Filtering

January 27, 2018
 by Paul McGowan

Our brains edit what we see and hear like a film director's viewfinder. You're probably familiar with this concept every time you take a picture of a distant mountain, seascape, or harvest moon and are disappointed with the outcome. What your vision records is entirely different than what the camera sees. Our brains edit out everything not of interest in the same way a telephoto lens removes extraneous clutter.

The same is true for listening. We often focus on a single aspect of music to the exclusion of everything else. We're so enamored with one aspect or another, like stunning dynamics, that we ignore evidence the channels are reversed.

Another way we use filtering is by suggestion. If I am sharing the listening room my fellow listener might ask if I am hearing this or that. "Did you notice the extended highs?" In fact, I did not, but now that they've been pointed out to me I can focus on them. I may or may not agree with the observation, but I can hone in like a microscope coming into focus.

Some would suggest this selective filtering is at the root of the dreaded Placebo Effect where we are said to hear that which is not real. I would propose it is often this filtering aspect of our ear/eye/brain mechanism at work instead.

Just because someone points out something you hadn't heard before, and then you do, does not mean it wasn't there in the first place.

"Check that out!" helps focus our filters and demonstrates we are more than just mere machines.

Subscribe to Paul's Posts

24 comments on “Filtering”

  1. In the world of non recorded sounds the cocktail party effect proves how we can focus on specific sound sources in a chaos of individual sounds being mixed. Stereo however offers an unnatural mix of sounds from individual drivers being placed at two different locations in the listening room and distributed along the baffle of each speaker. Stereo is based on fooling (acoustic illusion) our ears/brain creating phantom sources. There are specific test CDs that claim to create clearly defined sound sources that should be located at defined spots in the virtual room. In my experience not every listener gets the claimed results. And only few stereo systems can recreate the sound intended and heard by the mixing engineer. Thus how far away will be the sound of a music track from the original mix when presented in an ordinary listening room? How many stereo systems I could listen to presented a cettle drum as a gong or a violin as a viola! I think that the majority of listeners of recorded music doesn’t focus on such details but rather on the PRAT factor of a stereo system - as owners of ghetto blasters do too.

  2. When we listen to music I'd suggest the most rewarding approach is to have a system that is transparent but does not emphasise any aspect of the sound reproduction. We can then be open minded and potentially discover new aspects of a recording. Stating the obvious perhaps but if there is something that is immediately 'impressive' (e.g. dynamics) this may obscure some subtleties of the performance (e.g. counterpoint, accompaniment).

  3. Paul, this seems like a straw man: "Some would suggest this selective filtering is at the root of the dreaded Placebo Effect where we are said to hear that which is not real." This seems fallacious to me, as I don't know anyone who would claim that the placebo effect has anything to do with selective filtering, which is a a separate and widely studied characteristic of human sensory processing.

    I think you might be referring to the "suggestibility effect" (or memory conformity), which is not the same thing as the placebo effect. Concise definition from Wikipedia: "Suggestibility is the quality of being inclined to accept and act on the suggestions of others where false but plausible information is given and one fills in the gaps in certain memories with false information when recalling a scenario or moment. Suggestibility uses cues to distort recollection: when the subject has been persistently told something about a past event, his or her memory of the event conforms to the repeated message."

    So when a couple of audiophiles are evaluating a new piece of gear, one might say, "Wow, those highs are much more extended." The other will then focus on the highs, but the key here is that a false memory of what the highs were like previously will be filled in by the listener's brain. The highs may or may not be extended, but the listener who had not focused on that aspect and has no memory of it will create one. The same thing happens with eyewitness accounts of crimes. If one person says, "The robber had black sneakers," other eyewitnesses who didn't in fact notice the sneakers will subconsciously be motivated to create that memory. That is NOT the placebo effect.

    1. I wouldn't agree on the straw man because that implies intentional bias which was never the case. However, I do agree with and recognize the power of suggestion, which is why I always temper my decisions over time and a wide variety of music.

      1. Good point, Paul. I certainly didn't mean to imply that you were being intentionally misleading. Whenever I read, "Some would suggest..." a red flag goes up in my skeptical brain, but I stand corrected on the misuse of "straw man."

