for Windows, Mac, Linux do not mandate a particular physical manifestation of a file, in that blocks, for example, need not be contiguous etc. This is not just theoretically true, but demonstrably true.

 

Absolutely true. Only few layers further up in the storage hierarchy though, at the level of an application or OS function requesting data from the device, the data returned is identical, regardless of the storage media. Timing is always asynchronous as well with mass storage devices, so the OS and application requests are not time sensitive. (i.e. it does not matter if the data takes 3 milliseconds or 30 milliseconds to be returned, not at this level.)

 

I never mentioned playback or sound or audible differences in my statements above. Perhaps you are extrapolating to a point you *think* I am trying to make? I am sure I've said the above myself.

 

I never said you did, but that is an implicit assumption given the topic we are talking about, no? It is easily proven that the data retrieved from two bit identical files is in actual fact, identical. I am not sure what you are taking offense to here, but accept my apologies for any inadvertent offense. None was intended.

 

 

Aside from the, what I hoped would be uncontroversial, points I made above, I expect that my own system would be immune to file format issues as well.

 

(grin) I would expect it to be too. When I catch oddities like that in my system, I assume something is wrong and work at fixing it.

[esldude]

. This is a playback media discussion. Please cease and desist with off topic troll comments

 

Yes please use on topic trolling threads. No off topic trolling.

[plissken]

Listen Don Quixote, you are thrashing at the wrong windmill. Let me repeat:

 

1) bitwise identical files are not necessarily physically identical.

 

TRUE or FALSE? This statement has nothing to do with a DAC.

 

2) physical differences between bitwise identical files are not preserved in file copy or transmission operations

 

How does your proposal address my statement?

 

My proposal is Non-Realtime operating systems reads data off disk into buffer where it's yet again transferred to another buffer. It doesn't matter if it's fetching from SSD, HDD, Flash. Once it's buffered the prior doesn't matter. People keep thinking there are realtime analog-esque events. They are far from it.

 

The original storage mechanism matters not at this point? A COPY of what is on the hard drive is being played. Not the actual file on the HD itself.

 

Open up notepad on your computer and then save it to your desktop. Open it up again and type "I don't really understand how computers work" pull the power plug. Plug it back in and boot up. Let me know what is in your txt file.

 

Again I've put up $1K to someones $100. I'm still waiting for one of you deluded audiophiles that believe this shit to actually get a pair of brass one's strapped on.

[plissken]

In your dreams ! It fails to take into account the very things that many subjectivists report, even Cookie Marenco and Dr. Charles Zellig and Jay Clawson , about copying of .wav files causing degradation compared with the original CD and everything ends up sounding the same just as you wanted it to, due to all your additional copying/moving around of the files.

You assume that everyone making these reports is just as full of shit as you are !

It's a waste of time anybody discussing these things with such an arrogant nobody as yourself who has Zilch qualifications in the area of computer audio. No, I don't have any qualifications in this area either, but I get people who are qualified in this area to verify them ! .

 

Mrs Marenco, Zellig, Clawson are all welcome to my $$ same as any of the rest. Because if they believe what you are espousing, they are just as mistaken. Period.

 

They do potentially have more to lose than $100 though.

[sandyk]

Yes please use on topic trolling threads. No off topic trolling.

 

Dennis

Getting back to the subject of playback media, have you found the time to further examine or pass around those CD-Rs that

I sent you? I emailed you copies of a report about the findings with fairly similar comparison CD-Rs from me, where the results were consistent with the differences between different playback media as discussed in this thread.

 

Regards

Alex

[plissken]

Do you really think that qualified people are so stupid as to play your silly loaded games ?

You are just another faceless and anonymous person on the Internet.

Any further replies from you will be ignored. I have far better things to do with my time.

 

Are they qualified about how NRT OS's, copy operations, network framing and packets work? I don't propose to tell Mrs Morenco how to master audio anymore than she would be qualified to teach me parallel array structures in C#.

 

Anyways, I'm offering to to send a USB drive to Thailand along with the offer of a $1000 bucks.

[jabbr]

The OS has the obligation to return the file bit for bit perfectly when it is read from the disk. That may involve any number of permutations, but when the driver releases the data it is exactly what was written to disk. (Ignoring failures of course.)

 

What is not conclusively proven is that storage media differences cause audible differences. I for one, tend to think that differences in storage media do - or at least *can* - cause an audible difference.

 

We are in general agreement. In fact I see the whole storage issue and the FLAC vs WAV issue as two sides of the same coin.

 

There seem to be enough reports from people who have heard differences that I tend to believe that in certain circumstances ...

 

(grin) I would expect it to be too. When I catch oddities like that in my system, I assume something is wrong and work at fixing it.

 

And so for these possibilities as well as a number of other issues some of which I've had in a very noticeable fashion on my own system, I've adopted and advocate the "NAA" architecture where the audio PC is low power, discless and isolated by fiber optic Ethernet.

I never said you did, but that is an implicit assumption given the topic we are talking about, no? It is easily proven that the data retrieved from two bit identical files is in actual fact, identical. I am not sure what you are taking offense to here, but accept my apologies for any inadvertent offense. None was intended.

As you know things are always more complicated that they appear *but* I think the storage media issue is a red herring because it would go away if the proper buffering and isolation techniques were employed.

[esldude]

Dennis

Getting back to the subject of playback media, have you found the time to further examine or pass around those CD-Rs that

I sent you? I emailed you copies of a report about the findings with fairly similar comparison CD-Rs from me, where the results were consistent with the differences between different playback media as discussed in this thread.

 

Regards

Alex

 

No I have not Alex. It will happen soon finally. Soon being say within a month.

[alfe]

Questions:

 

Can jitter be stored in a CD?

Jitter generated by the ADC during recording is stored?

Is there any ADC in the ripping or reading process?

If the jitter can be stored in a CD, why it can't be in HDD or SSD?

Lucky we are, MFM, AFM, SEM exist