DE4007841A1 - Audio signal processing for stereo headphone - adjusting stereo spectrum width or original audio signal and feeding back variable intensity - Google Patents

Abstract

Starting from an arbitrary, original audio signal, its stereo sound spectrum width is variably adjusted. It is then fed back with a variable intensity to the original signal. The two signal components of the generated stereo signal are variably mixed, with a delay, feedback, and their intensity. Then the two signal components are passed through separately adjustable filters to simulate time dependent, variable, two-ear, outer ear transmission functions. The stereo signal components are several times, variably, intensity dependent delayed, frequency distorted and mixed for simulating three-dimensionally, directionally independent, localised room reflections. ADVANTAGE - Almost real, natural audio signal reception and reproduction by stereo headphones.

Description

The invention relates to a method for ear-related audio signal processing for the Stereo headphone playback.

The following terms are used, their Meaning explained below:

• - outer tube
  Ear cup (also called outer ear relief) including ear canal
• - Outer ear transfer function
  linear distortions or unevenness in frequency response arising from the anatomical ear design
• - Out of head localization
  Naturally perceived, naturally perceived hearing events outside of the head or skull
• - binaural
  both ears (general)
• - fictional
  imagine yourself mentally or imaginatively
• - frontal plane
  Level of the head-related coordinate system that is perpendicular to the ear axis (e.g. hearing events perceived on the left, right, up and down)
• - hearing
  Summary of all signal processing processes from the outer, middle and inner tube during sound perception
• - Hall
  Sum of all reflected from walls, objects, etc. of a room, according to the original signal at z. B. listener incoming signal components
• - horizontal plane
  Level of the head-related coordinate system that is horizontal to the ear axis (e.g. including front, back, left and right)
• - listening event
  subjectively assessed sound source with regard to the location and the sound
• - compressor
  Device that compresses or compresses the dynamics (level fluctuation range) of an alternating tone signal
• - Head-related coordinate system
  coordinate system fixed to the head and ear design with the origin in the middle of the head
• - head-related stereophony
  Two-channel transmission method for the generation of ear signals, which offers a realistic spatial perception of hearing events when playing headphones
• - dummy head
  special recording microphone device, which also takes into account an average upper body and head anatomy
• - In-the-head localization
  unnatural hearing events felt within the head or skull, which occur when the hearing is overloaded with unknown ear signals
• - interaural
  compared between the left and right
• - loudness
  subjective assessment of acoustic stimuli with regard to their sound intensity
• - localization
  subjective perception of (acoustic) stimuli by means of the sensory organs
• - median level
  Plane of the head-related coordinate system that is perpendicular to the ear axis (e.g. including front, back, top and bottom)
• - ear axis
  straight line or axis defined between the left and right ear canal
• - ear signal
  Sound signals occurring in the ear canal or ear canal
• - location
  subjective connection of acoustic stimuli with its spatial location
• - directional characteristic
  the linear distortions that exist and are generated by the outer ear for each sound direction invention
• - Directional mixer
  Filter bank for electrical simulation of outer ear transmission properties
• - Sound event
  physical object (sound source), which, with regard to its spatial location, builds up complex sound fields
• - Stereo base width
  basic interaural signal delay difference of approx. 0.63 milliseconds (mean) existing in anatomically healthy people; on the other hand, mono signals have a stereo base width (or stereo sound image width) equal to zero, since the interaural signal delay difference takes the value zero milliseconds
• - synthesize
  Comb filter-like artificial sound image change especially below the frequency 1 kHz.

It is known that for the generation of moving hearing events with stereo headphones Sound reinforcement during sound recording in mixers by means of manual or electromechanical or computer-controlled panorama controller  a two-channel audio signal feed becomes. For example, similar a pendulum movement, between the left and right channel, sliding over the mono center one-dimensional stereophonic listening event built up with headphone sound, what is located in the top of the head.