    2. Placebo also, because we become so used to natural variations that given a new situation, e.g., a pill for some kind of ailment, we pay greater attention to feelings that we had come to filter out of consciousness, or at least diminish in attention.

  4. "I would propose it is often this filtering aspect of our ear/eye/brain mechanism at work instead."

    This marvelous filtering effect is defeated in "street" recordings. The background noise which seemed innocuous in the moment becomes an impenetrable mask over the desired conversation.
    Hearing aids also thwart this mysterious ear/brain mechanism, seemingly over-amplifying background noise.
    Can anyone explain how we can see and hear so selectively? And why mic/speaker combos cant' come close to replicating the filtering effect?

  5. Just as there are "wine tasters" who have a specialized sense for tasting aspects found in different wines, those who are audiophiles may be considered "audio tasters."

    It might explain why some simply hear no differences with fuses and power cords. They simply like their regular wine and are happy with it. I know it saves me money. 😉 So be it.

    Likewise.. some have been gifted with DNA that detects and hears aspects of audio that others are not equipped for.

    Not everyone can get a job as a wine taster. And, not every one can be a gourmet judge in a food competition. Some people are just better equipped to sense flavors and spices and textures better than others. Its nothing to be ashamed of. It can save some money and leave them content with what they are able to enjoy?

    That is why we find fast food audio listeners.. and the audiophile. Probably not about filtering so much. Its got to do with inborn sensitivities to sense what others naturally filter out.

    1. I agree with your analogy of audiophiles to wine tasters and gourmet judges. Whenever we hear a wine taster's or gourmet judge's opinion we have to always take it as just that--their opinion of what is most satisfying to their unique set of taste buds, olfactory sensors and brain cells.

      There are at least five identified basic tastes: sweetness, sourness, bitterness, saltiness and umami. [They didn't teach the last one to me in school, back when Pluto was a planet.] The degree of sensitivity to and satisfaction with each taste and combination of tastes varies among humans due to genetics, age, environment, experience and other factors. Variations in audio tastes similarly involve such factors. For that reason, there can never be an audio system that sounds best to all listeners. People will tend to prefer the sound of a system that is most pleasing to their unique ears. If realism is the most pleasing sound, they will prefer the system that renders voices and instruments most realistically. If on the other hand their "sound taste buds" prefer a different flavor, they will go for a system that presents that different flavor. Those with a sweet tooth will prefer added sugar. Filtering of audio, like the filtering of food through the use of spices and adjustment of its texture, is often an attempt to make it more appealing to individual tastes--whether that is to make it more real or to alter reality in some way. And so it goes.

  6. Once on a cruise I befriended a freelance photographer for National Geographic who also taught photography at New York Institute of Technology. We actually had similar cameras. On that trip I was shooting a Canon T70 and he was shooting a Canon F1. He taught me a lot but of all the things he taught me, he taught me how to look, how to see. What a difference it made. He pointed out shots standing right next to me I just missed.

    I think you can be trained how to listen. A knowledgeable guide is helpful but you can train yourself if you are determined. If you are a serious audiophile you have no more valuable asset than knowing how to listen and what things to listen for. I'm talking about serious music heard live and comparing it to recordings. I generally listen for several things live which is my reference. Composition, performance, tonalities, and acoustics. The last one I taught myself. Taking music appreciation courses helps with the first. Knowing top notch performers and listening to many recordings with them, hearing their comments helps with the second. Familiarity with what musical instruments sound like live and playing them yourself and with groups of people who play different instruments helps with the third. Taking the advice of audio of bimbo reviewers who are not trained to have this expertise is absolutely worthless. BTW, imaging is not an element of music.

    Do I make mistakes? Yes. Because I have to adjust my systems for each recording individually, my most common mistake is about a 1 db peak at 5 khz. I'm always on the lookout for this problem. In my type of sound system this gives bad results but it doesn't become apparent until listening for a fairly extended period and then this false brightness and clarity becomes obviously annoying. In learning to appreciate the difference in the tonality of sound using a graphic equalizer is a very valuable tool.

    Here is a recording that will demonstrate some of these factors, composition, performance, and the powerful tonality of a Steinway D piano. It does not demonstrate all that much about room acoustics though.

    https://www.youtube.com/watch?v=zGBXA1tBiLw

  7. We used to have a great filter in the pop audio community which is now lost.

    These were the A/R people that actually went and saw acts, listened to their recordings, and made recommendations to the big studios. They separated out the good from the bad so you didn't have to... and then one day the studio system crashed and they all vanished.

