Noise and Objective Testing

[Jud]

Recently, over in the headphones and speakers forum, in a power cord thread, I posted about the surprising degree to which my system sound improved just by essentially eliminating a couple of extension cords and getting my interconnects and power cords nicely dressed (arranged so instead of all running together, they crossed at 90 degrees where they crossed at all). I'd say the most evident changes were a greater sense of ease in listening, and greater ability to hear details like particular bits of song lyrics that were hard to make out before.

 

I'd been thinking for some time about the "great divide" between what the subjectivists say they can hear, and the lack of evidence of their audibility in the types of tests objectivists point to, such as blind or double blind testing, "null" testing, etc. The surprising extent of the change in my system owing to a $35 hardware store power strip that allowed me to run the power cords from my equipment more neatly is just the sort of thing I can imagine *not* showing up in the types of tests relied on by objectivists. The experience has helped to crystallize my thoughts about what might be partly responsible for the "great divide."

 

Now I actually don't think either of the two sides in this great debate are borderline crazy and prone to hallucinations (subjectivist), cloth-eared clods who can't hear (objectivist), or so utterly arrogant that everyone must hear the same thing they do (both). So what could possibly be evident enough in its effects on the ear-brain that the subjectivists can hear it, but escape detection in the objectivists' tests? What I think may be at least partly the answer is, in a phrase, system (and environmental) noise.

 

The first type of test I want to talk about with respect to noise is "null" testing. Esldude - if you're reading, hi, Dennis - posted very thought-provokingly and with some frequency about this. Dennis ran many tests on his computer with software that compared the signals through, e.g., two interconnects. In nearly all cases, the results showed any differences were at levels that should be inaudible compared to the music signal. Interestingly, the deepest "null" (the least difference) between two cables in Dennis's tests occurred not between two "identical" cables, but between two different brands. That fact kept nagging at me, until it was one of the things that fell into place when I thought about my power cord rearrangement.

 

Dennis wasn't testing with lab equipment, but with his own system. That system, being regular consumer equipment, almost certainly has an internal noise level higher than well set-up lab equipment, which must have relatively low noise to perform its task acceptably. (Thank you to mayhem, whose remark in the power cord thread about purchasing used power conditioning equipment from a lab helped me think about this.) Where the differences between pieces of equipment lie approximately at or below the system noise level, measurements of those differences will tend to merge into the noise and be more random. Thus you get somewhat randomized results like two different brands of cable measuring closer to each other than two pieces of the same cable.

 

But so what? If the music signal is much louder than the noise, you're not going to hear the noise anyway, are you? I want to suggest that this is the wrong question, and that the crucial inquiry is not how much noise we *do* hear, but how much music we *don't*. But (another "but") how much can very faint noise really detract from much louder music?

 

To discuss this further, let me go off on a couple of tangents.

 

- A big step in computer graphics was using math based on the Mandelbrot set to model natural objects like mountains. Both natural objects and mathematical objects based on the Mandelbrot set have the quality that they regress infinitely. If you looked at a mountain from far away, then used successively higher magnifications, you'd keep seeing more detail, from boulders to rocks to pebbles to sand grains, etc. The same happens with images of the Mandelbrot set - no matter how high the magnification, you keep getting more detail. There is apparently something more natural looking about a rendering that doesn't have a "hard stop," even at a relatively low level compared to the pixel density of the screen.

 

- In studies of people with hearing loss, it's been found that the main problem with rendering sounds such as speech understandable isn't low volume so much as even low levels of background noise. Thus better noise reduction trumps higher amplification in making a hearing aid that works well for the user.

 

- My wife likes to hum along with music she enjoys. This bothers me to an unreasonable degree. When I am listening to my stereo, I can hear her humming under her breath in the kitchen 30 feet away over the noise of sauce bubbling. (Others may experience similar distractions. Refrigerators, Bill?) By contrast, I doubt I would notice her at all talking in a low voice at similar distance over background conversation at a party.