Adequate listening events also provide for this studio effects units, which, based an admixture of changeable, delayed in the millisecond range Sound signals to the original sound signal, a dynamic Set up stereo panorama for headphone sound. In this regard, u. a. Effects like flanging, Chorus and phasing are generated.

Are related in the basic concept too Sound signal processing devices that have the effect from so-called Leslies, i.e. rotating Speakers, simulate in stereo.

All of the effects described so far (or Effect devices) simulate a moving, but limited listening event, that is with stereo headphones exclusively on one of the natural sound source localization possible Levels, namely the frontal level, the is spanned in u. a. three dots left, right on the ear and above, mono-centered in the head, one semicircular surface.

On the other hand, natural spatial hearing closes as is well known, all points of a three-dimensional Space (described by the median, frontal and horizontal plane).  

For static, i.e. immobile or as immobile simulated hearing events known that firstly headphones for a Out-of-head localization of hearing events free field or diffuse field equalized or be rectified in a directionally neutral manner (K. Genuit: Why Free field, notification of the radio exhibition Berlin 1983; Broadcast communications: Issue 1/1983, pages 17 to 26; Patent application: No. P 31 31 347.7-35; Advances in Acoustic DAGA 1987, pages 477 to 480).

The basis for these known methods is a replica of the directional characteristic one, statistically, "middle" human ear anatomy. In doing so Upper body, trunk, head and outer ear relief, for realization or equalization of artificial heads, directional mixers and Unified headphones considered.

There are primarily two today major types of artificial head developments, which equalized in free field and diffuse field Structure versions (Brüel & Kjaer catalog: "Head and Body Simulator 4128"; Patent registration: No. P 31 46 706.7; Broadcast communications: Issue 1/1981, pages 1 to 6). These artificial heads are the only medium to move freely Record sound sources and approximate realistic via stereo headphone sound to be able to reproduce.

In addition, there are only two directional mixers so far (including artificial head simulators) the predominantly arbitrary binaural head-related Build up listening events, any acoustic situations one in a room located, anatomically  considered, i. A. Simulate "medium" subjects. (Head-Acustics: information brochure and Report of the 13th Tonmeistertagung Munich 1984, Pages 103 to 110; Acoustic and Kino-Geräte GmbH: Information brochure and report of the 15th Tonmeistertagung Mainz 1988, pages 396 to 424). The generation of moving hearing events will also indicated in this context: a "dynamic exchange of parameters in real time" is feasible.

The relevance of customization of the system "dummy headphones" is under different points of view (J. Blauert: Spatial Hearing, Postscript, new results and trends since 1972, S. Hirzel Verlag, 1985; Acustica: Vol. 48, Pages 272 to 274). In particular, to the one-decibel-exact replica of the Polar pattern of the human Attention is paid to how they are used in audio signal transmission are to be considered. With frequency response deviations greater than 1 dB is otherwise an in-the-head localization of Hearing events inevitable because of the process the assignment of the hearing event to Sound source origin in the case of headphone Reproduction of the additional visual information to link with what you hear is missing (G. Plenge: on the problem of intracranial location of sound sources in human acoustic perception, Habilitation, TU Berlin 1973, pages 25 ff.).

So it is obvious that only in front of a (compatible with the recording system) Headphones depending on the direction, individually equalized stereo sound signal of a  Artificial head or directional mixer the always wanted out-of-head localization enables (F. König; patent application: P 39 22 118.0).

The measurement methods, the measurement setup and the so achieved results in connection with the Determination of the outer tube transfer function were also described in an overview (see above, Blauert: spatial listening, postscript). Modified as faster working and simplified procedure (for example lie after a few seconds of measurement already have graphics that draw conclusions about enable frequency-dependent distortions), F. König shows this with regard to the sound system variant "Headphones" (patent application: P 39 03 246.9 and P 39 12 582.3).

