Transformation of the early sound field
engine translation of the german patent description only for information purposes, no warranty for agreement with the original text!
Summary
01.The procedure reconstructs the direct wave of an acoustic source and its first strong reflections in the recording room temporally correctly and to a large extent faithful in amplitude and direction to a rendition area, even if the areas differ clearly in their dimension and in their reflection characteristics. Thus comes into being an acoustic perception at the listener, which is hardly to be differentiated of the primary sound field.
02. The acoustic differences can become balanced, because transmit the relevant parameters of recording and rendition area before the actual audio signal transmission startet in the common system. Therefore becomes possible to reconstruct the sound field in its spatial structure according to the Huygen´s‘ Principle of the wave field synthesis, alone by a in front of the listener set up loudspeaker matrix, correct in time and widely also correct in direction.
03. This matrix does not reduce the sound field, as with WFS applications usual, on the plane level of the listener. It turn into reconstruct in all three space dimensions.
04.The system is not bound to a number of channels and compatible to the well-known procedures. Also during their rendition a more realistic sound field with clearly extended sweet spot can be generated.
05. The reproduced recording room can be both, more largely or smaller as the playback room.
Transformation of the early sound field
06. The audio rendition with loudspeakers was improved clearly in the last years. Computer-assisted development and manufacturing secure now even in the consumer realm distinguished technical parameters. And nevertheless: Also the best facilities fails pitifully until today, compared with the original sound field.
07. The cause for this cannot be justified any more longer in the inadequacies of the transducers themselves. Obviously the reason lay in errors inside the transmission system, wherefore we do not correct upon the reproduction of the audio event.
State of the art
08. Exactly as in the original sound field first the direct wave reaches us from the loudspeakers. On means of the run time differences between our ears we can determine the horizontal direction of the acoustic source within the range of approx. 160 to 3600 cycles per second immediately and very precisely.
09. It is important that this perception is not disturbed too early by reflections. If these arrive in former times as 2 ms after the primary sound, they not only prevent the precise detection, but also smear the transients of the vibrations and change thereby the sound impression completely.
10. Fig. 1 represents this problem in principle. It arises with conventional loudspeakers (1b) above all if they are set up closely at the walls of the rendition area (1c). Beyond these walls they produce strong mirror acoustic sources (1d), because they radiate within the basic tone range relatively unidirectional. With the listener (1e) early reflections (1g) with the direct wave (1f), produced by it, overlay. Therefore the same loudspeaker in an area sounds well, in the other one sound bad, such as like its special radiation characteristic apt with the room acoustic.
11 All attempts to then adjust the developing comb filter effects in the frequency response bring new problems. The impulse response often still more indifferent and prevented now an accurate detection of the point acoustic source completely.
12. If however first sound-strong reflections arrive in the time window between approx. 5 and 50 milliseconds after the direct wave, they do not falsify the sound impression any longer. In the opposite, they are assigned still directly to the primary sound and increase so the subjectively felt volume. If they arrive in addition from clearly different direction than the primary wave, they provide the transparent, open and spatial sound field, as we know it from material acoustic sources. Conventional loudspeakers cannot reproduce that, because they cannot represent this direction difference.
13. Straight that is the actual core of each room acoustics. Not only in the horizontal level of the listener, to which all conventional loudspeaker rendition procedure is reduced. An indispensable condition for a really authentic reproduction of the spatial perception is it that first strong reflections restored in all levels correct in time and direction.
14. We can determine the elevation of an acoustic source particularly over reflections at our pinnas and at the trunk. These detections are not as accurate as the run time detection. Besides they are individually very different, because the allocations are learned particularly over optical linkages.
15. Within the overtone range we must rely also in the horizontal level on this listening experience, because lie several wavelengths within the ear distance and are not clear the run time detection therefore any longer. Therefore we locate also here the acoustic sources amplitude-referred. We learned to connect diffraction and shading effects at the head with a direction allocation. We do not evaluate the frequency range over approx. 8 kHz.
16. How important the time-correct arrival of first strong reflections is below this frequency through, one can hear in many large concert halls around the world. If first strong reflections arrive too late, more as approx. 60 ms after the primary wave on the listener, language is smeared with difficulty intelligibility and music sounds indifferently.
17. After approximately 100 ms is the direction of the acoustic waves no longer of importance. We assign it then to the reverberation, which supplies important information for the condition of the area, whose spatial detection plays however a subordinated role.
18. Therefore it is little meaningful to limit the spacial representation too much to the reproduction of the reverberation. The representation remains improbable, as long as the loudspeakers impress first sound-strong reflections of the rendition area onto the audio rendition.
