www.ambiophonics.org

CONCERT-HALL SOUND CHARACTERISTICS

In order to recreate a realistic concert-hall or opera-house sound field at home, it is necessary to know what makes a great music auditorium sound the way it does. Literally hundreds of papers and books have been written on this subject, and while physical concert hall design is now largely based on computer simulation and known acoustic principles, there is still a lot of subjective opinion and art involved. This is also the case in creating a domestic concert hall.

Concert-hall listeners, not too far back in the auditorium, usually can detect left-to-right angular position of musicians on the stage, can sense depth or the distance they are from the performer, can sense height if say a chorus is elevated on risers, can sense the size of the space they are sitting in, and sense its liveness. Some people can also sense where they are in such a space and what is behind them. When listening to recorded music at home, we want our system to provide us with the same sonic clues that the concert hall provides to its patrons present in the hall during a performance.

In this chapter we explore what makes a hall sound both real and good, so that we can determine which features of a hall we must absolutely duplicate at home in order to fool our ears into thinking that we are in a concert-hall space that is palpably real. We also need to know enough about hall parameters so that we can optimize the ambience controls of our domestic concert just as we do our stereo volume, balance, tone controls, etc.

Direct Sound and Proscenium Reflections

First, for a listener in the audience, there must be an unobstructed path for direct sound to travel from the stage to the listener's ears. This direct sound is then followed by early reflections form the back wall of the stage, the side walls of the stage, the ceiling and, to a lesser extent, the floor of the stage. These first or early reflections come at the listener from roughly the same quadrasphere as the direct sound, i.e., the front 150 degrees or so. Depending on the depth, width, and height of the stage, and its sound reflectivity, these early proscenium reflections arrive from 10 to 300 milliseconds after the direct sound and are fairly strong.

Sound-Signal Correlation

At this point we must introduce the concept of sound-signal correlation. A piece of music, on paper, such as an organ fugue, has a correlation value that represents how the present sound relates to the previously heard sound. The extent of this self or internal structural correlation, called autocorrelation, depends only on the score and the length of time over which correlation is looked for. The intrinsic autocorrelation value of the music, when it is performed, will be modified by the amplitude, delay, angle of incidence and number of reflections experienced. Correlation factors go from 0 to 1 where 1 means the next sound is completely predictable and 0 means there is absolutely no relationship between one note or transient and the next or even no relationship between the beginning of a note and the end of it.

We are also very concerned with the correlation between the sounds reaching the right and left ears. This correlation factor is called Interaural Cross Correlation (IACC). The existence of IACCs less than 1 makes stereophonic and binaural perception possible. Thus, there are autocorrelation factors that describe the signals impinging on a single earm and there are the interaural cross-correlation factors that describe the sound differences between the ears.

An example of simple autocorrelation properties is the round "Row, Row, Row Your Boat as sung by two voices outdoors. If we look at the sound over just the short period of time it takes one voice to sing "Row, Row," and the other voice to sing "Merrily, Merrily," the voices will appear to be entirely uncorrelated. But if we look at the relationship over a period of minutes, we would discover a higher value of autocorrelation since each voice eventually sings exactly what the other voice has just sung. If one voice is a tenor and one a soprano, this correlation is weakened, and if the tenor sings out of tune, softly, in French, and is indoors in the next room, the correlation factor begins to approach zero. Most people would prefer to hear such a performance with an autocorrelation factor higher than zero but still much less than 1. A "1" would imply that the tenor and soprano where singing precisely the same notes and words at the same time, in the same room milieu, and in the same vocal range.

Autocorrelation and Musical Sounds

Different types of music have different autocorrelation values when looked at through a window of three seconds or longer. For example, an organ playing in a cathedral will have a significantly larger value than a solo guitar playing outdoors. The reason all this is pertinent to concert-hall sound is that the autocorrelation value of music determines the type of ambient field that will make it sound best. Thus a concert hall may be well designed for orchestral music but be a horror for a string quartet. The advantage of the home concert hall is that, unlike the real hall, we can, if we wish, adjust our home hall to suit the autocorrelation value of each musical selection.