For recording and playback true to direction Public address systems were a. by Blauert / Boerger / Laws, Kürer / Plenge / Wilkens, Pleiderer etc. made efforts to the annoying side effect of "in-the-head localization" when playing sound recordings Reduce headphone sound (patent application: No. P 22 33 316.0, 26 28 053.0-31; Publications: No. 19 27 401, 22 44 162, 25 45 446, 25 57 519; Radio technology, Issue 6 + 7/1984, special edition, patent no. P 31 12 874.2-35).

The latter method in particular (by P. M. Pfleiderer) should be emphasized as it is practical Use, not as with "processor for out-of-head localization ", one with the scientific term corresponding hearing event Localization offers (see above G. Plenge,  Habilitation thesis, 1973). It is about in this device around an effects processor that room acoustics after the stereo sound builds up, but not one of the possible sound direction perceptions (from top-in-head location different) in the natural three-dimensional Location of hearing events implemented. Add to that proven the directivity of the Outer tube at (see Blauert above, spatial Hear, Postscript).

In addition, "has been a new kind since 1970 Presence filter "(J. Blauert: television and Kino-Technik, 1970, Issue 3, pages 75 to 78) and "a model for description of outer ear transmission properties " known (K. Genuit: dissertation, TH Aachen 1984).

Regarding stereo headphone sound of the outer ears under the stipulation, listening events locate essentially horizontally in front was able to use a special Proven arrangement of sound transducer systems, that the sound recording history no influence from tinned audio to the head-related realized there Pre-location has (F. König; patent application: P 39 42 127.9). Specifically describes this is the fact when using any, for example mono, intensity or runtime stereo sound signals with a suitable headphone sound transducer system arrangement (in the viewing direction front-below the Placed outer ear reliefs, linear distortions to mix up) head-related listening events to be able to move forward where an embossing of a z. B. individual  Stereo base width of i. A. approx. 0.63 milliseconds Delay time arises (no relevant Crosstalk from the left to the right ear noticeable). After due to the all-pass character of stereo headphones on the two Ears only those audible delays of the Sound recording (mostly between one and at least 100 milliseconds) come into effect, d. H. from sonicated persons to be felt at all and thereby the generation of direction or simulation of hearing events (here horizontally in front) unaffected remains, has therefore compliance the delay time value of e.g. B. 0.63 milliseconds as insignificant Ingredient in human binaural Direction detection crystallized out. The same conclusion is also after an analysis at natural, not head-based sound: the two on the head of an anatomically normal designed People at a distance of about 18 centimeters located ears produce the described Stereo base width of 0.63 milliseconds, which, however, regardless of the direction of sound remains approximately constant (the pair of ears is only partially mobile). Therefore only variable, "automatically" in the process of hearing data to be collected and evaluated, such as B. the known linear distortions the outer ear relief (see Blauert above or Genuit), for human direction detection be held responsible.  

Are completely unknown in the above context (spatial hearing combined with continuous Panorama left-right or flanging effects etc.) Discussions about moving "listeners", which one for example around the vertical Axis rotating artificial head would correspond. Here would, however, after a short time rotational or Signal transmission problems occur because when the artificial head not z. B. on top of the head cover has a transmitter (e.g. using a Infrared transmitter-receiver system or one System for frequency-modulated transmission via carrier signal), cable or similar twist in itself and thus a continuation of the artificial head rotation prevent. The same applies to one Combination of moving sound sources with moving artificial heads (when playing via Stereo headphones as "fictional listeners" too consider). According to the descriptions by Blauert and Plenge (see above) has this in particular following evidence: The relative probability the occurrence of an out-of-head localization of stationary sound events Subjects with head not fixed back, because a sound event localizing person if you are unsure about your whereabouts a reflexive alignment movement of the sound source of the head (turning the head so that one of the two ears, which is a measurement base correspond, briefly closer to that Sound event is located). Another decline the out-of-head localization of hearing events is therefore too with movable sound sources expect.  

