Audio Blind Testing

[audiventory]

5 hours ago, fas42 said:

you mean the player, DAC, preamp, speakers, room, setup skills have absolutely nothing to do with it ... ?

 

I meant that comparison of 2 resolutions or DSD vs. PCM is technically non-correct (methodologically impossible) in general case. It do not depend on equipment.

[fas42]

6 hours ago, gmgraves said:

 

No, my results have answered my questions, not I. And if things change over time, then they change over time, but if you can't notice or measure the changes, what difference does it make?  I mean it's a lot like the classic dilemma; "If a tree falls in the forest, and there is nobody (and nothing) around hear it fall, does it make any sound?" The answer, is of course, that it doesn't matter and neither does this.  

 

Things changing over time are very, very important to me ... because the first time I experienced a step jump in subjective quality of playback, demonstrating what was possible, a huge frustration was that this major difference in the subjective presentation was very fragile - it would vanish like the morning mist, as I was listening - over and over again. I could not get an upper hand on this behaviour at the time, but it made me enormously aware that the slightest "imperfection" in the setup could be enough to degrade key areas of integrity - which I've spent years tackling, on and off.

 

It is entirely possible that a system may be so capable in every other way that a non-optimised interconnect won't be audibly significant - but I haven't come across such a situation yet. When a majority of systems can produce competent sound, without taking special care, then the relative importance of such things will be easier to determine.

[fas42]

4 hours ago, audiventory said:

 

I meant that comparison of 2 resolutions or DSD vs. PCM is technically non-correct (methodologically impossible) in general case. It do not depend on equipment.

 

I was agreeing with you, but made a poor attempt at humour, regarding how people can "know" what is the cause of something being better or worse in a complex system, when there is almost no information ... sorry for the confusion!

[audiventory]

21 minutes ago, fas42 said:

I was agreeing with you, but made a poor attempt at humour, regarding how people can "know" what is the cause of something being better or worse in a complex system, when there is almost no information

 

It is reason, why I very careful to results of any tests. Here example of careful test http://www.enjoythemusic.com/magazine/viewpoint/0716/High_Resolution_Audio_Perceptual_Evaluation/

I consider it as learning of certain case (hardware+software) only. But the test methodology and execution are done very carefully.

[pkane2001]

7 hours ago, audiventory said:

 

I meant that comparison of 2 resolutions or DSD vs. PCM is technically non-correct (methodologically impossible) in general case. It do not depend on equipment.

 

Wait... Ultimately, what matters to me is the analog output. Why is it that I can't compare analog output produced from DSD to that produced from PCM? I assume that you mean that one can't draw a general conclusion about DSD vs PCM from such a comparison, but I can certainly do a comparison in my own set up, no?

[audiventory]

1 hour ago, pkane2001 said:

Why is it that I can't compare analog output produced from DSD to that produced from PCM?

 

We can compare format implementation.

But we can't compare formats as itself.

 

Example #1:

 

It mean that all the cases are possible:

1) PCM better DSD

2) DSD better PCM

3) No difference

 

We can't claim that DSD better PCM (and contrary) for all cases.

 

Example #2:

 

1) 24 better 16 bit

2) 16 better 24 bit

3) No difference

 

We can't claim that 24 better 16 bit (and contrary) for all cases.

[pkane2001]

6 minutes ago, audiventory said:

We can't claim that DSD better PCM (and contrary) for all cases.

We can't claim that 24 better 16 bit (and contrary) for all cases.

 

Of course, there's no general set of factors that can be applied across the board for such a comparison , nothing that would declare one format as overall better than the other.

 

But, if you pick some specific factors to compare, you can certainly make a comparison and reach a conclusion that one format is better than the other in respect to these factors. For example, noise levels, frequency response, dynamic range, etc.

[audiventory]

9 minutes ago, pkane2001 said:

But, if you pick some specific factors to compare, you can certainly make a comparison and reach a conclusion that one format is better than the other in respect to these factors. For example, noise levels, frequency response, dynamic range, etc.

 

Format (description and figures) declare approximatelly technical potential.