 

All right, so what are all these tangents in aid of? Well, the points I want to make are these: (1) Humans notice a *lack* of detail below a certain level - even a very low level - as unnatural, because nature "goes all the way down." Nature doesn't reach a level at, e.g., -80db where electrical noise obscures the rest. So even random noise at levels so low it might obscure only very faint details may alert something in the ear-brain that "Hey, this isn't actually Gillian Welch sitting in my living room strumming a guitar." (2) Obscuring even low level detail makes us strain to capture nuances that may have an outsize effect on understanding of the whole, like background noise making it difficult for people with hearing aids to understand speech. So I'm thinking low level background electrical noise, though not readily discernible on its own, may be in part responsible for phenomena such as "listening fatigue," as we strain to get the lowest level details we know should be there, but which we somehow can't make out - and which in turn has an effect on our enjoyment or understanding of the music out of proportion to the relative loudness level of the noise. (3) The obscuring of low level detail by background noise will likely be exacerbated by the level of distraction caused by that noise. So for example, I am much more distracted by my wife's humming while wanting to hear every last nuance of a recording than I would be by her quietly talking against relatively unimportant background chitchat at a party. Thus a digital audio phenomenon such as "pre-ringing" may have a distracting effect far out of proportion to its loudness level, because there is no such thing in nature as an echo that occurs before the sound that causes it. This happens in our audio systems precisely while we are listening closely, seeking auditory cues that the music we are hearing is a reproduction of real musicians in a real space. Thus when it happens, it (or at least its effect on the illusion of reality) may tend to be much more apparent than its loudness level alone might suggest.

 

I want to go on to talk about some ideas for listening tests, but as this note has already reached eye-crossing length, I'll stop here for now.

[sdolezalek]

Maybe you scared us all away... When I first read your post, I thought "there is so much here, it will take me an hour to think up something thoughtful to say in response. So now its 24 hours later and no one else has responded, so I might as well take a shot:

 

First, I think you are on to something. Second, I don't believe it is the only important something -- there are others and they may all matter. When I think of silence versus noise, I think of it in two ways: (i) is silent really silent (the "black background stuff") and (ii) is signal somehow overlaid with noise that is not audible when there is no signal. Proper shielding and proper cable dressing should help with both of these.

 

But what about slew rate or "response rate" for lack of a better term? We know that when a loudspeaker cone is fed a signal it takes a moment to respond and then actually over-reacts and then takes a moment to calm down. That absolutely introduces distortion, but interestingly not a type of distortion that our ears seem to care much about. Having said that, Meridian has done some interesting work on their disc players to eliminate pre- and post-ringing that is introduced in the system and that we do somehow notice (or at least we really notice when it is no longer there).

 

Then there is the whole square wave notion -- does it matter just how quickly we can go from zero to loud and back to zero? We refer to slew rates in amplifiers and Shunyata has tried to introduce a similar concept into the speed with which their cables pass forward bursts of power. This too is something we can measure, but again our ears don't seem to be that finely attuned to it.

 

Next comes timbre, or what I think of as timing errors. For some reason, our ears seem to be incredibly sensitive to these distortions whether we hear them as changes in pitch or otherwise, but we seem to be far more sensitive than equipment measurements would imply.

 

I believe you introduced another interesting variable in an earlier post -- the effect of filtering that occurs each time there is a D/A or A/D conversion or a sampling rate change within our electronics. It is filtering we aren't supposed to be able to hear, yet somehow do. I wonder whether the same is true of why we like Audirvana Plus 1.4 running in Direct Mode - that by eliminating as much processing as posible and taking it to the lowest core level possible, we do get a cleaner signal and noticeably so.

 

What it leaves me with is that much of what we are trying to do in improving sound is a case of subtraction. It is not about more goodness, it is about less badness, mostly in places we didn't even know it existed. In other words we are constantly trying to strip away some form of distortion or interference, much of which we never knew was there in the first place. It is why I often refer to an improvement as a "lifting of the veil." Until you hear it with the impediment removed, you didn't realize what you were missing.

 

So now we have two posts too long to digest. Anyone else care to join the conversation?