It is also known loudness assessments of broadband signals using the recommendation according to DIN 45 631 (Acustica: Vol. 55, Pages 63 to 67). It is today without a known exception to a measuring method for (Environmental) noise analysis.

Finally, it is known to be broadband and extremely fluctuating depending on the level highly dynamic signals, in the recording studio area total level evaluating peak limitation and Compressor devices, readjust. In application on Show signals with a broad spectral distribution such dynamics-reduced processes, which especially to improve speech intelligibility from u. a. Broadcast production contribute to irritating dynamic pumping effects. Pump effects are e.g. B. by blowing sounds Speaking in front of a microphone triggered (as a pop sound known). In this context Spectral components not belonging to the useful signal generated that frequency once below the for the language essential spectrum to lie come and also, based on the current Total level, deliver the largest share of energy. With the dynamic reduction applied today are u. U. various syllables from words incomprehensible. In addition, these common devices find Dynamic reduction in modification also for Suppression of high-frequency audio noise their use (so-called Denoiser).  

In summary it can be said that a realistic natural sound signal recording, (broadcast) and playback situation via stereo body with partial Success is achieved, but a complete one Success in the existence of a conglomerate of different definable errors fails, the causes of which have so far been mostly wrong order (e.g. priority "Playback volume" before "individual adjustment"), as well as incompletely corrected.

The invention is based on the object Create procedures of the type in question, which is a realistic natural sound signal recording, (Transmission) and playback situation via stereo headphones with maximum success allowed.

This task is solved by the by the characteristic features of the Claim 1.

Advantageous developments of the invention are listed in the subclaims. It is not necessarily according to the invention by one monophonic input signal to be processed to go out because there is an opportunity inferior, often found stereophonic signals of a sound engineering Undergo revision.  

So it is an advantage that to collect the data used according to claim 5 about the "Medium" directional characteristic of a selected one Person the following points should be observed:

• - Three-dimensionally spatially directional Public address of this selected subject by means of suitable acoustic test signals.
• - Implementation of the subject sound in the median, horizontal and frontal plane with a directional resolution of less than 30 degrees.
• - Implementation of the subject sound in an anechoic room.
• - determination of both ears Transfer dimensions of the selected subject using suitable test signals.
• - Recording of binaural group transit times and interaural time differences using of these suitable test signals.
• - Realization of the suitable test signals by means of an extension of the standard measuring frequency sequence according to DIN 45401 to a quarter third stepped logarithmic frequency scale, which is at least the frequency dependent Resolving power of the geometric Provides ratio twelfth root of two.
• - The determination of outer tube transmission dimensions takes place in a frequency dependent Level investigation range of at least 20 Hz to 20 kHz with a level test accuracy less than 0.5 dB.
• - The determination of outer tube transmission dimensions is about four millimeters inward Auditory canal of the selected subject before Tympanic membrane made.  
• - The determination of outer tube transmission dimensions is characterized by a low-distortion, probe-like Device that is centered in the Ear canal mounted sound delivery tube Test signal samples to an outside of the Microphone located in the ear canal, realized.
• - Determination of the selected subject from a Large number of people (at least eight) through correlative comparisons of outer ear transmission dimensions, which on average about the maximum of yourself in the multitude of individual outer ear transmission dimensions statistically increasing frequency-dependent Has curve contours.
• - The at the outer ear transmission dimension Determination used probe-like device has a frequency range of 20 Hz up to 20 kHz over an accuracy of the sound conversion and transmission properties of less plus / minus 0.5 dB.
• - That for the probe-like device The microphone used has a dynamic range greater than 58 dB.

Regarding the simulation of three-dimensional spatially resolved moving hearing events, u. a. according to claim 6, it is advantageous that a suitable sound incidence angular resolution that for stereo headphone playback simulating hearing event directions is selected (Steps less than 30 degrees in the median, frontal and horizontal plane; preferably 10 degrees). This enables via stereo headphone playback once the simulation of moving widely Sound events according to claim 11 and / or the  Simulation of the lateral alignment movement of a unfixed human head during the localization of, for example, auditory difficult to grasp sound sources, which one Increased security and speed in the general direction assignment (a decrease in in-head localization of Hearing events are expected).