19. From the shown problem it becomes clear that to the natural reproduction of a primary sound field demands must be made, which conventional loudspeaker rendition cannot fulfil satisfying: The spatial progressive rate of the sound field get left completely to the interaction of loudspeaker and monitor room. Because of their to a large extent unidirectional radiation within the basic tone range conventional loudspeaker arrangements cannot ensure, that in relation to the direct sound portion comes prematurely arriving reflections of the walls of the rendition area is sufficiently inconsiderable.
20 Therefore they reproduce not acoustic conditions at the recording place, but impress to the sound event the acoustic behaviour of the rendition area, which must lead inevitably too completely falsified results. The transient behaviour of the recording room, thus time and direction-referred reproduction of its first sound-strong reflections, they can represent not nearly.
21. Such a purposeful ranking of the radiation can cause only one loudspeaker arrangement, whose expansion extends at least over the half wavelengths of the radiated frequency. It must be able to align the wave fronts in the horizontal and in the vertical level steered. Like among other things in " Sensivity ton of sound SOURCE elevation in Nontropic Auditory Cortex" of left XU, Shigeto Furkawa, John. C Middlebrooks into the journal OF Neurophysiology of volume of 80 No. August 1998, pp. 882-894” was represented, the accurate reproduction is particularly in the visual field important, because we notice errors here fastest.
22. Conventional loudspeaker boxes are thus already unsuitable for an authentic reproduction, because they are too small, in order to make a purposeful directional radiation possible in the basic tone range. With coherently headed for loudspeakers such a directive effect can be achieved only with long sound lines, then in addition, only in its longitudinal axis and not controllable. The procedure of the wave field synthesis offers by far better possibilities, to which there is straight in the last years world-wide intensive research work. The results of the European research are in summary represented on the Internet side http://www.emt.iis.fhg.de/projects/carrouso/ . The single emitters of a group of loudspeakers are headed for differentiated in such a way for the WFS by a computer synthesis individually and that they generate a cut-out of a primary wave front in accordance with the Huygens` principle again. Their diaphragms are expenditure-steered in addition accurately at the time, at which also the wave front of the material acoustic source would go through its space point.
23. In contrast to single loudspeakers, which can generate only unstable phantom acoustic sources, so virtual acoustic sources are produced, which do not differ in their stable locatableness for the listener from the primary wave front.
24. With procedures can generated convex and concave wave fronts, so that the restriction of conventional loudspeaker rendition, during which an acoustic source cannot be closer at the listener, as the loudspeakers, is waived.
25. Also parallel wave fronts can be produced, as the starting point of the virtual source is shifted infinitely far behind the loudspeaker front, without lowering the amplitude accordingly.
26. The expenditure for this synthesis is large, because the loudspeakers may not be from each other installed because of aliasing otherwise arising effects with frequency response falsifications at individual listener places in to large distance. Aside this acoustic source can be only represented, if it is seen from the listener within the loudspeaker field. Because of this truncation the loudspeakers are set up around the listener in circular or rectangular arrangement, in order to be able to represent acoustic sources from all directions. Therefore it is today usual to reduce the procedure to horizontal emitter lines. That involves however significant restrictions: The reproduction of the elevation of acoustic sources is not possible any longer. This reduction of the procedure on the infinite level of the listener becomes again and again, for example also in protocol of the 7. Int. Conference OF digital audio Effects (DAFx'04),Neaple 2004, page 127, Abs.3.3, publishes in h ttp://www.irt.de/wittek/hauptmikrofon/theile/Theile_DAFX.PDF ,
as the most significant disadvantage of the loudspeaker arrays criticizes. In the vertical level the horizontal lines have the same unidirectional radiation behaviour as individual boxes. Fig.2 represents this error schematically. With this comparison of the propagation of the acoustic wave of a material source with the rendition of such WFS line it becomes clear that the spherical waves emitted of the speaker line reproduce the wave front in the horizontal level very accurately, however in the vertical level very strong deviations developed (2d;2g).
27. Over the reflection of the ceiling these wave fronts in our dwelling example arrive already 3.34 milliseconds after the direct wave with the listener. On the other hand this reflection arrives in the acoustically favourable recording room of our example only 22.7 ms after it, thus in the psycho- acoustics favourable time window.
28. Exactly like conventional loudspeakers the WFS line impressed the unfavourable acoustics of the rendition area to the sound event thus in the elevation level. The actual wave front of the recording room is not represented in this level at all or only with a signal from direction of the acoustic source. The open sound impression of the recording room is lost completely.
29. Besides really parallel wave fronts cannot be generated by such a sound line. Always develop cylinder waves, whose sound pressure with doubling of the distance decreases by 3 dB. Also this restriction is admits and in the same section of the DAFx'04 of minutes mentioned. In the reversal conclusion it results from the fact that the level of the sound line increases clearly, if the listener approximates closely, which must lead inevitably to the restriction of the Sweet Spots.
