Curious Digital Audio Electronics Question

[Eric McMillan]

Hello all, perhaps some minds here can help.
I am 40 years experienced electronics technician, having specialised in the Audio electronics world in all flavours.
I have developed a mixture of materials that can be applied within cable connectors or as an enclosure fitted around electric/audio cables exactly in the manner of the now common ferrite clip-on filter assembly.
I have been exploring using a disused Android phone as a music player, ie iPod substitute.
The standard process is to transfer Wav audio files from PC hard drive to Phone flash memory via a USB cable, and then the phone can be disconnected from the PC and the phone headphone/line output signal connected to a standard hi-fi system for playback of the stored Wav file.
I have found that the playback is changeable according to the particular USB cable used to transfer the Wav file to the phone.

I have developed a protocol which is as follows.
On the PC hard drive I create a set of folders - Orig, 01, 03, 05, 07, 09 etc.
On the Phone I created a set of folders - 00, 02, 04, 06, 08, 10 etc.
I can choose any particular Wav file and copy this into the Orig folder.
Using a standard off the shelf USB cable I copy this Wav file to the Phone folder 00.
I then change out the standard USB cable and replace it with a custom USB cable incorporating the filter mixture contained in both connectors.

I then copy the Wav file from folder 00 to folder 01, and then copy this newly created file to folder 02 and repeat this process such that the folder number designates the number of transits that the Wav file has traveled the custom USB cable.
I can then disconnect the phone and use it as a player, and select whichever Wav file at will.
So the result is that Wav file 00 playback sound is different from the 'filtered' wav files contained in folders 02, 04, 06, 08, 10, and that the subjective difference increases according to the number of transits.
I can run AB comparison tests perfectly easily with A being the #00 file, and B say the #10 file.

So to cut to the chase, I am finding that particular Wav file copies alter the behaviour of the playback system, inducing a 'set' to the system, and according to the peak level encountered.
To clarify, I can play the 00 (A) file multiple times and after the first play, subsequent plays do not change.
If I then switch to say #10 (B) file, the system retains the #00 set/sound until the #10 file reaches higher/peak amplitude and the system then takes a new 'set' corresponding to the #10 file.
Replaying the #00 file restores the system 'set' back to the original condition.
So to summarise, A and B playback sounds are different, and further ABABAB playback sound is different to AABBAA playback sound.

All files are returned as being identical when compared in binary file checker softwares.
The subjective differences between the files is great enough to be perfectly reliably identified in ABX testing protocol.
The subjective differences include rather better centre image placement and stability, greater clarity of depth/distance information and overall distinctly better clarity and musicality.

So, it seems that there is some 'deeper' information contained in the Wav file that dictates/modifies the excess and/or 1/f noise behaviour of the playback/transducer system.
Some informal external soundcard loopback testing showed significant reduction (6dB+) in VLF 1/f noise.

Applying versions of the clip-on filter to all analogue systems such as phono and guitar amplifier significantly transforms such amplified sound.
Applying filters to the signal multicore cable and power feeds of large scale stadium PA system produces same subjective improvement, but includes significant (5-10 secs) delay.
Applying filters to transmit and/or receive antennas also alters subjective sound.

Anybody have ideas of what's going on here ?.

Eric.

 

[NTL2009]

Have you performed a bit by bit comparison of the original and later files? If there is no change in either the data or the meta data, then I have nothing to offer. I think there is data that describes the pre-emphasis and other settings for the codec, if this data got scrambled, that would explain a difference in sound and measured noise levels.

Perhaps I'm thinking of CD Redbook, not. WAV?

5-10 SECOND delays? Do you mean microseconds? Nanoseconds?

 

[Eric McMillan]

Thanks, I have determined this - "All files are returned as being identical when compared in binary file checker softwares.", and yes, I mean seconds.
What I'm talking about is deeper than binary data and standard radio theory, but thanks for trying to be helpful.

 

[NTL2009]

Interesting. I am an amateur musician, and I realize and believe that some of us can be sensitive to differences in sound sources that other people are not. The ear/brain is amazingly sensitive in some ways. And those sensitivities can be developed/trained to high levels. Similar to how a concert pianest plays at a muvh higher level than a beginner, critical listening is a skill that can be developed.

However, I also believe in science, so I believe that for us to detect a difference between two sound sources, a physical difference must actually exist. Otherwise, the difference exists only in our brain (the perception part of the equation of 'sound'). That is unlikely to pass a true, well designed ABX test at a level of significance. I have heard stories of trained listeners, either consciously or sub-consciously, tuning into the different relay click sounds of an ABX box, and identifying the sounds sources that way.

The differences we may hear may be difficult to measure. We may not know what to measure, or how to measure it. Maybe we don't even have equipment sensitive enough to measure it (I think our nose is more discriminating than some lab equipment in ppb detection of some compounds?). But I still stand by the belief that a physical difference must exist for us to hear it, whether we can measure it or not. It is just common sense.