It is also an advantage that how it is after one of claims 12 or 13 is expressed, a digital in the inventive method co-generated code that is a time-dependent changeable information about the simulated Includes listening event direction, playback side (e.g. in a compact disc player) as follows finds: This direction information code in the creation of canned clay recorded with, so is playback side in a suitably equipped stereo sound reproduction device a subsequent customization the listening events to person-specific Outer ear conditions, d. H. to the anatomical Design of the entire auditory system possible. This is advantageously done again by means of two variably controllable digital filter realized.

It is also advantageous that the used analog amplifier, the elements analog-digital (A / D-), digital-to-analog converter (D / A converter) as well as the chain of digital signal processing regarding the quality of the electrical transmission behavior at least the requirements like comply with:

• - General useful signal-to-noise ratio greater than 96 dB,  
• - including a gain factor of 40 dB results for the analog amplifier a useful signal-to-noise ratio of at least 136 dB,
• - dynamics greater than 60 dB,
• - linear distortions (max. deviation) less than 0.5 dB and
• - nonlinear distortions less than 0.1%
• - Larger according to an A / D and D / A conversion 16 bit.

Furthermore, it is advantageous that for playback of hearing events generated according to the invention a compatible equalized for recording, d. H. for example free-field or diffuse-field equalized, Stereo headphones are used.

The advantages achievable with the aid of the invention compared to the prior art at External ear sound using stereo headphones consist in particular in that

• a) Any simulations of three-dimensionally spatially moving hearing events, including "average" direction-dependent linear distortions of a person as well as room acoustic conditions, which
  • a1) the movements carried out by a hearing person with respect to a constant, ie location-specific, fixed sound source or
  • a2) the movements of a sound source that is not static in a room builds up, are made possible,
• b) a subsequent adjustment on the reproduction side from moving hearing events to individual Polar patterns (starting from the "medium" polar pattern) enables becomes,
• c) a comparatively low amount of hardware to implement the invention Procedure exists
• d) parallel to the distortion of the binaural Directional perception clan image effects in Connection with the binaural Stereo base width (i.e. changeable Left-right delay time) be asked
• e) z. B. undesirable in pop music productions highly dynamic instruments, voice or Noise dynamically compressed depending on loudness will,
• f) no switching noise during Simulations of moving hearing events occur and
• g) according to the invention for hearing provided essential basic information on general sound event direction location, in the form of points a) and b) and a simulation of distance perception the fictitious listener of sound sources and their distance from walls in acoustic Clear, is completed. Essential Meaning for a familiar, natural Out-of-head localization of hearing events is also that of u. a. Blauert and Plenge (see above), mobility the head of hearing persons (supplementary also that of the sound source). After the Advantages according to the invention explained a)  and b) this is taken into account here. Therefore the relative probability goes a perceived localization of the head of Hearing events significantly back.

The invention is based on Drawings are explained in more detail; in these shows:

Fig. 1 is a block diagram with an apparatus for performing the method according to the invention, which generates a head-related stereophonic sound signal from an underlying monophonic original sound signal by means of sound image synthesis and simulations of moving binaural auditory events

Fig. 2 shows a spatial representation of a moving sound source vector of the simulated listener and the acoustic emission thereby occurring or Schallimmisionskomponenten.

