For a long time, I thought that improving an audio system meant moving progressively toward some kind of final destination: a system capable, at last, of reproducing music in the right way. I am far less convinced of that today. The more I listen, record, teach, and compare the real sound of instruments with what remains after it has passed through microphones, electronics, and loudspeakers, the more it seems to me that the real question is not how to find the correct way to listen, but how to understand the many different ways listening can happen—and, above all, what exactly we are asking of sound at any given moment.
When I am working on a recording, for example, the system has to be almost unforgiving. It does not need to make everything sound beautiful; in fact, that would be a problem. It has to reveal an imbalance, a noise I had not noticed, excessive bass, an attack that has become too aggressive, or a perspective that no longer corresponds to what I imagined during the recording session. In that moment, listening means checking. The satisfaction does not necessarily come from beauty, but from the confidence that what I am hearing is not hiding too much. It is almost a professional relationship with reproduction: rather than being carried by the music, I am observing it.
But I do not always want to listen that way. In the evening, when I put on a record simply because I want to hear it and not because I am working, I may want something different: a voice that feels slightly more present, a piano with greater physical body, strings that do not force me to analyze every rough edge, a soundstage that offers depth and continuity before geometric precision. Some might see this as moving away from fidelity, but fidelity to what? To the recording? To the hall where it was made? To what the engineer had in mind? To what we remember from a concert? To what the performer was hearing while playing?
The word fidelity becomes less simple the more we try to define it. A recording is not reality stored inside a container; it is already a choice made through microphones, distances, acoustics, levels, editing, dynamics, and balance. Reproduction arrives much later, as the final step in a long sequence of interpretations. Perhaps, then, an audio system does not always need to pretend to be invisible. It may also have its own way of telling the story of a recording, as long as we are aware of it.
Living constantly with a grand piano makes it even harder to believe in a definitive form of reproduction. Play a note and then listen to a recording of that same note. In the room, the sound comes from the string, the soundboard, the cabinet, the floor, and the space itself. It changes if you move your head half a meter. It changes with the lid open or closed. It changes for the player and for the listener. So where, exactly, is the correct place from which to record a piano? There probably is no single answer. There are perspectives. A microphone placed very close tells us something an audience would normally never hear. A more distant placement reveals more of the room but loses some of the instrument’s mechanical intimacy. One approach may favor body, another attack, another the sense of space. Every recording chooses a point of view, and every playback system inevitably chooses another.
Cinema has made me think about this as well. The most immediate assumption is that more loudspeakers necessarily mean more realism, but that is not always true. A well-designed multichannel system can create an extraordinarily convincing space, with movement, depth, and information arriving from different directions. At the same time, two excellent loudspeakers can create a surprisingly wide and continuous scene in front of the listener. These are different experiences. One relies more on the physical placement of real channels; the other depends on the way our brain constructs sources that are not physically there, including the phantom center from which a voice can appear perfectly stable. Technology creates possibilities, but it does not automatically decide which one we should prefer.
For years, the language of high fidelity paid enormous attention to equipment and much less to the room. Yet moving a loudspeaker by only a few inches can produce a greater change than replacing an expensive component. Walls, windows, bookshelves, instruments, curtains, doors, sloping ceilings—all of them take part in what we hear. A modern system can measure the room and correct at least some of its effects, which is an extraordinary possibility, especially when the system is being used as a working tool. But correction does not mean erasing the room. We still listen inside a real space, and perhaps that is exactly as it should be.
Another opposition that interests me less and less is the one between analog and digital. A record forces me into a gesture: I have to choose it, take it out, place it on the turntable, lower the stylus. I cannot jump distractedly from one album to another with the same ease with which I scroll through a screen, and that slowness inevitably changes the way I listen. But that is a quality of the experience, not proof of technical superiority. Digital music, on the other hand, makes it possible to reach very different interpretations in seconds, compare performances, study, and discover recordings I might never have bought physically. One medium asks for concentration through ritual; the other offers knowledge through access. I no longer feel the need to decide which one is more authentic.
Older equipment has taught me something too. Some components have performance limitations that modern designs can easily surpass: more noise, less power, less control, less impressive measurements. And yet sometimes I still want to listen to them. Not because everything old must be better, and not out of compulsory nostalgia, but simply because they have a personality. At times, even a limitation can become part of the way we perceive music. Of course, it is easy to turn every flaw into poetry, and distortion is still distortion even when it comes from a beautiful object. But listening is not a permanent laboratory, and sometimes pleasure matters.
I have therefore started to think about audio systems less in terms of hierarchy and more in terms of function. When I record, I want a system that makes me doubt and reveals a mistake before it becomes permanent. When I listen in order to study a recording, I want to hear the choices contained within it. When I choose an album for pleasure, I may instead want those choices to stop mattering for a while so that only the music remains.
Precision still matters. Neutrality matters. The ability to measure, calibrate, and understand more clearly what we are hearing is a remarkable achievement. But pleasure is not a measurement error. I no longer believe that we need to choose between truth and emotion as if one necessarily excludes the other. I think we simply need to understand which of the two we are looking for at that particular moment.
Perhaps that is what interests me now more than any idea of the perfect system: the possibility of changing perspective without confusing functions. One kind of listening to understand, another to feel, knowing that the same music may ask both things of us.
In the end, the question I keep returning to is no longer how good is this system? but something much simpler:
what am I looking for in the music I am about to hear?