30. The sound pressure of a parallel wave, generated in two dimensions, remains direct against it completely constantly, all the same whether the listener stands 10 meters far away, or before the emitter surface.
31. Practically such was developed a two-dimensional WFS emitter surface already 1997 at the University of Tokyo , published for example in http://www.irt.de/wittek/hauptmikrofon/wittek_wfs_litreview.pdf. Thus an idea was carried out, which experts had already in the thirties, which considered it realizable at that time however never - the arrangement carries out the principle of the acoustic curtain.
32. The loudspeaker diaphragms reproduce thereby the sound pressure of a directly arranged surface in the recording room, so as if on a side of a wall microphones would be arranged, which in each case head for the associated loudspeaker on the other side of the wall. Within this window the rendition is perfect. However the restriction applies also, that the shown virtual acoustic source must be placed seen from the listener within the loudspeaker arrangement.
33. Nevertheless the acoustic impression of the recording room cannot be produced. It obtains the restriction on the cut-out rather the impression that one would in-hear by an open window into the opera house, why the listener does not lose however the acoustic perception nevertheless that he stands still in the free one.
34. The usual view of a transmission system begins with the microphone and ends with the loudspeaker. The acoustics of the monitor room and the position of the listener have at the most still influence on the adjusted total frequency response.
35. But the rendition area has the most relevant influence on the quality of the total rendition. The largest portion of signal distortions in the entire transmission chain develops, after the loudspeakers did their work. 36. Because however the acoustic parameters of this transfer element are included with no well-known output system differentiated into the transmission chain, there is also no possibility of correcting its errors purposefully.
37. It became clear that the principle of the wave field synthesis can be the only starting point with the attempt to transform the sound field of a recording room into a completely different rendition area. The principle may not be reduced however to a single speaker line, if we want to produce a three-dimensional sound field.
Setting of tasks
38. Task of the invention is it to indicate a procedure and a device for the transformation of the early sound field, on which the loudspeaker arrangement frontally before the listeners illustrates not only the acoustic sources, but also their first sound-strong reflections in the recording room in all three space dimensions as correctly as possible in a rendition area.
39. This task is solved with the characteristics of the requirement for procedure 1 and/or the characteristics of the requirement for device 8. Favourable arrangements of the procedure are indicated in the claims 2 to 7.
40. The invention is based on the realization the fact that for the illustration of a recording room in a rendition area the first strong back throws of walls, soil and cover are to be considered not only regarding their running times and amplitudes, but also regarding their directions of arrival that such back throws in the rendition area can be produced also by wave reflexions at the walls, the soil and the cover of the rendition area, and that this all with a two-dimensional loudspeaker arrangement arranged before the listener range succeeds, whose loudspeaker signals are prepared according to the principle of the wave field synthesis with additional consideration of physical characteristics of the recording and of the rendition area.
41. In order to be able to illustrate the recording room as perfectly as possible in the rendition area, also all necessary information must be conveyed regarding the characteristics of the recording room in addition to the audio signal.
42. But even the rendition of conventional sound records can be obtained improvement, if the audio source in the acoustic environment of a suitable area is represented, whose data the system from a stored library can call up.
43. Beyond that the invention can be used also in order to represent smaller recording rooms in a larger rendition area (requirement 4). It is even possible at the same way to illustrate in a larger rendition area several small rendition areas in which different rendition signals will deliver (requirement 5). Also the sound can be spatially adjusted changed listener positions (requirement 6)
45. For the realization of a three-dimensional reproduction we must develop a two-dimensional loudspeaker arrangement, which is set up frontally before the listener, possibly hidden behind a screen wall. Their expansion should go however clearly beyond the surface of such screen.
46. Their single emitters should work in a frequency range from approximately 160 Hz to approx. 8 kHz linear and have if possible a diaphragm diameter under 2 inch, in order to be able to produce in the entire work area a homogeneous hemisphere wave. Over 8 kHz two high-quality treble- speakers take over on the left and on the right, because the spatial detection is subordinated in this range. A Subwoofer supplements within the bass range.
47. The distance of the matrix loudspeakers to each other determines the Aliasing frequency. It should not be above 20 cm, to 10 cm would be better, which means however the quadruple expenditure.
48. Also with approx. 20 cm distance is already clearly three-figure their number in usual areas, so that the single emitters must contribute only little to the total sound power. Few Watts of achievement are therefore sufficient for the single output stages, which must be assigned to each loudspeaker, completely. They can be supplied with a common operating voltage; also the digital control can by a common bus be made. Commercial circuits can filter today as De multiplexer up to 192 channels from a usual Kat5 cable from such a bus and they are quite inexpensive thereby.
49. The heading for bus comes, as with the linear WFS systems, separately from the WFS Renderer, an efficient signal processor arrangement, those made of contents and form, thus that up to then finished data of pure audio signal and the positions of the associated virtual acoustic sources, which produces individual heading for signals for the loudspeakers.