    Now we have a plethora of new content recorded in someone's bedroom, much of which is unlistenable.

  8. One of my pet peeves is the erroneous use of the term Placebo Effect in audio applications to describe 'hearing that which is not real'. It is no such thing. The Placebo Effect describes a situation where a specific, measurable, change to a system occurs in response to a stimulus which is not expected to give rise to such a change. For example, if Ploxobloxohexamoxin was a new drug designed to reverse baldness, and a double-blind test incorporating a sugar pill placebo showed that a statistically significant number of patients on sugar pills reported significant baldness reversal, then it would be correct to describe the sugar pills result as being a Placebo Effect. The key here is that there is a measured reversal of baldness. If, instead, the patients on sugar pills convinced themselves that they had actually grown hair when in fact they had not, this would be something entirely different, and not a Placebo Effect. [In fact it would imply that the methodology of the DBT itself was weak.] This latter scenario is the one which the anti-audiophile DBT crowd inevitably has in mind when they shout "Placebo Effect!".

    Rant over!

      1. Well, I'm reluctant to offer a proper term of use because the issue as I see it is that the situation it really describes is typically that of an inadequate (or sometimes completely inappropriate) test methodology. When you set up a test designed to measure something, you have to believe the methodology is capable of resolving the variables that it sets out to measure. Alternatively, the methodology should not allow the test subject leeway to report inaccurate outcome data. Either way, the fault is with the test methodology and not with the test subject.

        On the other hand, if you can devise a particular test methodology which delivers a (completely or partially) invalid result in a self-consistent and reproducible manner, then you can assign a specific 'term of art' to describe that circumstance - maybe it could be the "McGowan Effect". However, the key is that your new "McGowan Effect" would describe a specific test methodology, and not some attribute of the phenomena you are purporting to test. Since your original objective would have been to observe the actual attributes of the phenomena, you (or someone wishing to take your work forward) would have to then devise a new, improved, test methodology which eliminated the possibility of encountering the "McGowan Effect".

  9. I don't think the Placebo Effect or, to be more accurate, the use of a placebo, has any relevance to audio.

    If a placebo were used in audio testing, I presume it would be to ask the listener to say whether he/she prefers item A or item B, when in fact the item is the same in both cases. It is a pointless test as it would prove nothing useful, and adds nothing to ABX testing, which will tell you all you need to know.

    I went to a nice concert this evening in a small local church, 60 seats with a superb acoustic, 6 professional period instrumentalists playing 17th and 18th C dance music (Handel, Rebel, Blow etc.,). They played Vivaldi's Trio Sonata "La Follia", a favourite I'd listened to only on Thursday. Either I have a rubbish stereo or there is nothing like live music. The more I read about attaining live performance sound from home audio, the less possible I believe it to be.
    (My wife studied French Court Dance at ballet school. She always tells me how difficult it is. It was socially mandatory in Louis XIV's court, he employed 70 masters of dance at Versailles alone. I learned something this evening.)

  10. Very rarely do staged audiophile listening tests deliver definitively on the promised subject matter. Mostly their purpose is to intentionally mislead their target audience. I recall reading about a manufacturer of violins, a long, long time ago (I wish I had the link - and I'm not talking about the recent Indianapolis/Paris experiments). He claimed to have invented a procedure to make instruments whose sounds were indistinguishable from those produced by Stradivarius, Amati, etc. So he assembled an audience of prominent musicians and placed them in a room. He then invited a professional violinist in an adjacent room to play his violin as well as a Stradivarius (his idea of a 'blind test'). It was reported that the roomful of prominent musicians couldn't tell the difference.

    Well, fancy that! ... 🙂

    1. Doubters of the validity of the conclusions of the Stradivarius test would say that more than one Stradivarius copies would have had to be included (as all handmade instruments vary in sound quality and poor copies do exist) and the instruments would have to be judged in closer proximity, not just in an adjacent room. Violin players who listen to the instrument inches from their face hear (and feel) distinct sound qualities that may not be distinguishable from far away.