Implementation can release the potential or not.

 

If we compare DSD and PCM we can't compare noise level, dynamic range, frequency response. Because noise shaping may be applied different ways.

 

When we compare 24 and 16 bit implementation, we compare also clock generators' stability, noise of amplifier chips into DAC, etc.

 

When we compare mathematically modeled signal that stored into 16 and 24 bit files, we can discuss about format comparison in digital domain.

 

When we compare mathematically modeled signal that stored into PCM and DSD files, we can't discuss about format comparison in digital domain, because implementation of noise shaping is matter.

[pkane2001]

2 hours ago, audiventory said:

 

Format (description and figures) declare approximatelly technical potential.

Implementation can release the potential or not.

 

If we compare DSD and PCM we can't compare noise level, dynamic range, frequency response. Because noise shaping may be applied different ways.

 

When we compare 24 and 16 bit implementation, we compare also clock generators' stability, noise of amplifier chips into DAC, etc.

 

When we compare mathematically modeled signal that stored into 16 and 24 bit files, we can discuss about format comparison in digital domain.

 

When we compare mathematically modeled signal that stored into PCM and DSD files, we can't discuss about format comparison in digital domain, because implementation of noise shaping is matter.

 

Yes, so, as I said:

Quote

 I assume that you mean that one can't draw a general conclusion about DSD vs PCM from such a comparison, but I can certainly do a comparison in my own set up, no?

 

[marce]

19 hours ago, sandyk said:

 

 When Marc gets his act together, he really should try a decent low noise Linear PSU for his antiquated (In SQ terms) beloved SBT ! He may then start to get a bit more adventurous and discover that his theoretical based dogmatic assertions on what is, and isn't audible, aren't always correct, as it's resolving abilities further improve.

I don't doubt that Marc is a whiz in the area he works in, but you can't always directly transfer all of this to other fields such as Audio, or even Digital Video for that matter, which also includes Audio.

Er don't forget I have been doing this for a for last 30 odd years, that is audio systems and before that an amateur musician playing in Brass bands when I was a kid, bass in various bands when a teenager and older.

Worked on audio and video over the years, don't forget we have professional audio companies on our books.

So years of listening and learning, the biggest thing I learned was don't trust your perceptions...

[gmgraves]

10 hours ago, fas42 said:

 

Things changing over time are very, very important to me ... because the first time I experienced a step jump in subjective quality of playback, demonstrating what was possible, a huge frustration was that this major difference in the subjective presentation was very fragile - it would vanish like the morning mist, as I was listening - over and over again. I could not get an upper hand on this behaviour at the time, but it made me enormously aware that the slightest "imperfection" in the setup could be enough to degrade key areas of integrity - which I've spent years tackling, on and off.

 

If I understand correctly what you are saying, I see people making this claim all the time, and frankly as a many year owner of tube gear, I don't buy it.  People cannot detect gradual changes in an audio system's sound (an analogy is the frog in the pot of heating water). Many's the time I have been astounded at the difference when changing-out older tubes in an audio system. I never noticed the gradual deterioration of SQ until I changed tubes, and then the difference was considerable and immediately noticeable. Also, human beings have no long-term memory for sound quality either. So when I hear people talk about how their components "burn-in" over many weeks or months, I look upon that statement with jaundiced eye, because nobody can possibly remember what something sounded like weeks or months before. If one insists that they can, then I say that person is either delusional or a liar. 

 

Quote

 

It is entirely possible that a system may be so capable in every other way that a non-optimised interconnect won't be audibly significant - but I haven't come across such a situation yet. When a majority of systems can produce competent sound, without taking special care, then the relative importance of such things will be easier to determine.

 

Can you please explain what a "non-optimized interconnect", and how does one know that they have found an "optimized" one? There are thousands of interconnect brands and models out there. How does one know that the interconnect that they just bought is "optimized" over the thousands of other choices of interconnects that they could have bought?

[Albrecht]

1 hour ago, marce said:

Er don't forget I have been doing this for a for last 30 odd years, that is audio systems and before that an amateur musician playing in Brass bands when I was a kid, bass in various bands when a teenager and older.

