Fidelity

January 25, 2019
 by Paul McGowan

The definition of fidelity is faithful. In our field, we promote the idea of a higher sort of faithful called High Fidelity.

In audio, we have different levels of fidelity: reasonable, good, and high.

And that’s what we write about and discuss on a continual basis: the level of faithfulness achieved by our systems. How true to the source? How faithful to the music?

The next time you sit in judgment of your system consider whether it’s merely faithful or if it’s achieved a higher level worthy of the moniker.

High Faithful doesn’t have quite the same ring as High Fidelity but both get the job done.

Subscribe to Paul's Posts

27 comments on “Fidelity”

  1. One day I had a customer come in to listen to some equipment and through the course of conversation he boldly stated that he wasn’t so much into High Fidelity more into Infidelity! I quickly played some music and attempted to get him on the road to redemption. Cant remember the outcome.

    1. Obviously he walked away without buying a new system. He probably didn’t like the system he had, but was only interested in going around and trying out other systems. Not the type of customer ready to commit to a new purchase.

    2. Well, you got him into the store somehow so that’s what’s important, even though he probably had no intention of buying anything at the time. You’re definitely not going to sell anything if they don’t come in.

    3. Seeing how the NAMM show is this week, I’ll say that musicians and producers are most often looking to make “a sound” they like. Sometimes one small component of that is fidelity along with choosing the microphones, mic preamps, compressors, limiters, effects, A/D converter, and processing that pleases them. Cables and power filtering might be the only things engineers might seek fidelity with rather than a more pleasant sound – the opposite of playback where salesmen/customers might want to balance a system with cables!

  2. If acuvox is right the only way to faithfully record and reproduce the sound of a voice or an instrument is to separately record each instrument and voice in mono and reproduce each sound source via a dedicated speaker. This concept seems most reasonable to me. Just look how accurately (faithfully) a voice can be reproduced by a cell phone. Things get highly complicated trying to faithfully reproduce the complex sound patterns of a chorus and an orchestra and the even more complex sound waves reflections in a concert hall. No wonder that sound engineers heavily manipulate the recorded signals including additional reverberation in order just to get a “good” sound from a more or less decent two channel only stereo system. After some decades of search for high fidelity I am convinced that hifi is a myth or a most professional marketing story. Having had some twenty pairs of different loudspeaker designs in my listening room I now finally found a pair that was able to render a faithful sound pattern of a recorded kettle drum. Obviously the other speaker designs were more or less bad compromises.

  3. There are so many variables in play when it comes to the content we listen to, being locked into the final product left to us by the engineer who did the mix, and that is not the benchmark I use to define “fidelity.” What I do utilize, is the original content of that recording, superimposed graphically with the playback from the source. This reveals the accuracy of the product from the source, compared against the original content provided to the source, especially in complex audio passages. And this is where components used in the source reveal their accuracy. For my source, this required utilizing a revised choice for the single ended I/V buffer and line driver for the output to the amp. Once revised, accuracy greatly improved.

  4. Faithful is a term used when talking about my Labrador’s, family, friends, and possibly a big geyser.

    Accurate and enjoyable are terms I use when talking audio. When I say accurate I refer to pace, rhythm, tonal quality etc; not in reference to “live” or some coined phrase of ideal.

    Highly Enjoyable is probably most important.

  5. I agree. I don’t think there that there is an “absolute sound”. Each listen experience is it’s own universe. Whether produced real time or recorded and played back. Copper once had a good discussion going about whether you would rather listen to live and real time or stay home “in the comfort of your living room” and listen to playback through your stereo. Mostly I vote stereo.

  6. How about going to listen to live music at venues that are as comfortable as your living room. Been to two such venues in the last week.

    Looking in Gramophone, “fidelity” was used as early as the second edition in May 1923, referring to “tonal fidelity”. In 1953 J H Leak & Sons Ltd used “HIFI” as an abbreviation for their telegram address. HiFi then became a generic term – fairly meaningless.

    Only later did mid-fi emerge, in the 1980s, to separate supermarket audio from the proper stuff, but “hifi” still applied to most budget stuff.

    I think “hi-end” then appeared as price drove exclusivity.

  7. You know, labels mean less and less, to me. If the Music coming out of the headphones don’t make me cringe, if the sound from my speakers sounds pleasant, and if the Music is engaging me…I quite frankly don’t give a hoot whether it is called High Fidelity or just a rig that friends say, “That sounds good to me”.

    Audiophiles, like shutterbugs share some basic similarities.
    Walk up to the photographer at his gallery opening, and compliment his eye, and he will probably point out on the print what he would have done different to make it better.

    Stereo guys can rarely just sit back and enjoy the Music, if they are thinking about what your next improvement will be.

