2. The spatial sonic field
2.1 Space in Channels
In conventional audio, we transmit the spatial impressions as the signal difference between separate channels. Studio productions have reached a very sophisticated stage today in this matter. The playback is mostly very pleasant and often each voice or instrument is audibly better than in the live performance. In most cases, the goal of playback is no longer the reproduction of an existent event; the recordings are much more a product of art. Basically, though, we are not able to restore the spatial impression of the genuine event in the conventional way. If we take a closer look, we mostly manipulate mono signals in studio productions. The sound sources are recorded from closely-distanced microphones or directly from the pick-up. The positions in the stereo panorama become specified by panning during post- editing, and synthetic reverb, which ought to complete the room impression.
As long as we are pursuing the goal of pleasant perception, this sort of procedure is satisfying. However, for reproducing the clarity of a voice in an excellent acoustic environment, the hum of a bee close to our nose, or to create the emotional impression of the spatial sonic field in one of the famous concert halls throughout the world, the perception of phantom sources is collapsing.
In the previous chapter, we described the reason why it isn't possible to capture the correct spatial distribution of sound waves and reflections in the recording room by a microphone alignment. Therefore, we have to verify the question; whether the conventional basic approach of channel-orientated procedures is the right way. In any case, these channel-oriented approaches will lead to phantom acoustic sources. However, these always move dependent on the listener's position. Correct spatial depiction of the sound event remains impossible in this case. The spatial sonic field isn't describable as signal deviation between some single points because this difference is unequal at any spot in room. For authentic reproduction, we are obliged to reconstruct the spatial structure of the wave fronts with volume. That´s verifiable by the fact of whether or not the movements of a listener across the playback- room causes the same changes in perception, as an equidistant movement of a listener in the recording room. No other indication is a mark of true spatial audio.
2.2 The mirror source model
This task becomes solvable by another way, different from Blumlein`s procedure. We have to realize that not the sound source itself, neither a soloist nor an instrument, contains any spatial sonic field. If we take a closer look, it is indeed only the spatial radiation patterns that are different. However, all spatial perception is caused by the reflections of the source signal in the recording room. These reflections seem to originate by mirror sound sources behind the recording room surfaces. Their coordinates in the recording room are dependent on the arrangement of the reflective surfaces and on the position of the primary sound source.
The starting points of the first reflections are apparently arranged outside the recording room. These primary reflection starting points are the source for further reflections behind the opposite walls.
As visible in this simple animation, the second reflections don't radiate a mirrored acoustic pattern of the whole recording room, as described in some publications. This fiction leads to the nebulous picture of the spatial sonic field, which is extensively discussed in endless disputes concerning the perfect rendition.
Also the second, third and all further reflections do not radiate other signals as the pure dry audio of the primary sound source! The level of the first reflection is depending on the adjustment and directional radiation pattern of the primary sound source. That reflection is the source signal for the second reflection. Of course, the reflective behavior concerning the surfaces in the signal path superimposes their own behavior onto the subsequent signal. Additionally, the sound pressure level will be decreasing according to distance with respect to the primary source. This distance will escalate with each number of reflections, wherefore the later signals become increasingly damped in the upper frequency range from air damping. Even so, each single reflection radiates no other signal like the pure, dry audio signal of the primary sound source.
As far as we would be able to determine, the reflective behavior of the recording room and the alignment and position of the primary acoustic sources therein, including its directional radiation pattern, would possibly define all positions and sound pressure levels of the arising mirror sound sources.
This data-set would completely describe the spatial sonic field pertaining to the sound source in the recording room. By means of that stored data, in theory, we would be able to recreate the complete spatial sonic field in a reflection free volume, or, for example, in a snowy meadow. For this purpose, we have to play through a huge number of single loudspeakers the dry recorded voice of the tenor, for example. Any one of the loudspeakers would represent one of the mirror source positions in the recording room. The signal content of each of those sources, from the signal of the primary sound source, would be filtered by the reflective behavior of the surfaces in its signal path. At increasing distances in regard to the primary source, the number of loudspeakers will escalate exponentially because each of the mirror sources is causing a number of secondary mirror sources. In principle, an infinite number of loudspeakers would be needed. But, let's say within a 3000 foot distance from the real audio source, which is equaled more as a 2, 5 seconds run-time across the chilly air of the snowy meadow, weakened from 20 or 30 reflection losses and air damping according to the long distance, the signal level falls below the audible -60 dB limit, which is marking the reverberation time. Thus, we can dispense of all further loudspeakers.
Nevertheless, such an experiment would not be feasible in practice due to the fact that we would need an individual group of loudspeakers for each of the primary sources in the recording room due to the different positions at which the mirror source would arise. Moreover, all loudspeakers must change their position in case the tenor, for example, takes a step across the stage.
On the other hand, any movement of the listener across the snowy meadow would cause the same changes in perception as, accordingly, changes of the listener position in the recording room. The tenor's voice will be louder and drier if we are to move towards the loudspeaker which is faking its primary sound wave. The first reflections considerably change their incident angle and sound pressure level, while later reverberation from the more distanced loudspeakers is hardly changed in level and direction.
The superposition of the direct wave front and early strong reflections will cause the same comb filter effects as accrue in the recording room at the same position. Even the initial time delay gap will deliver the correct value for authentic perception of the source distance.
The concrete sound source positions will engender correct Doppler effects and parallax changes if the listener moves across the snowy meadow. These deliberations show, in principle, that its possible to configure a virtual copy of the genuine sonic field. All we need is to leave the worn-out conventional ways. The technical realization of the described approaches would have been possible long ago. The following chapters will describe the way this can be achieved.
