Getting things right

August 21, 2016
 by Paul McGowan

In yesterday's post I wrote of the classic 3-way: a 2-way with the addition of yet another frequency dividing network and driver, the midrange.

Now that we have a clear idea of how all this works, let's think about what it takes to make things sound good.

Imagine a musician standing in your room. He has an acoustic guitar and sings. You are seated the same distance from the musician as you would be from your speakers. Everything sounds natural because, well, it is. The sound emanates from two sources: the guitar and his mouth. Our singer has quite a range and between the two sources of sound, we're covering areas that will eventually be shared by the top of the woofer, the midrange, and the lower parts of the tweeter - if we record the performance.

And we now understand that each of the three drivers do not abruptly end, while the next takes over in a clean transition. No, it's more like a relay race where the baton is handed over from the first runner with overlap from the second.

The point where each of the drivers meets has both drivers playing at the same time. Eventually, this overlap goes away, only to be added onto again by the next in line.

This crossover, this sharing of two distinctly different sound sources, does not happen in real life. Our singer has but only one mouth and one guitar from which sound is emitted.

So the question we must ask is how can the original performance ever be duplicated?

The answer, of course, is simple. It cannot. However, we can get close and that's what we'll start on tomorrow.

Subscribe to Paul's Posts

12 comments on “Getting things right”

  1. Again it is most obvious that a crossover is a most imprecise knife for vivisection of the music. Sewing the resulting parts together is a most difficult job for our ears being most precise instruments for localisation. Sitting in the near field they even can hear a voice oscillation between tweeter and woofer depending on the actual frequency. Having a pair of stereo speakers reproducing the singer's voice is a big challenge for our ears because they are confronted with two identical singers singing from different spots in the room resulting in heavy crosstalk phenomenons and often in different reflections from the surrounding walls. Concerning the guitar which doesn't radiate a single spherical sound wave into the room as a driver things are even more complicated. Instruments emit multiple sound waves with different directivity. Besides the damping effect of ambient air this instrument-specific directivity is responsable for the different perception between near field and far field recording / listening.

  2. Once upon a time I was an audiophile. I did everything audiophiles do, I thought everything audiophiles thought. I read magazines, ad copy, shopped, went to audio shows, drooled over equipment I couldn't afford, believed that eventually things would get better and better and better until one day technology would overcome all obstacles. In short I thought the "experts" knew what they were doing.

    But I did something else too. I went to the best engineering school I could find. For four years I worked like a dog. In this school they devised a course of study that would give me a very wide range of knowledge in many areas of science and engineering concerned that anything less would one day become obsolete. They gave me tools that were honed razor sharp. These tools were the tools you need to use to know not what to think but how to think, how to solve problems. They made it as hard as they could. Despite careful screening of applicants, 1/3 of the Freshman class never made it to graduation just as they predicted.

    And then one day I used these tools to understand sound and acoustics. And nothing has been the same for me ever since. I was no longer an audiophile, not since I saw the light.

    BTW, this is a very hard problem. Few if any in this industry are up to figuring out what the problem is about let alone solving it . Sorry to say it but the current solutions aren't even close. A well regarded expert said after attending a live opera a few years ago that the best he and his very capable team of experts could come up with after a lifetime of work was canned music. Hint, the problem basically relates to a branch of mechanical engineering called fluid dynamics. It was a very tough course and it had to be integrated with other fields of study and a lot of math, yes an awful lot of math.

    1. Fluid dynamics or better non-steady (instantaneous) gas dynamics of a compressible fluid is only one side of the medal. The other even more complex side is psycho-acoustics! 🙂

      1. Yes, the physiological response to auditory stimuli including the neurology and the way the mind interprets sound is part of the problem. It is useless to reproduce those physical aspects of sound we don't respond to and a fatal flaw to omit those aspects we do. But that never stopped an audiophile who wants FR out to 40 or more khz when even young people with normal hearing rarely can hear beyond 20 khz and those whose hearing was impaired by physical injury such as exposure to loud sounds, illness, or age are much more restricted. Nor those who want dynamic range capability far beyond that of the music they try to reproduce nor less noise and distortion already so low it can barely be measured let alone heard. There are other aspects to sound and hearing that have nothing to do with frequency, dynamic range, noise, and non linear distortion that are completely ignored and they are just as critical too. Einstein said doing the same thing over and over again and expecting different results is insanity. But that is exactly what people who engineer what are purported to be high fidelity audio systems have done for 60 or more years, convinced that they will get different results if only they could do it better than they already have.