    Contentment: Enjoying the system you could afford, and enjoying the Music you have collected.
    When in doubt, just relax and listen to Music.
    Suppress the desire and acquire syndrome, and stop looking for the secret in another, more expensive do-dad in magazines.

  8. I guess it was inevitable that sooner or later I’d think about what is meant by fidelity in audio systems. How does one determine fidelity? If you wanted to test the fidelity of color film say balanced for 6500 degrees K, you’d take a page with a color chart, you’d put it on the ground at high noon when the sun produces 6500 degree light, take a photo and compare the results with the original. The difference is distortion. Even if you couldn’t see a difference you might be able to measure one to determine the deviation. Suppose you wanted to check the geometric distortion of a lens. You’d take a perfectly rectilinear grid on a sheet of paper, photograph it and determine if the copy had various types of distortion such as barrel or pincushion distortion. In short you’d compare the copy to the reference to see what the differences are.

    But in Audio you have no way to do this. In a live versus recorded demo in the same place at the same time you might be able to do this for this specific case. Most people who try it fail. Remarkably Roy Alison succeeded with AR3. I heard it myself twice. My experience with that speaker in other places was entirely different. Acuvox claims he can do the same and I believe he can. But these are specific cases that are not typical of the way this type of equipment is used.

    So you were never at the original performance so you have no way to know what it sounded like. If you had been what you heard would have depended on where you were listening in the room. And there was no way to create a reference of the sound field anywhere in the room for later comparison with another machine. In short there is no reference recording, no reference audio system, no reference at all to go by to judge fidelity either objectively or subjectively. The term in this case has no meaning. It’s beyond the current state of the art.

    So what can be done? As I see it experience with live music at live venues can impart certain impressions that can be later recalled. The test for me is not accuracy or fidelity in an absolute sense but convincability. Does sound resemble what you would have expected to hear at a live performance of similar music where it was performed? To do this you need to have those experiences as a mental reference. If you don’t have them or can’t recall them then what does your judgment of equipment mean? People who have tried to measure it have failed because their measurements don’t correlate with what they hear. Clearly the state of the art of measuring sound has a long way to go. Even people who you’d think were experienced and whose judgment you might trust are not necessarily very good at it. John Atkinson proved that he couldn’t hear the difference between his carefully made recording of a Steinway D piano and the piano itself he’d recorded only moments earlier while his audience could. And he said he’d wondered what they were trying to tell him, to understand their message. So here’s Mark Waldrep’s latest posting, “Who can you trust?” He did not like my reply which was nobody. How odd it is that Waldrep’s goal is high resolution, high definition, whatever that means but by his own admission it is definitely not high fidelity. Now what in hell is he talking about? Can someone explain that to me?

    BTW, the battle rages on between Waldrep, Whiner, and Audioquest. They’d like to drag Paul into it but he’s too smart to play their game.

    1. “But in Audio you have no way to do this. In a live versus recorded demo in the same place at the same time you might be able to do this for this specific case. ”

      But the same is true in photography. You take a picture of a forest, a meadow, whatever and look at the result. You have no precise idea of the accuracy of the captured image.

      You can only see the distortions when you photograph a well-specified object.

      I think you can do the same thing with audio, with the difference that you’re measuring the complete chain (including microphone). Cameras (well, digital ones including their software) are a complete chain; the playback portion that we have at home isn’t a complete chain.

      1. At least in the example of the camera you have a reference test under controlled conditions. You can expand the test using different lighting, different objects, different anything but in the end you are comparing a particular piece of equipment or system performance to a reference standard.

        ISO 9000 is a system that assures the product you buy will be as close to the prototype in every way as is possible. The design includes performance specifications and tolerances. Units that fall outside those parameters need to be repaired before being sold or discarded as scrap. But to perform the ISO 9000 process you have to have a rational basis of comparison. In this case there is no rational basis of comparison. It is an event that occurred at another time and another place with no record of the technical details that satisfactorily explains it. Therefore the only thing you can rely on is memory based on experience and imagination of what a similar sound experience would be like. This is no way to run an airline. You’re flying blind…. in the fog…. at night…. through a storm….; with bullets fired at you…. and one engine is out…. with ice on the wings…. and you have a splitting headache and are nauseous….. and the copilot is dead…..and…..

        1. “At least in the example of the camera you have a reference test under controlled conditions. You can expand the test using different lighting, different objects, different anything but in the end you are comparing a particular piece of equipment or system performance to a reference standard.”

          You can do the same thing in audio. You can create standard sounds and record and play them back. The comparison is for the whole chain – sound creators (coloured objects), microphone (lens/filter/photomosaic), comparison technology (deltas betwixt bits used to emit sounds vs bits recorded; image vs standard test patterns).

          This compares what’s captured with the reference sound. It has inaccuracies and errors, but so does the photography example.