Significance of the Hall IACC

While hall reverberation characteristics are the key factor in coping with autocorrelation problems, it is really the interaural cross-correlation value particularly of the early reflected sounds that largely determines the quality of a concert hall and provides the best aural clues to hall presence. In the concert-hall ambience world, the IACC value largely represents what happens in the milliseconds after the arrival of a direct sound sample. Hall design research has shown that the IACC should be kept as small as possible (greatest signal difference between the ears for as long as possible) for the most pleasing concert-hall sound. This should come as no surprise to audiophiles who have always believed in maintaining as much left-right signal separation as possible.

To quote Professor Yoichi Ando, (Concert Hall Acoustics, Springer Verlag, 1945), "The IACC depends mainly on the directions from which the early reflections arrive at the listener and on their amplitude. IACC measurements show a minimum at a sound source angle of 55 degrees to the median plane." To translate this, the average person's ears and head are so constructed that a sound coming from 55 degrees to the right of the nose, impinging on the right ear, will not produce a very good replica of itself at the left ear due to time delay, frequency distortion and sound attenuation caused by the ear pinna shape and head obstruction. The IACC value for this condition is typically .36, which is a remarkably good separation for such a situation.

Ando points out that 90 degrees is not better because the almost identical paths around the head (front and back) double the leakage and, therefore, do not decrease the IACC effectively, particularly for frequencies higher than 500Hz.

By contrast, if an early reflection or any sound arrives from straight ahead, the IACC equals one since both ears hear almost exactly the same sound at the same time, and this is desirable for the direct sound from sources directly in front of the listener. That is, the direct frontal sounds should be more correlated than any reflective signals that follow in the first 100 milliseconds or so. As reflections bounce around the hall, the IACC of the reverberant field increases. The rate at which this inter-ear similarity increases determines how good a concert hall sounds when a piece of music with a particular autocorrelation value is being performed. That is why a pipe organ sounds better in a church than in a disco.

The lesson to be learned from all this correlation stuff is that early reflections in the home listening room should have as much left-right signal separation as the recording or ambience processing allows and that many early reflections should come from the region around 55 degrees.

More on Early Reflections

Some front proscenium reflections in the concert hall come from above. However, such vertical reflections strike the pinnae of both ears from pretty much the same angle with the same amplitude and at the same time. Thus these reflections are highly correlated at the ears and, therefore have little effect in adding to the spatial interest of a concert hall. In our discussions of domestic concert halls, we will, therefore, assume that early reflections from above are often deleterious, can be safely ignored and indeed, experiments with raising front reflection speakers overhead show this to be counterproductive.

In general, since music performance tends to take place on a horizontal performance plane, sonic height cues for a listener in the tenth row and further back are likely to be inaudible. For this reason and because with only two channels there is little that can be done to preserve direct sound frontal height cues, we forego height in the Ambiophonic concert hall.

To quote Ando again on early reflections: "The time delay between the first and second early reflections should be 0.8 of the delay between the direct sound and the first reflection." That is, later reflections should be closer together. "If the first reflection is of the same amplitude and frequency response as the direct sound, then the preferred initial time delay is found to be identical to the time delay at which the envelope of the autocorrelation function (coherence of the direct sound) decays to a value of 0.1 of its initial value." Ando found that first reflection delays of from 30 to 130 ms. were preferred, with the exact listener preference proportional to the duration of the autocorrelation function or the average or the average time over which the music is related to itself most strongly. That is, listeners prefer later initial reflections for organ music or a Brahms symphony and earlier ones for a Mozart violin sonata. Such a preference is perhaps intuitively obvious: for most organ music, if the first reflection arrived too soon, it would be ineffective, since the same direct note would probably still be sounding. We will make use of these rules of thumb when it comes time to set the early-reflection parameters for a given recording in our reconstituted concert hall.

We can all agree that different types of music sound best in different types of halls. For instance, symphony orchestras usually sound good in concert halls, string quartets sound better in salons or recital halls, and organs are more at home in churches or cathedrals. While one could use room treatment, panes, etc. to construct a home listening room that could very accurately mimic Carnegie Hall, this room would not be appropriate for a listener whose record collection also includes jazz, opera, madrigals, lieder and solo piano. Any home music theater must be capable of adapting quickly to each type of music being played. Fortunately the convolution technique described in later chapters makes this possible if one knows how halls work so that one can then operate the convolver intelligently.