Worked on audio and video over the years, don't forget we have professional audio companies on our books.

So years of listening and learning, the biggest thing I learned was don't trust your perceptions...

Well if it's years of comparing lo-fi, consumer products to each other, - then you're likely missing the point. And of course, (not sure if you've explained anywhere what you mean be perceptions. In fact, I think that everyone would do better to take a step back and THINK before they knee-jerk the tired cliche of our hearing ability sucks and cannot be trusted). If one is engaged in comparative listening tests with plenty of controls, - one comes to an understanding fairly quickly that whole systems vary significantly in performance.

[sandyk]

2 hours ago, marce said:

Er don't forget I have been doing this for a for last 30 odd years, that is audio systems and before that an amateur musician playing in Brass bands when I was a kid, bass in various bands when a teenager and older.

Worked on audio and video over the years, don't forget we have professional audio companies on our books.

So years of listening and learning, the biggest thing I learned was don't trust your perceptions...

  Marc

Yet you are still using an antiquated SBT with a SMPS ?

Are you afraid that you may learn something that unsettles you if you power it with a decent Linear PSU  or a low noise SMPS with the earth side of it's supply earthed ?

There is quite possibly another U.K. member who is willing to loan you a more suitable PSU to try with that mediocre SQ device.

Yes, I have heard a couple of SBTs owned by a friend who worked in I.T. , both before, and after modifications and a better PSU which did result in a very worthwhile improvement. The unmodified SBTs also sounded  a little lacking when compared with another I.T. friend's (Greg Erskine) Transporter which is still easily outperformed by more modern DACs.

 

Alex

[sandyk]

2 hours ago, gmgraves said:

Also, human beings have no long-term memory for sound quality either. So when I hear people talk about how their components "burn-in" over many weeks or months, I look upon that statement with jaundiced eye, because nobody can possibly remember what something sounded like weeks or months before. If one insists that they can, then I say that person is either delusional or a liar. 

George

 Up until that point I was in agreement with you about slow degradation not being noticeable, but I don't use Tube amplifiers, and semiconductor amplifiers don't slowly degrade in SQ unless due to something like electrolytic capacitors in the PSU drying  out . That is why these days I often check PSU electros for domed tops and signs of leakage.

This became more necessary due to the Pirated electrolyte episode that helped to destroy so many PC Motherboards.

 

I rely heavily on the depth of image and listen for a 3D type presentation with good source material for equipment evaluation after modifications etc., mainly using headphones for this.  I DO notice if this aspect is as good as it was even a year earlier when doing this. If it isn't, I check my B.P. before digging deeper into it.

Perhaps I am a LIAR ???

 

Alex

[fas42]

2 hours ago, gmgraves said:

 

If I understand correctly what you are saying, I see people making this claim all the time, and frankly as a many year owner of tube gear, I don't buy it.  People cannot detect gradual changes in an audio system's sound (an analogy is the frog in the pot of heating water). Many's the time I have been astounded at the difference when changing-out older tubes in an audio system. I never noticed the gradual deterioration of SQ until I changed tubes, and then the difference was considerable and immediately noticeable. Also, human beings have no long-term memory for sound quality either. So when I hear people talk about how their components "burn-in" over many weeks or months, I look upon that statement with jaundiced eye, because nobody can possibly remember what something sounded like weeks or months before. If one insists that they can, then I say that person is either delusional or a liar. 

 

 

For me, the change is a profound, subjective one; and has very distinct, rock solid characteristics. The most 'impressive' one is that the speaker drivers become completely non-locatable - if you had a blindfold on, you could stumble around the room and no matter where you were in relation to a particular speaker, you would be unable to point to a driver making the sounds. When I first got a system working this well that quality lasted about ten minutes, and then steadily vanished - it was back to normal audio, meaning it was obvious that the right, and left speakers were the source of the sound.

 

Which means that all I worry about is whether that 'switch' is on, or off - invisible speakers, or normal? The problem is, that everything in the entire system has to be working with a high level of integrity, and just one tiny thing not quite good enough is enough to kill this illusion.