          It turns out that doing piecewise testing of the audio chain (frequency response, distortion of various sorts, speaker dispersion,..) is much simpler and provides a good guide to whether the equipment works as desired. We are now getting to the point with the piecewise ‘correctness’ that semi-perfection in electronics (5Hz-100KHz at 0.0001% THD etc) is very easy to achieve at reasonable cost. It may be time to look much harder – but it’s feasible. Just like in photography.

          1. Unfortunately it won’t work. The problem is that most of the sound you hear at a live performance is reflected sound due to the room acoustics, often a very large room. The ability to record and replay these reflections with anything remotely approaching scientific precision is well beyond the current state of the art. These reflections are 90 percent or more of what you hear live. I devised such a measuring system 45 years ago but never built one.

            1. Er, no.

              You’re confusing “current practice” with “what we could do if we wanted”

              There’s no reason whatsoever why we couldn’t measure reflected sound and see how well it’s reproduced.

              One can place bets on the outcome, of course

                1. Ummmm.. just do it?

                  Obviously, in one of some range of ‘standard rooms’.

                  The difficulty to my mind is the transducers. They need to be linear and accurate. That’s hard. Each speaker doesn’t, however, need to be wide range in frequency and dynamic range.

                  1. Just do it is not a method.

                    Here’s an added kicker for you. Different reflections of the same sound even from the same direction have a different spectral content or what you might think of as frequency response. This is due to two factors. The first is that as sound bounces off a surface the degree of energy reflected is a function of frequency. Generally higher frequencies are better absorbed rather than reflected than middle or low frequencies by practically all materials. The other is that the air itself selectively attenuates high frequencies depending on humidity. In a concert hall with an RT of 2.5 seconds the sound can travel over half a mile or more through the air before it becomes inaudible. Therefore using a pulse as a source signal to measure reflections won’t work because there is no way to separate the time of each reflection’s arrival from its spectral content. The trick to get around that is one of my few remaining secrets. So now how would you do it?

                    1. I think we’re talking at cross purposes.

                      I’m talking about measuring a standard room with sound sources. You put the sources in specified places in a room with specified reflectivity and emit sound and measure. We know what the reflections should be in time and shape from computer modeling (it’s easier to remove the walls, but if you insist…)

                      You’re talking about how to accurately record an orchestra in an auditorium. I dunno. Ask John Atkinson??

                    2. As I pointed out, most of what you hear from a symphony orchestra at a concert hall or for that matter any sounds in any large room are reflections. Without those reflections the sound is well…. dead. But the reflections must have the same kinds of qualities in the playback as you’d hear in a live concert. Therefore if you expect high fidelity you must figure out how to either capture them and reproduce them accurately or reconstruct them as I do. But if fidelity is to be high you need to understand the specific details of the reflections. Two comparably sized rooms are Boston Symphony Hall and Avery Fisher Hall. The sound of musical performances between them as heard in the audience are world’s apart. Boston Symphony hall is said to be the best room for listening to music in the United States and one of the two or three best in the world. By contrast Avery Fisher Hall is and always was very poor, a terrible disappointment that no one has been able to fix so far in over 50 years.

                      As much effort as has gone into understanding concert hall acoustics, the state of the art of modeling and measurement still leaves much to be desired. I have my own model on which my own ideas are built and they are very different. So are the results. The prototype in my home is as far as I know a one of a kind experimental machine. I don’t know that there’s another one like it anywhere. It’s not a trick or hype. There is a lot of math and engineering behind it. The difference is obvious immediately as soon as you hear it. When properly adjusted I’d call it “convincing.”

  9. What’s today’s blog entry about – Faithful to what? The entry suggests two very different answers, then encourages us to make a judgment between “merely” faithful or some higher order of the term, without suggesting what fidelity means or the reference point(s) for judging. What’s yours, Paul?

  10. Hi fidelity? All I have to do is to shift one speaker the distance of the thickness of a line made by a pen (on my audiophile desktop) and the sound changes one step closer to the real thing.

    Hi Fidelity is an elusive goal. For if its possible? Some would need to calculate a perfect formula for each component. Then, to get everything exact,precise, and right. I would be content just to feel good after listening.

  11. What I learned from this post: Hifi is a good, maybe morally superior concept for relations or marriages, less so for the reproduction of musical art. For enjoyable audio experiences, most addicts prefer the expertise of a talented sound magician (SE or mastering engineer) who makes trivial things sound bigger than life.
    Is this not Hifi anymore because it is manipulated?

    Bare, unedited recordings are mostly quite boring, only good for testing or documentation. Give it a try, record yourself, and you will see.

    Setting up a good recording environment and decent editing is an art, and not trivial. When listening to different productions, and not knowing how they were staged and edited, it is impossible for a layman to figure out if they are “faithful” or if they just sound pleasant.

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