      I once accompanied a trained violinist to a violin shop to select a violin. We went through the store's stock of same brand and model violins to find the one that, to our ears, clearly had the most superior tone and strength. A concert violinist who plays solo for L.A. Phil and other venues said the violin and bow combination we picked sounds as good as her 20X more expensive violin and bow. So the truth of the matter, simply using one copy of an instrument for a listening test (double blind or not) leads to a result that is applicable only to that copy and the particular copies to which it was compared.

      1. The recent Indianapolis/Paris/New York experiments addressed all of the points you mentioned. They used multiple models of Stradivarius violins, plus other Old Italian instruments, and modern high-quality luthier-built violins. They used professional concert violinists and audiences of accomplished musicians. They conducted separate evaluations of the instruments based on the opinions of the violinists playing them, and the audiences listening to them. And, of course, both performers and audience were in the same room (they used both carefully-selected recording spaces and established concert venues).

        The results were interesting. Certain individual violins were consistently favored by both players and audiences. Likewise, certain other instruments were consistently disfavoured by both players and audiences. So a consistent picture emerged which ranked some of these instruments very highly and some very poorly. It turned out that the highest ranked instrument was a modern, luthier-built violin. Some Stradivarius instruments were highly ranked, but some were consistently poorly ranked.

        The tests received wide publicity of the "Modern Violin Trashes Priceless Stradivarius" variety, and while they received recognition for having been carried out in a scientifically rigorous manner, significant criticisms were received from the violinist establishment. Chief among these was that the test methodology did not necessarily enable the most cherished and prized properties of the instruments to be correctly evaluated. The actual tests involved playing short passages one after the other, rather than extended concert recitals. Therefore the results favored "first impression" reactions over a "considered appreciation" viewpoint. And that is a fair criticism. Particularly when you consider that the violins which ranked consistently highest were also those judged to have sounded the loudest, with the lowest-ranked instruments being judged to have sounded the quietest.

        We all know that when auditioning audio components using A/B testing considerable attention has to be paid to volume matching in order to avoid predictably false results, because "louder" is routinely perceived to be "better". Interestingly, one of the highly-prized property of the violins from the Old Italian masters is that they are often reputed to sound "quieter under the ear" ... in other words the violinist playing them perceives a lower volume of sound.

  11. Years ago I was in an audience that was being entertained by an excellent professional hypnotist. The first thing he did before starting the show was to have a simple test to establish who could be hypnotized. He then explained that not everyone can be.

    For those who could be? The power of suggestion under the needed circumstances cause them to sense something that others could not..

    That phenomena which can happen may explain why some automatically presume that those who have the DNA to detect differences must be suggestible and naive. When in fact, they are having denial that others can actually hear something that their own DNA denies them.

    I am coming to believe that the "it must be measureable" crowd might suffer from envy of those who have something they do not.
    In their anger they throw this "measurement" wrench into the gears of those who are sharing their enjoyment of what they do not understand.

    Just the same.. What good may come out of this?

    How was color blindness discovered? Obviously, those who are color blind can see. They can see the same objects as those who are not color blind. That form of blindness could have remained undetected for thousands of years. Maybe it was in ancient warfare with the use of banners and signal flags that caused its discovery. When someone totally botched a command given from the command.. For when a red flag meant one thing, and a blue meant another? The one receiving the command could only see one color. So its out of warfare this form of blindness may have been discovered. Just like the warfare we face in audio when "IT MUST BE MEASURED FOR!" people begin conflicts where there was none until they jumped in.

    Because of this kind of conflict we find in audio forums and blogs, someday science may discover a form of deafness in people that can hear sounds otherwise just like everyone else.

  12. This is usually why I always like to have a partner with.me when I go listen to new gear. First it's fun and second I always try to get their opinion before I give mine. And usually we come to some type of conclusion but we try to wrangle the thoughts without influencing the other person. If my conclusion differs - then I push it up against what they said and we get into discussion as to why we think as we do. More often than not we explain what we are hearing in different terms but with the same meaning.

    Anyway - it's supposed to be fun and I think audiophiles miss that sometimes. Everyone gets caught up with being 'right'. :).

    Btw Paul - your posts have once again dropped out of my mail. I stopped getting them awhile back again.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301
1-800-PSAUDIO

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram