[markC]

This is easy to prove. Record a digital stream from a CD player or whatever to a hard drive five times in a row, then bit-compare the files. If the files are all the same - and they surely will be - then there was no corruption or mis-reading of ones and zeros.


What if the same errors occurred each of the 5 times?

[andyr]

Try to open the mind of the Taliban so they start to respect women and other religions. Now, that would be a real challenge

LOL - good one Kyrill!!    I think it would be easier to stuff a camel through the eye of a needle than achieve that one, mate! 

I mean, the entry requirement for joining up surely is a "closed mind" and a mind-set which is at least 500 years old!     IE. if you respect women and other religions ... by definition you can't be Taliban! 

Regards from downunder.

Andy

[turkey]

I mean, the entry requirement for joining up surely is a "closed mind" and a mind-set which is at least 500 years old!

That's true of many religions.

[andyr]

That's true of many religions.

Except Buddhism, I would suggest. 

Regards,

Andy

[kyrill]

This is easy to prove. Record a digital stream from a CD player or whatever to a hard drive five times in a row, then bit-compare the files. If the files are all the same - and they surely will be - then there was no corruption or mis-reading of ones and zeros.


What if the same errors occurred each of the 5 times?

the storage of digital data is practically perfect but in music reproduction the streaming of digital data has a host of different problems listening wise Ears/brains are much more "sharp" and pin point precise to what constitutes a believable musical copy of a performance than measurements will ever show. True is too that a host of psychological influences play a role. Just like that for every good scientific researcher in the social domain there are many many more "naive" people perfectly inadequate to be a neutral inquirer

[turkey]

I was simply disagreeing with your statement that data transfer in a computer is flawless.

Maybe that's why both Windows and Mac OS crashes sometimes?

It's possible, although software bugs are the more likely cause. Device drivers are typically a point of failure. (I have had Windows crash too many times to count. On the other hand, I have never had Mac OS X do the same.)

[Ethan Winer]

What if the same errors occurred each of the 5 times?

The very nature of jitter and edge detection errors is randomness. So five tries would never have the same errors at the same places. Even two tries is enough to prove the point.

--Ethan

[JoshK]

Let's stay off the religion topic please, this is the lab.

[Browntrout]

Ethan I think you have an incorrect understanding of digital systems, from what you write that certainly appears to be the case. The storage of digital data is not at all perfect, errors occur all the time, it is the re-reading and statistical analyisis of the readings that provides the certainty of retrieval and this can only occur when the frequency of errors is within the working limit of the system.
  Here is an interesting article you could read...http://www.soundonsound.com/sos/jul98/articles/digitalbasics3.html

As for how different error correction technologies/mathematical formula sound depends ultimately upon how accurate the original read is from the disc/hard drive/memory chip this can be dependant upon many factors including ambient light, background radiation, the stability of the voltage being supplied to the memory chip (and noise in that supply) and applicable to both cd's and hard drives, the rotational characteristic.
  Digital data retrieval is flawed, however the mathematical logic that is applied to this data (Boolean algebra) and the use of truth tables allows for multiple flawed readings to result in one accurate one assuming that the data is readable in the first place.

[Ethan Winer]

The storage of digital data is not at all perfect, errors occur all the time, it is the re-reading and statistical analyisis of the readings that provides the certainty of retrieval ... Here is an interesting article

That article concerns data redundancy on CD media, where we've been discussing edge detection and jitter which is an entirely different issue. My comment that you can prove valid data transfer by recording the same stream a few times and comparing the files is valid. If you believe that bit-identical files can sound different, I respectfully submit it is not I who does not understand digital audio.

I'm still hoping Scotty will let us know if he's ever actually designed electronic gear, and I'll now extend that question to you. Audiophile type magazines are probably not the best source of technical information about audio science.

--Ethan

[Browntrout]

     'If you believe that bit-identical files can sound different, I respectfully submit it is not I who does not understand digital audio. '

What is a bit-identical file? Do you mean two seperate stored files that read the same or do you mean two seperate files that are actually identical down to the last bit of data?
  You see from your last post you show your missunderstanding again. Two files that read the same will not actually be the same, the different files result in the same outcome, they appear identical but they are not, they have errors (and differences) but the error correction makes them appear to be perfectly identical.
 