To summarize, the front-side early reflections are the most useful in either a real or simulated concert hall and some, at least, should be centered on 55 degrees. The frequency response of this reflected sound should be similar to the direct sound. If the walls are symmetrical, then the IACC for a centrally located listener is increased, because identical reflections from central sound sources arrive at both ears simultaneously. Our listening room, like a concert hall, can be made more exciting by using an asymmetrical room shape and asymmetrical early reflection signal generation. Finally, as many concert hall designers have suggested, strong early reflection from the ceiling and rear walls should be steered or diverted to com from a direction that minimizes the IACC. We will accomplish this by room treatment and by sending only measured or desirable early reflections to the side and rear surround loudspeakers. The listening room must, therefore, have sufficient sound absorption treatment to avoid increasing the IACC by inadvertent and uncontrolled diffusion.

Reverberation

After the mostly frontal early reflections come the rear, ceiling, and rearward side reflections and reflections of these reflections form the proscenium and all the other hall surfaces. Once these reflections are so close together that the ear or even measuring instruments cannot distinguish them they are called collectively "reverberation" and form a reverberant field. The reverberant field has many parameters that concert hall designers tinker with ant that we will be able to season to taste at home. They are the sound level at the onset of the reverberant field, its density, its frequency response and such response changes with time, its angles of incidence, its diffuseness (i.e., its directionality versus intensity), its rate of decay, and its interaural cross correlation. Combinations of these reverberant train parameters allow a listener to perceive the liveness and, to some extent, with the help of the early reflections, the volume of the structure.

The reverberation preferences of concert-goers are again dependent on program material. Chamber music, jazz combos and string symphonies usually sound better with shorter reverberation times. (For the record, the official definition of reverberation time is the time it takes for the sound pressure of a single impulse to fall by 60 dB or to one-millionth of its initial strength.) Large choral works and organ recitals usually benefit from longer reverberation times, with opera stagings somewhere in between. In numerical terms, reverberation times range from over 3 seconds for cathedrals to 1 to 2 seconds for opera houses and concert halls to .5 to 1 second for recital halls or bars. Since the home listener may perhaps have a wide-ranging record collection, we must take care to see that the home concert hall can be quickly optimized for the specific recording being played.

Depth Perception

The ears' ability to detect distance is not as good as that of the eyes'. Depth localization depends on a hazy feeling for absolute loudness, timbre differences with distance (such as high frequency roll-off), time-of-arrival differences between direct and reflected sound and, if indoors, the ratio of direct to reflected sound. The first three of these factors are easily captured on recordings directly by microphones or can be manipulated by recording engineers, using delay compensation for spot microphones. Nothing in the Ambiophonic playback arrangement alters recorded depth perception base on these first three factors.

The fourth depth localization factor is sometimes difficult to preserve. If a recording is made outdoors or with microphones that do not pick up many reflections or much hall reverb, than any ambience added later during reproduction will affect all sound source positions equally. For example, increasing the level of the reverberant field makes the listener feel he is further back in the auditorium rather than increasing the distance between the from and rear instruments.

However, as a practical matter, I do not sense any loss of depth perception in my own domestic concert hall. This may be because most recordings are not dry enough to make the effect audible. But more likely, in the average live concert hall, the stage and its shell are so reflective that the direct sound of all instruments, whether located at the front or the back, has about the same ratio of direct-to-reflected sound. This front-to-back stage depth, as opposed to average distance to the stage, particularly for a balcony listener, is not easy to perceive in the typical hall. Also, in some recordings, multiple spot microphones are placed so close to their sound sources that almost no difference in the ratio of direct-to-reflected sound of any instrument is actually recorded. To compensate for this ambience pickup is then relegated to other remotely placed microphones, so again all instruments recede together. In the home reproduction system, as in the concert hall, it is unlikely that any lack of differential depth perception will actually disturb the illusion of being there.