50. Thus the hardware is fixed, with which we want to produce the transformed sound field on the rendition side. At first sight this attempt seems, above all offering no prospects, because the loudspeaker arrangement is to be limited to only one wall of this area.
51. The "Truncation" - effect, thus the restriction of the radiation on the range within the loudspeaker arrangement, would make a solution impossible during such an arrangement, as long as we regard it individually as output system. If we see however the rendition area no longer as disturbing factor, whose we must removing with empty egg cardboards or thick black a cloths as possible, completely new possibilities result: The truncation effect is always related to the position of the listener. Goes for example to the right wall, in front beside the matrix, it knows virtual sources within completely different range, until far left behind the matrix hear.
52. An acoustic source positioned there meets naturally also right side panel and of is reflected. Already with, in the home range problem-free realizable, diagonals of 3 to 4 meters the directive effect of the matrix is so pronounced also in the basic tone range that - the Truncation effect is thanks - which are met listeners hardly still directly by the wave front. 53. Functions naturally exactly the same with the left wall, the cover and the floor.
54. From the mirror acoustic sources of the loudspeakers in the rendition area, which disturbed us in such a way so far, in this view completely independently controllable acoustic sources became, which offer us completely new possibilities for the organization of the sound field.
55. In order to use it, must their data admits to be. After the installation of the frontal loudspeaker wall therefore all relevant parameters of the rendition area must be entered into the system.
56. For the later sound field synthesis beside the geometrical data of the rendition area also the reflection factors of the walls, separately indicated for each octave of the rendition range, are seized. In practice that quite simply to be, clicking the respective material in an appropriate computer program is sufficient. The possibility, instead of processing this data based model the impulse response of the areas, is here less suitable.
57. A position of the listener must be specified. We absolutely need a firm point of reference for a transformation. Only at the ears of the listener at this place our view of system may end. Here the wave fronts must arrive if possible in the same temporal and spatial progressive rate, as at one point in the recording room, at which a virtual listener is. 58. This fixed rendition place is not to become the only place in the area, on which a high-quality reproduction is possible. It must be rather the centre of large a Sweet as possible Spots for the rendition.
59. From the recording side we need the audio signals of the individual acoustic sources for the transformation of first sound-strong reflections. For each acoustic source a separate channel is necessary, whereby single sources can be summarized, if they spatially closely neighbouring are. Transformation must enter also its momentary position, its polar directional characteristic and its momentary adjustment. The polar directional characteristic can be assigned to the static data, because it does not change during the period of the transmission. These data are conveyed before the beginning of the actual audio transmission.
60. They do not have again to be transferred each mark, if them admit on the rendition side are. Directional characteristics of all possible acoustic sources for example, could be put down into libraries on the rendition side and assign by appropriate codes of the respective acoustic source.
61. The momentary position and the momentary adjustment of the acoustic source must be assigned to the dynamic data, which are updated during the transmission in short time intervals. Finally it makes different sound impression clear for the listener, if Caruso turns unexpectedly and hits the sound against the decoration.
62. As on the rendition side also a position and an adjustment of the listener in the recording room must be specified for the transformation. That can be the acoustically most favourable place. Together with a picture transmission it is however more meaningful, if this position and adjustment correspond with the picture transmission.
63. These listener positions are set for the sound field reconstruction to the coincidence. With change of the coordinates of the listener in the recording room change then distance, angle of incidence and level of all wave fronts with this listener both for the direct waves of the individual acoustic sources and for all acoustic waves outgoing from the mirror acoustic sources of the area. Accordingly also the levels and running times reconstructed during the rendition change. Thus the acoustic perspective of the listener can be shifted on any point in the recording room by the change of only one data block. And if the listener wanted, he could do with its remote maintenance also.
64. The positions of the mirror acoustic sources, which a sound event in the recording room produces by reflections at its delimitation surfaces, play a key role with the transformation of the sound field. Succeeds letting with the rendition-lateral synthesis virtual acoustic sources proceed apparently from these positions then they produce the same impulse response, which effect affect the listener in the recording room at the listener position in the rendition area.
65. Because also the direct wave with the acoustic curtain can be rather perfectly copied, such a completely realistic image of the sound field develops, as far as also amplitudes and angles of incidence of these wave fronts agree. Only with the frontal loudspeaker arrangement we cannot generate these mirror acoustic sources however at all space positions. 66. The invention is more near described in the following one on the basis the design figures. In particular is stated, as succeeds, with the independently controllable mirror acoustic sources of the loudspeaker matrix in the rendition area the recording room very similar a sound field producing.