[sandyk]

Back in the early 50s ( and before that) we were taught in a Postmaster General's Dept. Training school (Australia, later called Telstra) that power cables should cross telephone cables at 90 degrees.( Installation Circular No.4, as also used in the British PMG Dept.) The same applies for Audio too.

Even well shielded interconnects should be 10CM or so above mains cables. Perhaps some expensive screened mains cables are more effective, because people don't always keep a reasonable distance betwen mains cables and interconnects ?

As far as noise goes, sjoc2000 (Jim ) told me today that he was surprised to hear the difference with his DAC betweeen power supplies that were way quieter already than a normal power supply.

Now hearing differences from 300-400 uv. and >175 uv seems strange to me, but there it is.

That was with a John Linsley Hood PSU add-on in line,as per his tnread about power supplies.

Due to a combination of age and industrial type hearing damage, I am highly aware of the effects of rise time (slew rate) of a signal. Hence, all of my DIY gear is both very low noise and wide bandwidth.

What it leaves me with is that much of what we are trying to do in improving sound is a case of subtraction. It is not about more goodness, it is about less badness, mostly in places we didn't even know it existed. In other words we are constantly trying to strip away some form of distortion or interference, much of which we never knew was there in the first place. It is why I often refer to an improvement as a "lifting of the veil." Until you hear it with the impediment removed, you didn't realize what you were missing.

 

EXACTLY !!!

 

Regards

Alex

[AudGuy]

Yes, is it real or is it Hi-Fi? Background noise is ever-present and it takes on so many forms that it would be all but impossible to identify what these sounds actually are. When I am at the symphony I listen very carefully, perhaps too analytically but I somehow feel compelled to record this sound in my brain for future reference. And so that is what I am doing ... trying to get a broad sense of what I am hearing, complete with all that background noise. And there is lots of it in an auditorium of a thousand people. And there are noises from ventilators, and a whole raft of other not quite describable sounds all contributing to the mix.

 

Alas I have no filter mechanism so these sounds are all part of the real music that I am attempting to hear. And I do hear all this noise, when the orchestra stops or is playing so low as to not drown out all the other noises. But the noise is always there, whether I am cognizant of it, or not. Really, it is part of the music just like echoes and all those other out of phase sounds that the music makers are creating.

 

BUT, in a home environment listening to a recording there may be NO other sounds, certainly none of the ambient sounds that are present in the real auditorium. So how does this change the perception of the playback? Hundred dollar question ... I suspect there will be a variable answer from almost every listener.

 

So I am brave and take my SPL meter with me to a show and try to get a real sense of the sound pressures that I am hearing in the 7th row, center seats that we sit in. From inaudible to very loud, orchestra strikes can reach very loud levels. And the levels are changing by huge amounts, swings of 50 to 60db are not uncommon. Wow, I am not listening at home to anything approaching real levels. That will have to change.

 

So I get out a good sounding SACD and get it spinning and I listen, with the critical bit in mind that well, lets try and get this stuff to real levels and I start to turn up the volume. And up, and up ... and now, according to the amp I am at 0db gain, which should be pretty close to the levels that the piece was recorded at. And according to the SPL meter I am now hearing sound pressure levels approaching those I hear in the auditorium. And yes, it does now sound more "real" than it usually does.

 

And the sound is fantastic, the music is righteous and I have a new found appreciation of Hi-Fi ... hmmm, well loud has always been a big thrill, but this aspect of getting things to near real levels has me playing music I have not played in years.

 

BUT, it is not real, does not sound real and after a while I start to realize why. There is no background noise, all that racket that is heard in the auditorium is now missing as is the part of the music mix that it must be adding to. Can a Hi-Fi system ever sound any better than just as a good Hi-Fi system, is it possible for a Hi-Fi system to ever play like it was real?

 

Maybe not.

[sandyk]

AudGuy

I think the point we are trying to make here , is that low level system noise creates a fine veil over the music that helps to obsure the low level ambience, that makes the recording sound real. In a normal domestic system, even air conditioning on low has a similar effect. It's a completely different thing with a live popular performance , due to the much higher overall levels. Try to approach those levels at home for any period of time, and you will end up with hearing damage.