A) Hardware concept

The embodiment shown in FIG. 1 essentially describes a software-controlled hardware signal processing device which, by means of analog and digital signal processing, leads to multidimensional spatial, natural hearing events when played back using conventional stereo headphones. First, the single-channel audio signal, here e.g. . B. a microphone ( 1 ), brought by a broadband, low-noise, controllable preamplifier ( 2 ) into the appropriate voltage level of the subsequent various analog-digital converters ( 4 ) and ( 6 ). For this purpose, an input-side loudness analysis of the temporally changing sound signal is implemented, which, after digital signal processing, provides for a control or adjustment of the gain factor of ( 2 ). The prerequisite for this is that a relative adjustment or adjustment of the detected voltages to sone levels and frequency groups (in accordance with the loudness determination according to DIN 45631) is carried out first. This is based on a multi-selective bandpass subdivision into thirds ( 3 ), which results in a large number of partial voltages. Later the frequency groups necessary for the loudness determination are formed from it (by a CU ( 40 ), see below). In addition, the individual partial voltages are separated from one another in such a way that each frequency band can be selectively analyzed on a digital level ( 40 ) by means of a subsequent analog-digital converter ( 6 ). The connection to an EDP system (corresponds to the abbreviation for electrical data processing) is guaranteed by a bidirectional data bus ( 7 ). A control signal {code}, which has been input by hand and / or generated accordingly, is then passed to one of the digital-to-analog converters ( 9 ) or ( 34 ), {the selection made by the system user}, the {( 9 ) or ( 34 )} controls the preamplifier ( 2 ) or the two post-amplifiers ( 28 ) and ( 30 ). In this regard, an automatic limiter safety control {or control ( 3, 40 )} of the sound signal reaching the selective filter or the analog-digital converter ( 4 ) is integrated in order to prevent impulsive noises which occur, for example, at the start of the use of the device. An interface ( 8 ) manages the data reception from the analog-digital converter ( 6 ) and the data transmission from the digital-analog converter ( 9 ).

Specifically denotes the input or output-side variable, loudness-dependent Level adjustment the following points:

• P₁) compression and peak limit after the S curve commonly used today {Called hysteresis loop}, with which increasing low-level signals increasingly raised and vice versa, increasingly high-level signals are increasingly limited.
• P₂) This S curve relates to the invention now on the value of loudness {in sone; loudness level may also be used in phone} of broadband signals that frequency group dependent {according to DIN 45631} rated and summarized one of usually relevant, targeted Bumps in frequency responses {i. a. referred to as a compression curve family} deviating, hearing-related dynamic reduction makes.
• P₃) The S curve is characterized by a maximum and minimum loudness value limits which are to be regarded as "soft limits" {here z. B. 50 sone and 200 sone}.
• P₄) Thereby get low-level to be regulated Signals depending on loudness in various Frequency groups {e.g. B. the frequency range around the C⁵ valley} a special meaning, resulting in a lower upregulation effect arises.  
• P₅) Reverse to P₄), in loud, d. H. at the upper "soft limit" coming to lie Signals, the spectral analytical z. B. have narrow-band bumps, obtained due to the concealment effect an intensifying re-regulation effect.
• P₆) A the original signal in the spectrum distorting effect is excluded.

The audio signal, which has been converted into a digital code by means of the analog / digital converter ( 4 ), is now upgraded by several, monodirectional drivers ( 5 ) working in parallel for larger connection loads. A driver module with so many channels is used for this that the i. A. Binary resolution from the analog-to-digital converter coming signal ( 4 ) is covered {ie e.g. 16 bit results in 16 data lines and 16 driver channels}. Due to the increased fan-out {measure of the resilience of a digital module output with a certain number of connectable further components}, such drivers offer the basis for "copying" or "reproduction" of the digital code in the application example according to the invention.

The amount of the fan-out is determined by the "imaginary" {explanation, see below under "fictitious"} desired three-dimensional spatial transparency determines which, in other words, the sum all possible representable directional components of a spatial hearing event to be simulated and thus corresponds to the number of code copies.  