It shows:
67. Fig. 1 a schematic representation of the incorrect rendition of conventional loudspeaker boxes, during which the mirror acoustic sources of the loudspeakers of the reproduction within the basic clay/tone range always impress the acoustics of the rendition area.
68. Fig. 2 a schematic representation of the rendition according to the principle of the wave field synthesis with horizontal loudspeaker arrays with incorrectly reproduced sound-strong reflections in the elevation level. 69. Fig. 3 a schematic representation of the shift according to invention of the position of the mirror acoustic source of the cover reflection of the recording room into the represent able range within the mirror acoustic source of the loudspeaker arrangement in the rendition area.
70. Fig. 4 a schematic representation of the positioning of the virtual source for the cover reflection of the recording room by reflection of the position of the shifted source after Fig. 3f to that, which system admitted, coordinates of the cover of the rendition area.
71. Fig. 5 a schematic representation of the shift of the virtual source for the reflection of the rear wall of the recording room in the represent able range and the following double reflection of this position to that, which system admitted, coordinates of the rear wall and the cover of the rendition area.
72. Of it the figures 1 and 2 were already described in connection with the conditions of the technology. In the figures 3 to 5 now details of the invention are represented.
73. In Fig. 3 the outside right parallelepiped (3a) represents again the recording room. The smaller rendition area of the example (3b) is used in such a way for the computation of the starting points of the virtual acoustic sources into this hall that the positions (3í), specified for our transmission, are congruent the listener in the hall and at home. At this common position we want to produce agreeing signals.
74. The starting point of the direct wave (3c) lies in the range of the loudspeaker field (3g), which represents an acoustic curtain for this source. It is therefore accurately represented. However the position of the reflection of the primary acoustic source is appropriate for the independently controllable mirror acoustic source (3h) in the recording room (3d) outside of the represent able range, far from the range of the loudspeaker arrangement, but nevertheless not too far of the range (é). 75. Thus we must shift it in a first step into this range, in order to be able to represent it as well as possible. That happens on a circular path around our fixed point of listener (3í). This shift must take place however sufficiently far into this range inside (3f), because the listener would be otherwise in the boundary region of the respective Sweet Spots.
76. Important is above all that the distance to the listener and thus the time remained unchanged, to which the wave with the listener arrives. 77. Fig. 4 represents the second step of the rule, according to which the starting position of the virtual mirror acoustic sources is specified. Because within the range over the living room cover no material loudspeakers exist, we must reflect the fixed position now at the ceiling. Since the dimensions of the rendition area to the system admits are, the final position of our virtual acoustic source (4b) can be computed according to the simple rules of geometry.
78. Thus the WFS Renderer has all information, which it needs around the single loudspeakers to head for. It must retard the audio signal associated to the mirror acoustic source only around the computed sound running time from their starting point to the respective loudspeaker.
79. Naturally a part of this wave front is converted from the ceiling into warmth, why it never arrives with the listener. That Happens in the recording room in the same way, however. If the reflection factors are alike in both areas, the amplitude of our virtual source of mirror does not have to be changed.
80 If the materials of both covers are different, their frequency response can be corrected easily in the individual octaves around the difference of the reflection factors. For example if our recording room has a reflection factor of 0,9, our living room cover however only 0.7 with a certain frequency, then the level of our virtual mirror acoustic source must be raised around 20 log (0.9/0.7), thus around 2,2 decibel. Additionally this level is reduced after the laws of the ball wave propagation and by the amount of the airborne sound insulation for its virtual part of the distance up to the respective loudspeaker.
81. The polar directional characteristic of the associated acoustic source enters synthesis, as the amplitude of the virtual mirror acoustic source is reduced periodically by the amount of the level sinking in its polar directivity pattern in the associated solid angle.
82. After the same principle then the starting points of the virtual sources for the soil reflection move from down upward and for the side panels from left to the right and in reverse.
83. Fig. 5 represents the reproduction of reflections of the rear wall of the recording room schematically. While the reflection of the front wall can be produced directly within the matrix, the representation of the reflection of the rear wall (5d) without rear loudspeakers is more difficult. For these the wave front of the mirror acoustic source is reflected doubly, only at the rear wall (5f) and then at the cover (5g). Because of "truncation" - effect hardly meets it thereby the listener on its way there. The reflection factors of the rear walls are charged again, the absorption the rendition space cover becomes balanced by appropriate periodical level rise.
84. So far possible, the rear wall of the rendition area should be sound hard in the upper range. Among them a sofa, an open book board or a wall carpet could absorb the direct wave.