If the files are bit-to-bit identical, and the player doesn't treat them differently in any way, then I see no rationale for there to be any detectable difference in sound. You can do anything you want to a file, if the end result is a bit-for-bit copy, and the player treats the file the same, then there is no physical difference in the sound. Period. Running the file through filters, unless a bit is flipped, means nothing.

I'm leaving that caveat in there about the player. Perhaps there is something that the player is doing? The file name triggers it to respond a certain way? The place on the media (could affect access time, could create higher current to engage the drive head, creating noise in the system, etc), fragmentation of the file? I doubt these would affect the physical sound produced in a properly designed player (sufficient buffer, proper noise immunity) on robust media, but I cannot rule it out w/o testing.

You say you measured a noise level difference. Then there must be an explanation. But I would not be focused on what happened to the file if it is bit identical. The answer lies elsewhere. Try copying the original to the location/name of your X gen 'filtered' copy. Do you still get a measured noise delta? That might lead you somewhere.

A 5-10 second delay in a sound system as a result of some ferrite-like filters on cables makes no sense. Something else is going on if that is what you hear.

 

[NTL2009]

Some more info:

http://soundfile.sapp.org/doc/WaveFormat/

That source describes the 'header' of a WAV file, and there are settings in there that, if altered, would affect the sound.

Perhaps one of these bits got flipped during a 'filtered' copy, and your bit comparison was only on the sound data? That would explain a measured difference.

 

[Eric McMillan]

Thanks, for your interesting and considered input.
The wav files are proven to be bit identical.
The loopback testing was recording soundcard self noise if I recall correctly, this will be revisited.
On running systems the system throughput is changeable at will by applying the filters, this is old hat to me.

I have run the experiment of all the return copies placed into the same folder and identified by 'date created' file information.
Changes in file metadata would appear to be the only difference, and presumably inconsequential.
Change of playback system behaviour correlating with data file history is taking things to another level for which I do not have explanation...yet.

 

[Baluncore]

I then change out the standard USB cable and replace it with a custom USB cable incorporating the filter mixture contained in both connectors.

I believe you are saying that the detail of the USB cable used to transfer the digital file to the phone/player makes a difference to the sound of the file when played.

I find that very hard to believe.

The wav files are proven to be bit identical.

You are saying that the files do not change their actual binary content due to the mode of USB transfer, storage directory or play history.

PF is a science forum. You tell us there is a difference, but have presented no evidence that there is a difference. You have not identified the make, model of the phone, nor the operating system, music player application or chipset. You have not tried phones from different eras with different software and digital audio modems.

There are a number of possibilities here.
1. There may be no real change in the sound. You could be imagining the difference.
2. You may be a magician, performing a trick to fool us and a potential marketplace.
3. The parameters used when replaying a file through a technically imperfect phone modem may be learned during the first play.

I think you should try different phone/players and investigate the phone-player audio specification. It may have a low resolution and so be juggling gain and peak amplitude to do the best it can. It may be something to do with the way the file is being played on the phone. The player may be holding a hidden file with parameters used for replay. There may be only one of those files per directory or file systems. Can you rename a directory. Does the phone forget the history?

 

[Eric McMillan]

Different phones, different player softwares, different playback audio systems, different listeners, same result.
Look I'm not trolling, I'm looking to meet a physicist with the depth of knowledge to drill down into what's happening here.
I did know a very old nuclear physicist who gave me a head start on my 'quantum' filter findings in the analogue domain, these observations in the digital domain are unexpected but subjectively duplicate the analogue observations.

Eric.

 

[Baluncore]

I did know a very old nuclear physicist who gave me a head start on my 'quantum' filter findings in the analogue domain, these observations in the digital domain are unexpected but subjectively duplicate the analogue observations.

Are there any references in the literature or links on the web ?
Do you have a patent on your filter technique ?

 

[Eric McMillan]

No and no, I haven't found useful/relevant information yet.

 

[Baluncore]

Are you really asserting that your filter on a digital USB cable does not change the bit pattern of the music files transferred, but can change what the file sounds like when it is played later ?

 

[Eric McMillan]

Yes, this indeed is what I find.
In this equation is the behavior of flash memory, this could be a factor in this experiment.
However this does not explain other observations like clipping the non magnetic filter onto a car radio antenna causing change of sound, yes, I'm dead serious.

 

[Baluncore]

This is really not a tractable subject for rational physics.

 

[Eric McMillan]

Is quantum physics rational ?.

 

[DrClaude]

I really, really, really doubt that this could pass a double-blinded test. Such magical cables abound in the audio world, and none has been proven to do anything valuable for analog signals, so never mind digital signals.

Instead of trying to figure out what is happening physically, you should try to concentrate of finding out was is happening psychologically for people to perceive identical sounds as different.

In any case, this topic is not suitable to be further discussed at PhysicsForums. Thread closed.

P.-S.:

Is quantum physics rational ?.

Extremely rational. For instance, I do all my QM calculations on a computer, which is a very logical machine