gsomers

I, too, have noticed that I can "re-train" my ears over extended listening. This may be something akin to what occurs in catastrophe theory (which only addresses perception of loudness).

Julf, I am both a musician and a computer scientist. I live both sides of the objectivist/subjectivist bifurcation daily, so I feel your pain. However, look at your own words ... they will point out the shortcomings in the objectivist's audio world view. For instance, your "one clear criteria" is "reproducing the original recording as faithfully as possible" (I will leave out the qualifying hints that you add, afterwards, for the moment). Is that what we truly desire? Or do we want to reproduce the original "performance" (such as it is, for that is also a hornet's nest of argument) as faithfully as possible? The two aims may not be the same. Also, what is the "recording"? Is it two wave-modulated planar surfaces that meet at some angle in a vinyl platter? What of the possibility (since most performances are not captured direct-to-disk) that considerable distortion of the original performance has occurred before the lathe cutter even approaches the vinyl surface to create the master? (And one can substitute whatever alternative sound carrier you prefer - the discussion does not change - vinyl is merely one type of sound carrier). Consider the electronic chain of events that precedes the freezing of the sound waves in plastic? What choice do you have in the matter of the execution of faithful reproduction before the creation of the sound-carrying device that you will attempt to playback as faithfully as possible? So, let's uncomplicate the argument a little by shrinking it back to your original, although debatable thesis: that the goal is to faithfully reproduce the recording, to the degree that is possible. Now we can consider your qualifications as well: "with as little distortion or added noise as possible". Again, at one level, it is difficult to object to these goals. But how does one define "distortion" and "noise". As this is an imperfect world, designers are usually in a position of trading off some types of distortion and noise for others. You probably remember that, not so many years ago, much of the audio equipment industry was pursuing lower and lower distortion figures - they were literally stumbling over each other to add places after the decimal point to their published distortion numbers. This was almost always accomplished through the injection of massive amounts of negative feedback into the circuits. The pursuit of the vanishing distortion number came to an end when many customers began to react to other distortions that were side-effects of all of the negative feedback. As is usually the case in the design world, design choices pit one set of negatives against another. If you doubt that such choices are often influenced by fad and fancy, take a look at the cosmetics of the devices that we buy over the last fifty years. Your goal is an admirable one - few would argue with it. On the other hand, I believe you would have a very difficult time getting a large group of audio enthusiasts to arrive at a consensus about what constitutes "little distortion", "added noise", "faithful" reproduction, and "artificial coloring". As always, beauty is in the eye of the beholder. And the definition of beauty also changes over time. There are those who argue that techniques employed in "78 RPM audio reproduction" have never been equalled with respect to the sound captured. Now, perhaps you may not agree with their assessment, but if we substitute "vinyl long play reproduction", or "compact disc (44.1KHz sampling rate) reproduction", or some other type of your choice, you will find that you will have many that agree with you and many that do not. For you, the phrases that you employ are self-evident. But that is the rub - nothing ever is self-evident. And that is why we have the staggering variety of choices available to us. You could fill a room with audio enthusiasts who all believe that their choices meet your objectives only to find that the equipment that they love and listen to are all over the spectrum of possibilities in terms of design philosophy, implementation, and how they sound. That is because the "science" of audio is not a science in the same way that particle physics is. Thanks to the necessity of having the human in the picture to complete the audio path, the science of audio must end at the ears. From that point on, the pursuit of perfection in audio reproduction inhabits a universe it shares with other art forms - like painting, sculpture, and, yes musical performance. Arts also have a foundation in the concrete world of science (brushes, canvas, chisels, clay, wood, steel, etc.), but arts require the human to complete the circle. Audio is no different, although there may be more science in constructing a worthy audio amplifier than there is in constructing a worthy chisel. In the end, the ear decides the success or failure of an audio device. In the end, people must trust their ears. As a very young man (already keenly interested in audio reproduction), I tried an experiment at a concert that I attended. After listening a while, I closed my eyes in order to focus deeply on the music. I kept them closed for nearly the whole of one of the four movements of a Brahms symphony. After some period of time, with eyes closed, I began to perceive a change in the sound that I was hearing. I felt it became flatter and more artificial. The audio picture of a large symphonic ensemble performing in a large space was collapsing. When I opened my eyes, the music almost instantly regained its three-dimensional quality. This early experience explains why no subsequent audio experience has ever approached, for me, the tangible reality of an actual performance in a concert space - no matter if it is a living room or a cathedral. The complete "audio" experience is not just about what my ears take in - although that has to be the largest contributor to the experience. All such performances engage us at more than an auditory level. This is why, for me, the most breathtaking audio reproduction still falls short of the real thing. Perhaps, one day, humans will develop immersive technologies that will utterly persuade our senses that the experience is 100% faithful in its versimilitude. Now THAT would be amazing - at least for these ears :-) Despite this inherent limit, I still am able to derive great pleasure in the occasional recording that is reproduced well enough to cause goose flesh or move me profoundly. Such moments are justification enough for a lifetime of involvement in the audio world. When I share that recording with another, I am thrilled to find that someone else also derived the greatest of pleasure in experiencing it. However, I am never upset when the recording does not catch fire for others as it did for me. Nor am I upset that they may have disliked it. I accept this as the way things are - we have different DNA, that's all!