85. If our listener leaves at home his point of reference, then its distance to the virtual acoustic source and to their entire mirror acoustic sources changes. The running times to them and thus the impulse response at its current space position change in the same way, as them would change in the recording room for our virtual listener, if he leaves his place. If it sits down at home thus a place to the right, then it has the same sound impression as the right neighbour of our virtual listener in the recording room. The relative importance of this change of location was not transformed; therefore our listener cannot go in the sound field any more than five places to the right, because he runs then against a wall. 86. And still it has two restriction: It may within the ranges, in which wave fronts of the mirror acoustic sources on the way to the room walls are, thus directly on the right of or left beside the loudspeaker matrix, not with a high-quality rendition counts. Here wave fronts reach it at the wrong time from wrong directions, like that as we it are accustomed by conventional loudspeaker rendition.
87. And it cannot hear the direct acoustic source, if it leaves the truncation range of the acoustic curtain. This restriction is more important, because it limits the Sweet Spot for the rendition laterally, if a virtual source is closely beside its edge.
88. Nevertheless if an acoustic source is to be far laterally represented, then we can avail ourselves as compromise settlement again of our independently controllable, virtual mirror acoustic sources of the matrix: So can be positioned for example very far on the right of a virtual acoustic source, as in the described procedure its starting point at the right wall is reflected to the left.
89. The amplitude falsification by the moving inflection is compensated with the inverse transfer function of this wall. However the right side panel reflection of this source can be only generated from direction of the source. But in the correct distance, that will thus be noticeable the error only the experienced listener.
90. Further disadvantage of this method is it that the relative angle of the acoustic waves is larger to the axle of the matrix. Thus the Aliasing frequency shifts downward with given single emitter distance. A signal distortion, which is smaller however with a two-dimensional matrix than with the LINE array: If the single emitters are hexagonal arranged not in the raster, but, the emitters increase the horizontal Aliasing frequency by the factor 1, 4.
91. Depending upon vertical adjustment of the wave front they are not in the angle of the zero in phase, if a wave front goes through the matrix laterally. Thus the amplitude in this angle sinks clearly, but does not decrease/go back not up to zero. The Aliasing effect produces a comb filter effect in the frequency response at the listener place, for which our ear is only little sensitive, as long as there are no expressed zeros.
92. In the rendition area the described procedure created a sound field, whose impulse response in the important time period up to approximately 100 ms agrees after the direct wave with the impulse response of the recording room. Only some angles of incidence of reflections changed, if the dimensions of the areas are very different or if their starting points are far laterally from the listener or directly over him. Within this, in the Psycho-acoustic as "cone OF confusion" designated range is however anyway very imprecise our detection.
93. The response can be transferred then again with the primary sound signal, since its direction of arrival without meaning is. It reaches us then also from the mirror acoustic sources, which its irregular distribution in the recording room very close.
94. There should be a forbidden range for all virtual acoustic sources: If they are placed directly in the level of the loudspeaker arrangement, then an individual loudspeaker must produce nearly the entire sound power. Therefore all individual components of the matrix would have to be able to produce this achievement distortion-free. The plant became then substantially more complex. Therefore these virtual sources should be shifted in principle or two loudspeaker distances to the rear, so that several loudspeakers co-operate and so also with less efficient single emitters good total dynamics can be obtained.
95. In the system is not fixed actually, as many primary signals will transfer at the same time. Already an individual source, thus a mono channel can be represented very authentically, even if during the reproduction all additional information from the photograph side is present. The characteristics of the rendition side are anyway in the system stored. The dynamic additional information must be to the respective source related, become the space data together used.
96. With the today's digital transmission methods the transmission of this additional information is problem-free possible. If a transmission system does not permit however the simultaneous transmission, they can be conveyed and stored also separately, for example from the Internet transmission.
97. If a recording is transferred, then also the dynamic complementary data are well-known for the entire running time already before beginning of the transmission. They can be conveyed and stored thus before the start of the rendition. Over one time they can LINE is linked then easily with the rendition.
98. Then thus our virtual Caruso can move by the Scala of Milano and turn as desired its head, even if its record comes only from an old mono admission.
99. With the described procedure, with according to produced program material, a very authentic reproduction of a sound event succeeds. Few separate channels are sufficient already for many applications. Usually many individual acoustic sources are not at the same time active. Their rendition will not arouse the impression that only few mono channels would transfer, because each primary acoustic source produces its own, three-dimensional sound field.
100. Beyond the number of the channels sound events can be represented if they come from closely neighbouring sources and use a common channel, or if they not at the same time with another acoustic source active are. The complementary data change then temporarily on the other source.
101. As previously mentioned in foreseeable time only few transmissions directly for the described procedure produced to be. It is possible so long that conventional records also, supplemental by additional information, can be clearly more realistically reproduced.
102. How described further above, of the early sound field can be represented an old mono admission in a completely new acoustic dimension by the transformation. In this way also the centre channel of a surround rendition, in correspondence for associated picture information, could be clearly revalued.