[Ethan Winer]

Do you mean two seperate stored files that read the same or do you mean two seperate files that are actually identical down to the last bit of data?

That's the same thing. If you run a file compare utility on two files on your hard drive and no difference is found, then the files will also sound the same by definition.

You see from your last post you show your missunderstanding again.

I'm trying hard to be nice, and I haven't accused you of not understanding digital audio. It would be great if you'd extend the same courtesy to me and stick to the topic at hand. Thanks.

Two files that read the same will not actually be the same, the different files result in the same outcome, they appear identical but they are not, they have errors (and differences) but the error correction makes them appear to be perfectly identical.

I have no idea what you're trying to say here, but I'm sure you'll clarify.

In the mean time, maybe you can cite some references from the literature explaining how two files on a hard drive can be bit-identical yet not sound the same every time they are played.

--Ethan

[NewBuyer]

If the speaker cable picks it up it will only be a problem if it can find its way back into the amplifiers feedback loop. Most amplifiers have a Zobel network or filter network which bandwidth limits what can come back into the amp from the speaker cables. I would not have looked for RF at the amplifiers terminals it should not be present unless the amp operates at radio frequencies or it is oscillating.
What might have been measurable is increased intermodulation distortion at the output of the amp, a very small increase in IM distortion is very audible and very objectionable. 
Scotty

Hi Scotty,

I was actually considering trying these "High Definition Links" from Walker Audio, to see what if anything they did.  Note they are claimed to address and defeat EMI/RFI etc riding on the speaker cables, with (as they and their users say) very audible results.  (Please click on picture) 


Based on what you said above, do you believe these items are probably totally bogus then?  (I personally have no idea - just looking for trustworthy advice, if the claimed premise of their product is genuinely scientifically nonsensical...) 

[*Scotty*]

NewBuyer,we're a long ways off the original topic, but here goes.The "High Definition Links" appear to be stacked plastic film capacitors,which depending on their value will reduce RFI going back into the amp and also reduce it's high frequency response which may impact how the highs sound. The speaker could care less whether RF goes into it or not as it will not respond to micro-volt levels of RF. If the amplifier is well designed with good stability margins and a has proper network on it's output these filters shouldn't be required. Here is an image of a stacked non-inductive cap without the spade-lugs installed. If they are a filter network as claimed there will also be a resistor in there somewhere as well in series with the cap. I expect that most of the changes in sound that users have heard when using these devices can be attributed to reduced HF response from the power amplifier. When I played with this idea 5years ago I used a .47mfd 200V polypropylene/tin cap in series with 10 ohm 1watt metal oxide resistor. These two parts were soldered together and the parts were inserted into a dual banana plug from Pomona. You pop the dual banana plug into the back of the speaker terminals and listen. You can easily do a single blind test to see whether you like what it does or not. You can clean up sibilance problems with them to a certain degree but you may pay a price in high frequency extension. I don't know if I had more than $30 in both filters. I had the caps and resistors laying around in my parts drawers for a few years so I had no out of pocket expense. I tried them because I had a very fast amplifier that wanted to oscillate or ring on high frequency transients when used in combination with a certain speaker cable. I decided to not use this type of cable design with this amplifier rather than use the extra filter at the speaker terminals. It was better to side step the problem and use a speaker cable design that didn't upset the amplifier rather than degrade the high frequency response of the amp with the extra filter.

I wouldn't pay what Walker asks for his but they're not bogus. In most cases I would call this another form of tone control.
Scotty

[NewBuyer]

Extremely interesting and thorough info Scotty - thanks! 

I think you're saying that if an amp has an IMD increase from incoming mains EMI/RFI, this could be entering perhaps through the power-cable/IEC, or through the speaker wires.  The former (mains and power cable) could be a problem if the amp somehow has a less-than-sufficient PSU (i.e. with inadequate mains noise-rejection characteristics, etc).  The latter (speaker wires) could be a problem potentially with certain less-compatible amp/speaker-wire combos, or in the unlikely case of an amp with no Zobel or filter network.