 

Quote

Can you please explain what a "non-optimized interconnect", and how does one know that they have found an "optimized" one? There are thousands of interconnect brands and models out there. How does one know that the interconnect that they just bought is "optimized" over the thousands of other choices of interconnects that they could have bought?

 

The interconnect should have the electrical characteristics of a short, soldered link or track on a circuit board - anything less than that could cause audible issues. The technique I use is single core wire, shielded as best as possible, no longer than necessary, and soldered directly to the appropriate circuit nodes at either end - the connection has lost the properties of a plug-in cable, and everything benefits. Of course, this option is not open to most people - and this is something the audio industry needs to address.

 

If a plug-in cable has something about its construction that gets it closer to not causing audible degradation, in a particular system, then it will be better - and that is something that's mighty hard to predict.

 

[fas42]

8 hours ago, audiventory said:

It mean that all the cases are possible:

1) PCM better DSD

2) DSD better PCM

3) No difference

 

We can't claim that DSD better PCM (and contrary) for all cases.

 

Example #2:

 

1) 24 better 16 bit

2) 16 better 24 bit

3) No difference

 

We can't claim that 24 better 16 bit (and contrary) for all cases.

 

I don't buy into any of the hires, SACD, DSD thing - classic Redbook CD has always had everything for optimum sound. The differences in the output sound are because of the implementation of the playback circuitry, and if the latter is done well enough then format is completely irrelevant.

[sandyk]

26 minutes ago, fas42 said:

When I first got a system working this well that quality lasted about ten minutes, and then steadily vanished - it was back to normal audio, meaning it was obvious that the right, and left speakers were the source of the sound.

Frank

 That is exactly the reason I got involved in the area discussed in the DIY Audio attached link.

I found that the front end was slowly drifting in and out of optimum balance of the input Differential Pair of the amplifier, despite using 2 well matched transistors for both HFE (Gain) and VBE ( Voltage between Base and Emitter).

 

Alex

http://www.diyaudio.com/forums/solid-state/133018-current-mirror-discussion-15.html

 

[gmgraves]

3 hours ago, sandyk said:

 Up until that point I was in agreement with you about slow degradation not being noticeable, but I don't use Tube amplifiers, and semiconductor amplifiers don't slowly degrade in SQ unless due to something like electrolytic capacitors in the PSU drying  out . That is why these days I often check PSU electros for domed tops and signs of leakage.

This became more necessary due to the Pirated electrolyte episode that helped to destroy so many PC Motherboards.

 

My word, Alex! My point had little to do with tube amplifiers, I was just using those to illustrate how performance can degrade slowly over time and we not notice the degradation until we remedy it. I'm surprised that you didn't pick up on the fact that the tube anecdote was just an example of that phenomenon. I use only solid-state these days myself.

 

3 hours ago, sandyk said:

I rely heavily on the depth of image and listen for a 3D type presentation with good source material for equipment evaluation after modifications etc., mainly using headphones for this.  I DO notice if this aspect is as good as it was even a year earlier when doing this. If it isn't, I check my B.P. before digging deeper into it.

Perhaps I am a LIAR ???

 

I seriously doubt that you are a liar. All psychoacousticians agree that human memory of how something in a musical performance sounds is extremely short (strangely though, this does not apply to voices. We recognize voices that are familiar to us almost instantly and we never forget them. This is tied to some primordial survival skill). Oh, you can concentrate on one aspect of the sound and remember it as a general impression, but not in fact. What I mean by that is something like: "I went to the symphony last night and I couldn't get over how smooth the strings sounded." Now you will remember your impression of the strings, but you won't be able to remember what the strings actually sounded like, just your reaction to what they sounded like. In your case you have a mental impression of some aspect of past imaging, and you are comparing that to a current impression of imaging, but here's the rub. Your remembered impression of anything really specific, might be inaccurate. Specific impressions are very subject to the vagaries of human memory. An excellent example of this is the eye witness to a crime who is sure that he saw the defendant commit the crime, and is later found out to have fingered the wrong man. A lot of research has been done in this field recently. It turns out that memory in humans doesn't work like a recording (which was supposed for decades), that is to say, it's not continuous. The brain "refreshes" the memory every time it is brought to consciousness. IOW, it recalls it and refiles or 're-writes" it. When this happens, all kinds of non associative forces creep into it changing it subtly. Not saying that this is happening in your example, but it's possible simply because you are human. 