103. For the left and right channels, exactly as with the stereo rendition, dependent on the category an acoustically suitable, virtual recording room from the library is selected. In this area again a listener is placed, before whom the left and right loudspeaker boxes are set up as virtual acoustic sources, virtual panning Spots so mentioned. 104. The remainder of the reproduction does not differ from the one natural acoustic source at this position. Then still the grossest errors now of the "conventional" can be electronically adjusted, virtual loudspeaker rendition in a virtual area.
105. Similar applies to, the rear loudspeaker generated over the covering reflection. Thus altogether a procedure is available for the rendition of conventional surround productions, which is not only compatible, but that the rendition significantly improves.
106. Develops a substantially increased Sweet Spot, which is expanded since the entire rendition area nearly, because the angle- changes and the relative changes in distance are clearly smaller to that far distant "boxes", than with material loudspeakers in the rendition area, if the listener moves therein.
107. In the past description always the case was treated that a larger recording room in a smaller rendition area is represented, what the rule for home rendition surely is. By the loudspeaker arrangement early reflections are copied, which are retarded radiated in relation to the direct wave.
108. But also the reverse case is without modifications at the system possible, even the described algorithms for the positioning of the virtual mirror acoustic sources remains the same.
109. The Samples of the digitized audio channels will in Wfs systems anyway for some hundred milliseconds buffered, in order to be able to select it for the individual loudspeakers sequentially. Thus it plays also no role whether first the direct wave and then reflections are shown, or whether the matrix already sends reflections on its way, before the direct wave is produced. In the result an area is then represented, which is smaller than the output area.
110. That is not only interesting, over for example voices from the interior of a passenger car in the living room to crimping. Completely new perspectives open themselves thereby in the acoustic irradiation technology.
111. Into a too large hall, with which long detours of first reflections produce a psychoacoustic unfavourable sound field, a smaller can become "in-designed", acoustically favourable virtual area. Everything else takes place subsequent to the described procedure exactly. However the public surface must fit in it still this smaller virtual hall, outside of its borders comes to acoustic confusions.
112. Nothing speaks against it that the recording room is also at the same time the rendition area. Only to the signal processing of the system then higher demands must be made, they would have approximately in real time to work. If a speaker speaks into a microphone, to high deceleration time lets him come otherwise into the stutter.
113. Preference/advantage of such a system is it, which the sound of the wide matrix can be aligned very purposefully. The direct wave can be arranged during suitable matrix arrangement purposefully diagonally from above on the public; there it to the predominant part was absorbed. Also first sound-strong reflections can be aligned over smooth delimitation surfaces purposefully.
114. A very interesting application possibility opens: The sound can be aligned only to a sub range of the public surface. For example on the seats of the right side, where sit the foreign guests. Left the national language would be to be heard.
115. Not without cross modulation of the other range. Its volume was essentially determined by the reflection factor of the public surface. In a hall of middle reverberation time about 10 railways level difference would be to be expected with closely occupied public surface with polarizing inheritance frame, with heavy upholstered chairs approx. 15 railways would be realistic.
116. Because of the well-known party of effect then each listener would concentrate on his signal. Thus such values would be sufficient and the listeners the quieter foreign language signals would surely accept, if they could do for it without their annoying headphones.
117. At constantly changing public, for example in exhibitions and on fairs, the headphones bring substantial hygienic problems. The sound of the described matrix focused on an individual place could solve the problem, also with inexpensive single emitters. The sound field could be adjusted even the moved listener. With the well-known acrylic glass domes for this focusing that is not possible.
118. With practically developed plants with LINE it turned out arrays that the influence of artefacts, which result from the partial interruption of the sound lines is relatively small. Where a door is, the line can be interrupted relatively broadly, without it becomes subjectively perceptible. 119. Same will apply also to the matrix. As simple application it seems therefore even still meaningful to use the procedure to the frontal acoustic baffle of individually set up loudspeaker boxes.
120. The reproduction of such an arrangement will stay naturally far behind the naturalness of the sound field of the matrix, but with few single emitters the errors of conventional loudspeaker boxes, at least in the higher frequency range, described above, will already be less clear. And already the attempt to generate at least the strongest reflections at all will cause a clearly improved subjective perception. Even if it does not generate in the described way complex, but only by some standard Setups for firm deceleration time were produced. With such simple additional retarded signals even a usual d'Appolito loudspeaker arrangement can produce a surprising spatialness.