Alex

[4est]

Yes, errors of omission vs errors of commission...

[mayhem13]

Great threa Jud and some good points for both camps. Low ambient noise is a top three for me when listening at home.....quiet system, quiet room, and a quiet mind!

 

If anyone has ever been to recording/mastering session in a professional studio, that type of environmental, critical listening is what home HiFI is all about....IMO anyways and not about the live performance. Home listening and live performances and so far apart, I'm often surprised they're ever compared as a reference to each other. Otherwise, I think most material we listen to would be recorded live.....which can also be fantastic to listen to.

[Blake]

Home listening and live performances and so far apart, I'm often surprised they're ever compared as a reference to each other.

 

I concur. I do think that going to live events can certainly help to be a guide in educating yourself as to what a musical instrument sounds like in that particular environment. Then, when listening to music on your Hi-Fi system, you can reference your memory of the live event to help you decide whether the instrument in question sounds "life like". However, I think that is about as far as it goes.

[4est]

OT post: I have to wonder how much the increase in Jud's sonics came from simply unplugging and re plugging everything in which basically cleans the contacts. Over the years this has confounded me occasionally in an ABA situation where the return to A was not the same as the original. This is not to belittle or claim there was no change in Jud's experiment, but simply a question and a suggestion to periodically perform this simple task.

[mayhem13]

OT post: I have to wonder how much the increase in Jud's sonics came from simply unplugging and re plugging everything in which basically cleans the contacts..

 

None.

[Jud]

OT post: I have to wonder how much the increase in Jud's sonics came from simply unplugging and re plugging everything in which basically cleans the contacts. Over the years this has confounded me occasionally in an ABA situation where the return to A was not the same as the original. This is not to belittle or claim there was no change in Jud's experiment, but simply a question and a suggestion to periodically perform this simple task.

 

None.

 

I'm interested in both these responses. I don't feel belittled at all, Forrest, it's something I hadn't considered and I'm therefore curious about. Mayhem, why the response of "none"? It wouldn't necessarily have the effect of cleaning the contacts, or cleaning the contacts would not have that (large an) effect?

[Daudio]

cleaning the contacts would not have that (large an) effect?

 

I make it a practice of applying a contact cleaner to all my audio, video, and sometimes, computer equipment. Both signal and power connections. I happen to use Cramolin, but there are other good products easily available.

 

While there are good technical rationals for contact cleaner/enhancers. I have found the improvement in a very visual form. I also use them when repairing/tuning brass model locomotives (part of the 'high-end' of model trains) and I can easily see the improvement in the running performance of the mechanisms after treating all the drive circuit connections. No question of " did I hear that or imagine it" !!

[JohnMH]

Ditto on the contact cleaner. I use DeoxIT D5 from Caig Laboratories on all sorts of electrical connections. I think this is the current name of the stuff that used to be called Cramolin? Wonderful stuff.

 

Re: background noise - Last week I set up a little near-field system using a Velodyne MiniVee with a pair of Radio Shack Minimus speakers (the small metal ones) setting right on top of the sub powered by a little Dayton Audio class D amp. Listening distance about 3 feet. Plugged it all together and fired it up using the line level in and out on the sub and began playing some Mozart. I was immediately hearing a bunch of weird noises primarily in the bass out of the Velodyne. At first I thought it was 60 hz hum, but after fooling with levels, X-over freq., etc. and playing some other recordings I realized I was hearing background noise recorded in the venue. This particular one had substantial low frequency rumbling, perhaps coming from the air handling system.

 

As others have posted above, in a live venue there is a lot of ambient noise. One's ear / brain system is very good at tuning out the stuff that's not part of the music, especially when it's constant and fairly uniform like air handling noise. In our Fine Arts Concert Hall here on the University of Wyoming Campus it is standard practice to turn the air handler off during performances and especially if there is recording going on. The room is cooled and ventilated at intermissions and pre-concert.