I'm guessing that signal interconnects picking up some RFI might be another issue as well(?)

So I suppose that if a certain fancy power cord can substantially reduce radiated or conducted RFI (bidirectionally) that could otherwise be picked up by other system parts or components, then this could potentially be a benefit (in theory).

But whether such an alleged power-cord benefit actually audibly matters, and thus has ever been a benefit in practice, is what some have perhaps been questioning.  The previously described null test seems like it would be a great field-test of this... especially if somebody ever can show by this method, that a fancy power cord in fact really helps in their specific case.

[Browntrout]

Ethan I'm not being nasty saying you don't understand, it's not accusatory. There are lots of things I don't understand and I will admit that and some I think I understand and don't I'm sure.
 

  

[*Scotty*]

For my self, I'm not very concerned about RF getting into the feedback loop of my amplifier. RFI severity is a function of signal strength,the mains are an enormous antenna system and as such are of primary concern and it would make sense to low pass the the power at the service before the RF got into your house wiring. Conversely the speaker cables are a very small antenna and the potential for a RFI problem from this direction is much smaller. Resorting to a distortion analyzer and spectrum analysis of the distortion residual is really not necessary
unless you feel that you have to have concrete evidence to confirm what you have observed through listening to the power filter or power-cord. If I didn't make it clear in my previous post the overshoot/ringing problem I encountered with the amp was directly due to the RLC parameters of the cable and were not related to any RFI problem. In environments that have serious levels of RF energy almost any potential entrance point for it has to be dealt with, frequently through shielding everything. In some severely RF contaminated areas the amplifiers bandwidth has to be reduced to cope with the problem. This is a sub-optimal solution because if you have to resort to too much of a reduction in bandwidth you have basically dumbed the system down to the point that you have defeated the purpose of having a High Fidelity system. Bandwidth reduction in these cases is a balancing act between trade-offs.
Scotty

[Ethan Winer]

Ethan I'm not being nasty saying you don't understand, it's not accusatory. There are lots of things I don't understand and I will admit that and some I think I understand and don't I'm sure.

I'm sure too.

Seriously, it's always best to stick to commenting on the issues, rather than address what someone else does or doesn't know. If you think I'm wrong on how digital edge detection and jitter etc affect what's audible, then explain how I'm wrong and why.

--Ethan

[Browntrout]

I think your understanding of how a digital system works is incorrect based upon this:

"Quote
There is a transition period between the 1 and 0 where voltage is in between the 2 values.  At that point when it's read, there is some error allowed where the firmware will simply round.

I don't think this is accurate either. The receiving devices know what sort of voltage levels to expect, and they don't read in-between values. This is easy to prove. Record a digital stream from a CD player or whatever to a hard drive five times in a row, then bit-compare the files. If the files are all the same - and they surely will be - then there was no corruption or mis-reading of ones and zeros.

--Ethan

Firstly they do read in between values, they read all values, the basis of all digital systems is that they quantify readings into groups (usually two groups).
   Recording from a cd player to a hard drive then using software to show they're identical shows nothing other than the device is working properly. It tells us nothing of how many errors in reading there were or how many times interpolation occured.
  To say there was no corruption or miss-reading of ones and zeros is showing your basic missunderstanding. There were errors (I promise) but the algebra used in the correction removed them (identified them as anomalies).
  All the test you suggested shows is that the error correction is working within it's capability. Reading the link I posted will show you in more depth what I'm trying to get across.
  The only part of the article I disagree with is the idea that no degradation of signal occurs with increased error correction, hifi has shown this not to be true as the better sounding players use mechanical methods to try to obtain the best read possible thus requiring correction as little as possible. I think it has something to do with time but then I think everything has something to do with time.
   Cheers, Ben.

[JoshK]

With all due respect, Browntrout, you haven't shown any evidence that this is audible.  You claim it is, but where is your evidence?