 

[sandyk]

33 minutes ago, gmgraves said:

Your remembered impression of anything really specific, might be inaccurate. Specific impressions are very subject to the vagaries of human memory. An excellent example of this is the eye witness to a crime who is sure that he saw the defendant commit the crime, and is later found out to have fingered the wrong man. A lot of research has been done in this field recently. It turns out that memory in humans doesn't work like a recording (which was supposed for decades), that is to say, it's not continuous. The brain "refreshes" the memory every time it is brought to consciousness. IOW, it recalls it and refiles or 're-writes" it. When this happens, all kinds of non associative forces creep into it changing it subtly. Not saying that this is happening in your example, but it's possible simply because you are human. 

 

 I am well aware of this, after years later re-reading my notes that I had to prepare for a Divorce action many years earlier.

 

Much of the testing I do after modifications involves listening to material where there is an excellent 3D type image.

Either the illusion sounds very real , or it is missing something compared with how well you know that it SHOULD sound. 

 Occasionally, I will be surprised at how good it sounds this time, and realise that it is a further improvement.

That is very different to the situation you quoted.

Quote

 "I went to the symphony last night and I couldn't get over how smooth the strings sounded." Now you will remember your impression of the strings, but you won't be able to remember what the strings actually sounded like, just your reaction to what they sounded like"

 

[gmgraves]

3 hours ago, fas42 said:

The interconnect should have the electrical characteristics of a short, soldered link or track on a circuit board - anything less than that could cause audible issues.

 

Well, that definition fits any properly made shielded cable of less than 10 ft in length; meaning that at audio frequencies, there is no measurable or visible (as on an oscilloscope) difference between a signal at one end of a soldered link or printed circuit track or one end to the other of a short, audiophile "length" (18" - 2 meters) of coax terminated in RCA plugs! A cheap Radio Shack interconnect meets those standards.  

3 hours ago, fas42 said:

The technique I use is single core wire, shielded as best as possible, no longer than necessary, and soldered directly to the appropriate circuit nodes at either end - the connection has lost the properties of a plug-in cable, and everything benefits. Of course, this option is not open to most people - and this is something the audio industry needs to address.

 

If you are saying that RCA connectors suck, you won't get any argument from me. How this connector made the jump from a connector invented by RCA Victor in the mid-forties to transfer the IF signal from the tuner of an RCA model 4630 TV (the first commercial post WWII TV set) to the first IF amplifier on the main chassis, to the ubiquitous universal Hi-Fi interconnect connector, is beyond me! It's designed wrong (make hot before ground [earth] and break ground before hot???!!! This is an invitation to blown speakers and even damaged components!), in it's simplest form it is unreliable (it was designed to connected once and not touched again unless the TV tuner needed maintenance), and seldom makes a gas-tight connection. The earliest use as an audio connector that I've ever seen was in a 1949 RCA Victor console TV that my family had when I was a child. On the back of the TV was a single RCA jack marked "auxiliary audio". The set itself was a "console" model with a 12-inch woofer and a 2-inch cone "tweeter" mounted inside of an unfinished plywood bass reflex cabinet, sitting on three springs inside the bottom part (chassis in the top) of the furniture cabinet. The chassis had an 8-10 watt amplifier on-board using a pair of 6V6 output tubes in the push-pull "Williamson" circuit configuration. To invoke this rear-mounted RCA jack, one pulled the on/off/volume knob toward one. Normally, to watch TV  you rotated the control clockwise to power the TV and control the volume. Pulling the control instead of rotating it left the TV circuitry off and the screen blank and allowed you to control the volume normally. Around 1953, my dad built an early Heathkit FM tuner and that sat on top of the TV console and played through the RCA jack on the back. To my young ears, it sounded pretty good too! When we replaced the TV with our first color set in about 1958, the "auxiliary audio" jack was lost and the FM tuner fell to me and my nascent Hi-Fi enthusiasm.  