If I take a .wav file, re-encode it to .flac, and then transform the .flac file back to a .wav file, AND the resulting .wav file after the transform is identical, bit for bit, to the original .wav file, then the two .wav files are identical in an information sense. To argue otherwise would be to challenge information theory dating back to Shannon and before - it's possible that something so fundamental is in error, but the challenge of disputing something this fundamental to all of modern science would be formidable, indeed. They may occupy different zones of magnetic flux on a disk drive, but, if they could be read, repeatedly, and found to match bit-for-bit (in other words, if we can ensure that the reading mechanism is not, itself, fraught with errror), then the two magnetic flux complexes are representations of the same information. However, it would not surprise me that playing back the two .wav files could elicit different responses in listeners. That is because the two playbacks can never be identical - thanks to entropy, or, as we prefer to say, the uniqueness of any user experience at a particular point in time because two different points in time are never exactly identical in a physical sense. In fact, it would not surprise me that playing back the first .wav file twice in a row could elicit different response in listeners. The equipment has changed over the course of time passing between the two playbacks (aging, for instance, no matter how slight, has occurred); our ears have changed over the time passing between the two playbacks (as we have, in a more general sense). Every experience is guaranteed to be unique because experiences, no matter how identical they may appear in our human perception, cannot ever re-occur at the moment in time that they originally occurred; we cannot take back the clock to re-experience the original event. Such repeat experiences may be as satisfying to us, aesthetically, as the original, but, at some deep level, they can never be utterly identical because too much has changed, in a physical sense, even after a moment's passing. Of course, for the same reason, such repeat experiences may not be as satifying to us. I read the Absolute Sound article on .flac quality and smiled because, although it was written to be entertaining and attention-arousing, it cannot be contested or defended. And although there are many who have challenged the assumptions and mechanics of the testing, fundamentally, when human ears are the final arbiter of suceess or failure, there is no reliable repeatability. A lot of folks have been troubled by the published results (perhaps even lost some sleep over them). This should not be the case. In fact, one could argue that the authors have succeeded admirably, in the sense that they have entertained many readers enormously. Consider the many contributions to this thread, a situation which is being repeated in many other fora. I enjoyed the original article, and I have enjoyed the many responses to it, here and in other places - responses both pro and con. One may draw any conclusion from the article(s) that they wish - I certainly have drawn my own. The conclusion that may be questioned, if it is made at all (and I am not certain that it IS made) is that the differences perceived by the listeners in the testing can be blamed, unequivocally, on the theoretical limits of .flac encoding/decoding or that they are simply a natural consequence of our being individuals that are influenced by a great many things not connected to the espoused focus of the testing, and the uniqueness of the components in the audio chain that existed between the flac files and the listeners' ears. I have one scenario to pose to all. If I play back a vinyl source twice, either in sequence or at some distance in time, would anyone argue that the experience (no matter that the equipment is or is not the same) is identical on both occasions? Consider the nature of such playback and the potential that entropy can alter the experience - even if only slightly. A person's reaction to both experiences may be highly pleasurable, but how could one go about establishing that the two experiences are identical - good or bad? Ultimately, people worry altogether too much about what others think about the playback mechanisms at their disposal. All of us, potentially, can develop tastes with respect to the reproduction of sound that will influence our purchasing and enjoyment of playback experiences. Who am I to assert that someone whose taste differs dramatically from my own is in error - rather, I can only assert that such taste differs from mine. That we are offered such an embarrassment of riches, with respect to choices for sound playback, is a matter for celebration. If someone believes that their 2.5 million dollar playback environment is superior to someone else's 100,000 dollar playback environment, they are missing the fundamental point. If such an environment is the one that offers them the best opportunity to enjoy the sound playback experience (and if they can afford it), than I am genuinely happy for them, but such an environment may not offer me the same sound playback experience. And since that someone can never listen through my ears, they can never argue the superiority of their experience over my own. They should simply be happy in their ability to satisfy their own ears. The verity of this argument is borne out in the bewildering variety of playback equipment available to us, and the strong difference of opinion evidenced in the many different audio experience publications (both print and online) concerning such equipment. The objectivists are wasting their time because there is no repeatable obective experience. The subjectivists are wasting their time because the very nature of subjective experience argues that many may disagree with their subjective evaluation of the audio experience. Such argument, is, at best, entertainment for some. As Leonard Bernstein once said in one of his collection of essays, REJOICE.