 

Studio recorded material played back on a good system in a quiet listening space is quite different, especially from good digital sources. Sometimes I think that part of the appeal of vinyl and analog tape is that there is always some constant background noise. I wonder if our brains interpret this as more like a "real" sonic environment.

 

Excellent thread, Jud. Thanks for starting it.

[mayhem13]

I'm interested in both these responses. I don't feel belittled at all, Forrest, it's something I hadn't considered and I'm therefore curious about. Mayhem, why the response of "none"? It wouldn't necessarily have the effect of cleaning the contacts, or cleaning the contacts would not have that (large an) effect?

 

Repeated plugging and unplugging does nothing to clean contacts, but instead degrades the contact through arcing, tension relief of spring loaded receptacle contacts and burring of the spades and slots.

 

If anybody's interested, I came across a weave loom PET type product that's heat shrink able for DIY cable and interconnect projects. The only distributor I know of is Wurth but I'm sure there's others. This stuff gives really professional results with basic tools and skills.

[sandyk]

Repeated plugging and unplugging does nothing to clean contacts, but instead degrades the contact through arcing, tension relief of spring loaded receptacle contacts and burring of the spades and slots.... mayhem13

I am in 100% agreement with that statement.

As I am heavily into DIY, and often need to do that, I have found that RCA sockets for example, end up going high resistance and fail intermittently, due to the inner of the socket losing tension. The outer "earth" side of the RCA plug can also become a loose fit due to repeated insertions.I have replaced probably close to 20 supposedly high quality gold RCA sockets over the years due to this.

Some IEC mains sockets are not without fault either. I have also had intermittent no power problems due to an IEC socket that used thin metal inserts over the main contacts, instead of solid plated contacts. The same problem can arise with quality banana plugs used for speakers.Repeated insertions can cause them to go high resistance and fail as they lose tension and contaminants get in between the sections of the plug.

Regards

Alex

[Daudio]

I realized I was hearing background noise recorded in the venue. This particular one had substantial low frequency rumbling, perhaps coming from the air handling system.

 

John, I have a very nice record of a choral work that has a heavy rumble at some point that sounds like a subway underneath, or a truck outside, but low enough that it takes a good system to hear it. I think it may be ARGO ZRG-5240 'A Procession with Carols', Choir of Kings College. Cambridge (1961). A good test record for your system, as well as beautiful sound and music.

 

And I think vinyl has other advantages, then just a slight noise masking, and there are some very quiet surfaces out there too.

 

Yea, Thanks, Jud for an interesting and civilized thread

[mayhem13]

Fans can produce some serious LF content and noise. Check out the Thigpen Rotary subwoofer. Eminent Technology: home

[4est]

Interesting how things spin. I wasn't suggesting one wear out their plugs to clean them, only that the insertion wipes the contacts in the process. Someone as Jud whom holds onto his cables might ought to remake contact annually is all. Actually cleaning them at the time would be better still. Cramolin, Caig and Stabilant acts as a lube btw.

[sandyk]

Forrest

I suspect that originally, the inners of sockets may have used something like phosphor bronze which retains it's tension. These days even the premium gold RCA sockets don't like too many insertions.All in all, the RCA plug and socket appears to be a poor design, although there are a few very expensive variants that appear to be worthwhile.

I haven't tried them though, as they are scarce and quite expensive.

I do agree with your last sentence though.

Regards

Alex

[AudGuy]

Thanks Sandy, yes I think I did miss the point although maybe I was just relating a variation of the theme. Noise ...

My experience has been that classical music does not sound right when played at lower, somewhat sane levels.

[mayhem13]

Hey Jud, just wondering if you did the same for your interconnects and speaker cables at the same time and if their was an audible difference there too. Personally, I use the shortest length interconnects as possible and right angle connectors where possible....except for the glass TosLink which can't be curved excessiveley.

[4est]

Forrest

I suspect that originally, the inners of sockets may have used something like phosphor bronze which retains it's tension. These days even the premium gold RCA sockets don't like too many insertions.All in all, the RCA plug and socket appears to be a poor design, although there are a few very expensive variants that appear to be worthwhile.