Reference symbol list
1a acoustic source
1b conventional loudspeakers which can be represented ç rendition area 1d mirror acoustic sources of the loudspeakers in the rendition area
1e listener
1f first wave front
1g mirror acoustic source of cover reflection
3g loudspeaker arrangement
3h mirror acoustic sources of the loudspeaker arrangement in the rendition area, shifted by the mirror acoustic sources of the loudspeakers outgoing, sound-strong reflections of the rendition area à recording room
2b rendition area
2c acoustic source
2d of the acoustic source outgoing wave front
2e horizontal loudspeaker array to the wave field synthesis
2f listener position in admission and rendition area
2g of the primary wave front deviating radiation of the loudspeaker array in the elevation level
2h detour of cover reflection in the recording room ì detour of cover reflection in the rendition area
3a recording room
3b rendition area
3c acoustic source in the recording room
3d mirror acoustic source of cover reflection in the recording room
3e represent able range for the mirror acoustic source of cover reflection
3f into the represent able range,
4a Rendition area
4b at the cover of the rendition area reflected position of the virtual acoustic source 3f
4c loudspeaker arrangement
4d listener position, at that the impulse response of the rendition with the impulse response at the listener position í in the recording room agrees
5a recording room
5b rendition area
5c acoustic source
5d mirror acoustic source of the rear wall of the recording room
5e mirror acoustic source of the rear wall of the recording room
5f at the rear wall of the rendition area reflected position of the mirror acoustic source of the rear wall of the recording room
5g at the cover of the rendition area reflected position that, shifted into the represent able range, the mirror acoustic source of the rear wall of the recording room as starting point of the virtual acoustic source with the wave field synthesis
5h loudspeaker arrangement
5i depictable range for the reflection of the rear wall of the recording room of
Patent claims
1. Procedure regarding the transformation of the early sound field by means of a two dimensional loudspeaker set for the wave field synthesis (WFS) in horizontal and vertical level, arranged before a listener position in a rendition area, for the production of virtual acoustic sources within its for the listener position acoustically perceptible, from the size of the loudspeaker arrangement the limited range,
characterized by the fact,
that further virtual acoustic sources are produced, which outside of this limited range and by mathematical linkage of the geometrical data of material or virtual recording room and the rendition area source positions are generated, at which they are perceptible for the listener mainly after reflection at the delimitation surfaces of the rendition area at the distance and close of the position of the acoustic sources and/or its mirror acoustic sources, in order that primary Acoustic source in the recording room and/or at its delimitation surfaces (first sound-strong reflections) would produce produced , whereby a sound field develops, which perceived according the dimension of the sound field of the recording room.
2. Procedure after patent claim 1, by the fact characterized that the amplitudes of the described virtual acoustic sources are corrected periodically, in particular octaves, then that the difference becomes balanced reflection factors of the delimitation surfaces concerned of admission and rendition area by appropriate correction of the amplitude of the virtual acoustic source.
3. Procedure after patent claim 1, by the fact characterized that those is reproduced the momentary adjustment of an acoustic source in the recording room during the rendition by the fact that the amplitudes of the direct wave and its first sound-strong reflections as a function of the momentary adjustment of the acoustic source which can be represented are changed periodically in the measure, as it gives the polar directional characteristic of the acoustic source which can be represented.
4. Procedure after patent claim 1, by the fact characterized that by the space positions of the virtual acoustic sources wave fronts are produced, which are radiated prematurely in relation to the direct wave toward the delimitation surfaces of the rendition area, in order to make its acoustically noticed dimension smaller virtually.
5. Procedures after patent claim 1, by it marked that in a common rendition area discrete smaller virtual rendition areas are formed, in which different signals are produced.
6. Procedure after patent claim 1, by the fact characterized that the rendition sound field is reduced to a small space part, which it can be adjusted over change of the parameters of the listener position.
7. Procedure after patent claim 1, by the fact characterized that for noted sound events the data are linked temporally for the momentary position and adjustment of the acoustic sources in the recording room as well as the parameters of this area before the audio transmission separately transferred and during the rendition with the audio signal.
8. Device for the execution of the procedure after one of the requirements 1 to 7, characterized by a two-dimensional loudspeaker field, consisting of a multiplicity of loudspeakers or other suitable transducers, arranged frontally before a listener position, whose number, size and mutual distance are so selected that a wave field synthesis in horizontal and vertical direction is made possible, are developed, and by a WFS Renderer, which prepares the heading for signals for the individual loudspeakers and/or transducers.
Drawings
Fig.1 
Fig. 2
Fig.3
Fig.4
Fig.5 
Cooperation partner still searched
For practical realization of the described procedure a cooperation partner would be needed. The shown principles typifies the base of a 3d wave-field synthesis procedure, which seems to be the only practicable way for reconstruct the sound field in all 3 tree room dimensions physically. There are some supplementing ideas, which growing up by the author in work on the topic since several years. They are not revealed until today. By that becomes possible the reduction of the expenditure, the described divergence problem may solve, rough reflection surface can depict simplified and a transformation factor will become specify, by which change of the listener position in the rendition area fitted regarding recording room movements. For a partner with appropriate possibilities would result a fruitful cooperation. In that matter inquiries are welcome at any time. Please use by the contact page, incoming mails beocome confirm immidiately.
