Phantom sources and spatial accuracy

"Tonal accuracy is the best that can be hoped for in a traditional audio system; true spatial accuracy will never happen. Audio products should come bearing this disclaimer:


Because of that statement of Barry Wills in Audio 08/1994 some audio purists going protested possibly. Today amplifiers hardly produce distortion and some expensive loudspeakers produce straight response from subsonic to ultrasonic in the anechoic room. What should be the reason hence for such cheeky allegation? Let's search the problems in the principle of phantom source imaging and illuminate the facts:

Channel Separation

All channel based audio reproduction procedures, such as stereo and surround, rely upon phantom source based loudspeaker reproduction. We perceive such phantom sources between the speakers in order of level and time relations.

Unfortunately, such phantom sources don't show really the behaviour of real acoustic sources, as it seems on the face of it. We cannot to bear up our ear on it, unlike the real source the phantom source migrates with the listener position. Moreover, in fact we don't hear one audio source; we hear two audio sources. One of the main Problems of the phantom source imaging caused by the disadvantage, the wave fronts of the left speaker hits also the right ear and converse. That's the same for the natural sound source, but during phantom source playback that crosstalk happens twice, during the recording process between the main microphones and once again during the reproduction process.

That exalts the Interaural Cross-Correlation Coefficient (IACC) value, one of the most important factors regarding our spatial perception. We feel the sound field as spatial, if the difference between both ear signals as far as possible. While the signals utterly different, the IACC is zero. If no difference between, as by mono headphone reproduction, the IACC is one. Natural point sources in free environments reach its maximum IACC value of app. 0.3 under 55 degree outside the median. But if we place the loudspeakers in so far-flung, the sound field would break apart. For phantom source Loudspeakers in +- 30 degree positions is 0.6 a formidable value. That's the end of all spatial accuracy. Sound engineers have no other choice as to go over the top for spatial distribution of the sound sources in anticipation of relic of it remains during the loudspeaker reproduction process. The stage becomes splay from one speaker to the other, no possibility remain for radiate early strong reflections from outside stage positions, what arrange the acoustic attractions during the real performance.

Comb filter effects

Something to the effect that some people doesn't receive an impression of spatiality from an perspective drawing, just as well some people doesn't able to build an phantom source perception. That isn't astonishing, the sum of both loudspeaker signals are fairly different regarding one single sound source.

The summary signal of two microphones, placed on the listener position in the recording room in seven inches ear distance apart would be utterly linear for a central sound source. But if the source nearby the loudspeaker positions on + or - 30 degree the sum signal would shown a strong comb filter response with notches by 2.960, 7.040, 11.180 and 15.200 Hz. The reverse box fulfilling those notches by its mirrored contribution for the central phantom source, but our brain doesn't work so simply as a linear addition of two voltage signals. Something is different in the perception and some people simply doesn't able to sum up the signals to the phantom source perception.

Moreover all symmetry is disturbed, as far as we only turn our head a little. The complex filter function of our outer ear, head sharp and body sound deflection adds by the described comb filter functions to an unmanageable diversity of notches and overemphasis, far beyond the narrow limits, which we would claim even by middle class loudspeakers.

Beyond, from the loudspeaker box direction, by which also the center positioned phantom sound source wave front hits our ear now, its frequency response utterly different regarding the direction of a real frontal source. Excessive sensibility on the upper midrange caused apparent elevation of the center phantom source, while other frequency ranges less prominent in that direction.

Moreover two discrete sound sources causing utterly different reflection pattern in the playback room. The total difference in frequency response dependent of a lot of facts in the complex outer ear- head shape- body- box placement- box radiation and room mirror source structure. Which effects cancel each other or add together is hardly predictable. Mostly remain only the try and error method for satisfying results.

One- dimensionality

As like as a perspective drawing trigger spatial impression only in certain limits, likewise, the phantom source is shortened able to create the perception, a sound source would be more or less distant. Just like all ink remain inside the drawing, all phantom sources remain on the line between the loudspeakers.

That's comprehensible, if we are moving in the listening area. For example, if the violin in the concert hall straight in front of the timbale, we hear both instruments from the same direction. However, the real violin occurred left from timbal, if we move to the right wall at the concert hall. Both phantom sources between the loudspeakers, conversely, remain at her common starting point.

On that reason seems not realistic, to expect the spatial deepness of the real event from any phantom source based reproduction. We have only limited possibilities for fault the brain, regarding the phantom source distance.

Firstly, we can change the levels, of course. We are connecting higher levels in perception at decreasingly distance.
At the second, the portion of direct wave compared with the diffuse field level of reflections, is eminently important. In case of the sound source came closely the listener, the direct wave sound pressure increase accordingly the 1/r function. The diffuse field level remained unchanged, however.

Unfortunately, solitary loudspeakers cannot produce such relations, because its own radiation became the overlay from the playback room reflections. Thus, we cannot depict any source more closely the listener as the loudspeakers itself. That's the reason, why we are only surrounded in surround, not include in the sonic field.
The third reason for the disturbing source distance in the home cinema is the initial time delay gap (ITDG). We cannot avoid short sound detours in small rooms, in large rooms the first reflection arrives later.

Regarding very close sound sources, a further sign is important. In case of the signal in one of the ears is much louder as at another, the only possible reason, the sound source is very closely. The Headphones can be reproducing this, but not the loudspeakers.

Besides wearing headphones, only way out seems to be near field playback. Two possibilities exist: On first we can place the loudspeakers closely the listeners. That sounds pretty good but diminished the listening area and is hardly practicable in living rooms. Alternatively, more aligned radiation improve the conditions. The mirror sources are supply at a reduced level in that situation. But for midrange that need a large radiating surface, because the wavelengths are above a foot. ELS is a good choice in that matter. Nevertheless, a directive radiation independent frequency seems not feasible. However, the playback room acoustics lose of influence. At such conditions the phantom source reproduction provides sometimes surprisingly good results. However, the best perspective drawing isn't really an accurate spatial representation. We would need a hologram.  

Playback room acoustics

The most audible difference during conventionally playback procedures caused by the fact, the playback room acoustics strongly differs regarding the recording room acoustics. Phantom source based reproduction hasn't the ability for restore the genuine wave front. All sound pressure originates from some single points; hence utterly different wave propagation is caused:


In the main tone range the single speakers radiate hardly directed. Its wave fronts hits the playback room surfaces in utterly different manner as the wave fronts of the sound source in the recording room do. The first reflections arise in the mostly smaller playback room inside the psychoacoustic adverse range below two milliseconds. Yet the recorded early reflections doesn't differ from source direction, at least for all panpot mixed recordings. Inevitably the playback room acoustics determinate the reproduction by that reason, the spatial impression of the record room acoustics never will happen for phantom source reproduction.

All, what we can do for the spatial impression remain producing the appropriate reverberation of the recording room. Though, the reverberation cannot deliver any information's regarding the source position. The core of all spatial impression is the spatial distribution of the direct wave and of the mirror sources which originate the first strong reflections in the recording room. Its correct positions deliver the distance information regarding the source and regarding all reflection surfaces, what delivers the spatial impression. We are fare away by the phantom source reproduction; the presented image is really less then lifelike.

The described problems seem only avoidable by virtual acoustic sources, which described on the next sites. Such sources would be able for deliver a “holophonic” picture.


last update 2013-09-27