I haven't tried them though, as they are scarce and quite expensive.

I do agree with your last sentence though.

Regards

Alex

 

Well sure, as with all things, the new stuff is often junk. One needs to simply purchase a decent Hubbell hospital grade AC plug for $10-15 instead of $2 Levington's. Yes, typical RCA/Cinch plugs often suck, and yet if you cannot make/break a contact 20x before it breaks, you have had problems from the start. IMO, XLR is the single best reason to run balance cables. All I was really getting at is that it would seem that often I get a perceived bump in sonics when I take things apart and put them back again. Jud, as with myself, holds on to his gear and lord knows how long that stuff had been left plugged in. If one were to want to avoid this, I would heartily recommend Stabilant 22 on the signal and low voltage stuff, and Cramolin on the 110/220v power contacts. Avoid silver pastes at all costs as it degrades and leaves a conductive mess.

 

Sorry Jud, I didn't mean to derail this. I just thought I'd pose the question and you seemed to get my point.

[Jud]

Well sure, as with all things, the new stuff is often junk. One needs to simply purchase a decent Hubbell hospital grade AC plug for $10-15 instead of $2 Levington's. Yes, typical RCA/Cinch plugs often suck, and yet if you cannot make/break a contact 20x before it breaks, you have had problems from the start. IMO, XLR is the single best reason to run balance cables. All I was really getting at is that it would seem that often I get a perceived bump in sonics when I take things apart and put them back again. Jud, as with myself, holds on to his gear and lord knows how long that stuff had been left plugged in. If one were to want to avoid this, I would heartily recommend Stabilant 22 on the signal and low voltage stuff, and Cramolin on the 110/220v power contacts. Avoid silver pastes at all costs as it degrades and leaves a conductive mess.

 

Sorry Jud, I didn't mean to derail this. I just thought I'd pose the question and you seemed to get my point.

 

Well if it is a detour at all, it's been an interesting and useful one.

 

For everyone who asked whether I changed routing or replugged interconnects and speaker cables at the same time I did the power cords - no. I've done most of my digital-related acquisitions in the last year or two, as well as new speaker cables and a new amp. So simply by virtue of getting new cables and equipment I've replugged the interconnects associated with the digital chain, as well as the speaker cables, once per year or more. But I didn't do anything with the interconnects or speaker cables when I recently re-ran the power cords.

 

Since then, I actually have gotten a new USB cable (more to come in a thread I started over in the DAC forum specifically about that), and, in association with that acquisition, a new power strip and DAC power cord (more to come about that in the power cord thread, which for some reason is over in the headphones and speakers forum). And what makes me think about it and bring it up here is something you mentioned above, Forrest, about hospital grade plugs. For a different-than-usual point of view, have a look at what the seller/manufacturer of my new power cord and power strip (and just about all of my cables, including the new USB cable, under the Mapleshade or Omega Mikro insignia) has to say about "hospital grade" stuff: Clearview AC Power Products-Mapleshade and Power Conditioning Strips-Mapleshade .

 

I'll be putting up my promised/threatened second major post in this thread shortly, probably later today.

[4est]

Thank you for your patience Jud. FWIW, I am not sure I buy into the "sound" of a receptacle. I suggested hospital grade only for the more robust connection required. I am fortunate to have had some old 20a Hubbells around that I use. The cryo makes sense however, if nothing else it should make contact strength last longer. One thing to consider is that if you are going to redo receptacles, make sure you use the screw terminals and not the push in holes. Tell the sparky, and watch if you have to!

[mayhem13]

For frequently plugged and unplugged receptacles, Hubbel branded Twist Lock are an excellent way to go with greater surface area and less of a need for spring loaded connections where the twist insertion commands a premium contact point. They're pretty easy to install on power chords and if you know what you're doing, the receptacles are easy to swap. Will they provide any sonic advantage?.......the individuals subjective observations would have to be the judge as objective measurements and testing says no. I like the look and safety that they provide,many since I build them myself I won't be wasting hard earled $$$ on wild assed claims....allows me to sleep a bit better. Lol