 

[fas42]

51 minutes ago, gmgraves said:

I seriously doubt that you are a liar. All psychoacousticians agree that human memory of how something in a musical performance sounds is extremely short (strangely though, this does not apply to voices. We recognize voices that are familiar to us almost instantly and we never forget them. This is tied to some primordial survival skill). Oh, you can concentrate on one aspect of the sound and remember it as a general impression, but not in fact. What I mean by that is something like: "I went to the symphony last night and I couldn't get over how smooth the strings sounded." Now you will remember your impression of the strings, but you won't be able to remember what the strings actually sounded like, just your reaction to what they sounded like. In your case you have a mental impression of some aspect of past imaging, and you are comparing that to a current impression of imaging, but here's the rub. Your remembered impression of anything really specific, might be inaccurate. Specific impressions are very subject to the vagaries of human memory. An excellent example of this is the eye witness to a crime who is sure that he saw the defendant commit the crime, and is later found out to have fingered the wrong man. A lot of research has been done in this field recently. It turns out that memory in humans doesn't work like a recording (which was supposed for decades), that is to say, it's not continuous. The brain "refreshes" the memory every time it is brought to consciousness. IOW, it recalls it and refiles or 're-writes" it. When this happens, all kinds of non associative forces creep into it changing it subtly. Not saying that this is happening in your example, but it's possible simply because you are human. 

 

 

Which is why I have not the slightest interest in what a "system sounds like" - this is irrelevant. What matters is whether elements of recordings that I know well are reproduced without obvious problems, and virtually all systems I casually come across fail by this test - it's defective behaviour of the playback chain that's audible, that needs to be addressed; which automatically leads to optimum reproduction of all one's recordings.

[fas42]

4 minutes ago, gmgraves said:

If you are saying that RCA connectors suck, you won't get any argument from me. How this connector made the jump from a connector invented by RCA Victor in the mid-forties to transfer the IF signal from the tuner of an RCA model 4630 TV (the first commercial post WWII TV set) to the first IF amplifier on the main chassis, to the ubiquitous universal Hi-Fi interconnect connector, is beyond me!

 

 

It's a huge weak link ... the only solution is to completely bypass it; the first thing I do when optimising a setup. Otherwise, I might as well try driving a Ferrari with flat tyres ...

[gmgraves]

2 minutes ago, fas42 said:

 

It's a huge weak link ... the only solution is to completely bypass it; the first thing I do when optimising a setup. Otherwise, I might as well try driving a Ferrari with flat tyres ...

 

That may be just a wee bit of hyperbole, but I get your gist! A Ferrari with flat tires..... I'd rather see a church fall down! 

[sandyk]

10 minutes ago, gmgraves said:

Well, that definition fits any properly made shielded cable of less than 10 ft in length; meaning that at audio frequencies, there is no measurable or visible (as on an oscilloscope) difference between a signal at one end of a soldered link or printed circuit track or one end to the other of a short, audiophile "length" (18" - 2 meters) of coax terminated in RCA plugs! A cheap Radio Shack interconnect meets those standards.  

 Yet a device such as a DAT player or even a Nakamichi Tape Deck sounds better when numerous soldered steel links are replaced by copper wire links !

[gmgraves]

21 minutes ago, fas42 said:

 

Which is why I have not the slightest interest in what a "system sounds like" - this is irrelevant. What matters is whether elements of recordings that I know well are reproduced without obvious problems, and virtually all systems I casually come across fail by this test - it's defective behaviour of the playback chain that's audible, that needs to be addressed; which automatically leads to optimum reproduction of all one's recordings.

 

I'm sorry, I'm having a bit of trouble following this line of reasoning. If it isn't about what a system "sounds like", then what is it about? I mean, if recordings are "reproduced without obvious problems" then isn't that what the system "sounds like"? One that reproduces music without obvious problems as opposed to sounding like a system that doesn't reproduce music without obvious problems?