        1. [@Soundmind]
          Mark I guess Iknow what you are talking about, but I don't understand your guardedness against maths.
          Without maths we wouldn't understand anything going on in the physical world and nowadays in the psychological science too!
          I love maths, it's so easy!
          I fear psychology - to many heads, to many minds.

          1. bernd, I'm not sure I understand what you are saying. All of my ideas come from a mathematical model of sound fields that are in part presented in graphical form in my patent. Would you prefer to see them as a series of integral equations? The model is a precise explanation of acoustic fields, specifically those generated at one point and those arriving at another point as a result. The relationship between them is a mapping function. The arriving field is taken apart using a little trick that is obvious once you see it. The goal is to measure and reconstruct this field by reconstructing the relationships between its component parts. For more than one source, the arriving field is the result of the principle of superposition. By extrapolating the location of the field elements backwards or by defining them over all points on a closed surface, it seems to arrive at the same mathematical result as Wave Field Synthesis. WFS gets there through an entirely different method. I call my theory acoustic energy field transfer theory and the mapping function the acoustic energy field transfer function. The invention based on it I call an electronic environmental acoustic simulator. WFS starts out with the Kirchhoff-Helmholtz integral and is based on the Huygens–Fresnel principle.

            For me, acoustic energy transfer theory is much easier to understand and its discovery and invention predates WFS by 14 years. WFS also has a great deal of problems with measuring real world cases because the impulse response used cannot distinguish between time related and spectral related components. I've learned this from an acoustics engineer working with WFS in Australia on a professional blog site several years ago.

            In creating working prototypes each one, mine and theirs makes enormous compromises by comparison to their ultimate goals for practical reasons. Those compromises are very different from each other and I presume so are the results.

            Why does my idea work when other peoples' don't? Because we do not listen to speakers, amplifiers, wires, or even room. We listen to the sound fields that arrive at our ears. Unless and until the others working on the problem understand it and get it right, they will never solve the problem no matter how much money or effort they throw at it They are stumbling around blind and therefore clueless. They've done about as much as they can with what they know. They've gone far past the point of diminishing returns to the point of absurdity and their results show it. Canned music.

            1. [@Soundmind]
              Well Mark,
              I know quite well, as I've read all that was published on your patent (we discussed it quite a time ago).
              What I mean is that I've always enjoyed doing maths. Maybe because it was quite easy for me. I didn't study engineering but theoretical physics, so the most work during my time on university was involved with any kind of maths.
              I also think that one has to leave the beaten track to get really closer to the original musical event - on both sides recording and reproduction.

        2. Biology games Physics by using non-linearity in ways that have not been cracked yet. We do know that aural perception has a time quantum under 3 microseconds, and this applies particularly to the phase relationships of one time events.

          This means that well trained INDIVIDUALS (not the general populations that audiologists measure) can exceed the Fourier Uncertainty Principle by a factor of ten, meaning that a guard band of the times the frequency bandwidth is necessary to replicate the neuro-physiological response to live acoustic sound. A sampling frequency of 384,000 per second is appropriate.

  3. Luckily, there is a solution: you don't need "perfect" quality to have the "original performance duplicated" - it only has to be, "good enough", for the subjective experience to fulfill one's expectations. Problem is, the vast majority of systems, most of the time, _aren't_ good enough - there is a hurdle of quality that needs to be cleared; once beyond this point everything's good!

    Why is this so? The relatively recent research field of Auditory Scene Analysis, ASA, gives the explanation - if there is enough good information in the sound field then the mind fills in the gaps, literally creates the sound picture as would be experienced as if "live", in your head. The big issue, for audio, is that the brain is very fussy, if there are too many clues that the sound is "fake" then the mind rejects the illusion - what has to happen is for the number and type of disturbing artifacts in the sound to be reduced below a certain level, which then allows the illusion to fully manifest.

      1. There are various levels of the aural illusion - at the lowest level the sound is _always_ coming from the speaker, the kitchen radio thing; at the typical audio enthusiast level there is a "sweet spot" at a symmetry point with respect to the speakers where there are phantom images separate from the speakers; at the best level these phantom images always float free of the speakers, irrespective of where one is positioned relative to the speakers, the latter are 'invisible' at all times. I call this convincing reproduction, because it is impossible to "force" the sound to come from the speakers, no matter how closely one listens to them - the illusion is so strongly constructed in one's head that it dominates you knowing, intellectually, that the sound _is_ coming from the drivers. The other benefit is that the quality of "realness" is sustained when one wanders off outside the listening area, into another part of the house - the auditory clues are never confusing, so the illusion is maintained.

Leave a Reply

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

4865 Sterling Dr.
Boulder, CO 80301

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