In this context, the code duplications from the driver ( 5 ) to "fictitious", single-channel time delay elements {( 12 ), ( 14 ); k _mn × T _n , which corresponds to multiple delays} and direction generators {( 13 ), ( 15 ); Z _n including the regulation of the partial level transmission factor and the simulation of direction-dependent linear distortions by means of digital filters} via monodirectional data bus ( 11 ). It can be z. B. build a three-dimensional spatial direction matrix with a resolution in ten-degree steps per horizontal, vertical and median plane, which corresponds to 36 × 36 × 36 {corresponds to the value n} different outer ear transmission measures {Z _n } and to n-different leads k _m multiples {thus k _mn } space reflection patterns {T _n }. As the number n increases, the left {here z. B. ( 13 )} and right channel {z. B. ( 15 )} with k _mn × T _n and Z _n . Adequate to Z₁ and Z _{n also} applies to the direction generation after the preparation of the monophonic to the binaural, digital sound signal by means of the elements Z _OL ( 21 ) and Z _OR ( 23 ).

A few notes on the following discussions: All "fictitious" digitally processed audio signals are generally forwarded {see above} via a monodirectional bus ( 11 ). In addition, "fictional" describes the descriptive character of the selected element arrangement {according to ( 12 ) to ( 15 )}, because in practice delays {T _nt }, frequency response distortions {Z _n } and the partial level controls included here are carried out by means of computer-aided manipulation of digital data which ia a data link {or Data concatenation}. The same is symbolically expressed again in FIG. 1: The data transfer between the audio signal delay or audio signal distortion elements and the actual core of the data processing, the central unit {( 40 ); from the English: central unit = CU}, today is never handled via a bidirectional data bus ( 7 ), but software-controlled, faster, space-saving and more precisely directly from the central unit ( 40 ), which consists of an input / output system, data memory, command - and data processor {including a crystal oscillator}. This offers the possibility, inter alia, of fetching, without detours, the data sets required for moving hearing events, clocked from the storage medium ( 26 ) into the data processor ( 40 ) via the directional characteristic of a "middle" ear, and to link them in a program-controlled manner with the time-varying digital input signal. Therefore, the end product of this system according to the invention forms, among other things, a completely menu-controlled animation program which offers the user any binaural-acoustic conditions.

An alphanumeric keypad ( 41 ) for user-related input, a display device ( 39 ) and a permanent storage medium {( 26 ); e.g. B. a hard disk or hard disc} complement the CU ( 40 ) as a functional EDP system.

A special peculiarity according to the invention are the elements T _OL ( 16 ), T _OR ( 22 ), T _L ( 19 ), T _R ( 18 ), R _L ( 20 ) and R _R ( 17 ), which are variable in their effect and which full system bandwidth generate {unique} signal delays and attenuations. In the simplest case, T _OL or T _OR are intended for realizing the binaural stereo base width of approximately 0.63 milliseconds delay time. Together with the elements T _L , R _L , T _R and R _{R, which} are switched "fictitiously" from left to right and vice versa below, even more complex effects are possible:

• I) T _OL remains constant at the value of 0.63 milliseconds, while T _OR in fine steps between zero and 100 milliseconds z. B. sinusoidal or triangular waveform modulated in value {increasing / decreasing}.
• II) T _OL is modulated between 0.63 and 100 milliseconds like element T _OR according to I), T _OR has the value zero milliseconds.
• III) T _OR remains at a value of the delay time of 50 milliseconds, T _OL has the value zero milliseconds.
• IV) T _OR according to III), only that a value of 0.63 milliseconds is taken.
• V) The effects from I) to III) are realized with the elements T _R and T _L as required, with an additional subdivision according to V1) point I), V2) point II), V3) point III), V4) point IV) and applies these effects obtained by T _L and T _R are freely adjusted in its intensity by means of R _L and R _R.
• VI) The effects of V), only that the temporal modulation flank is left-right inverse {z. B. T _L has a continuously increasing delay time over a period of time, while T _{R has} a decreasing delay time}.

An optional exchange of the elements T _OL with T _OR or T _L with T _R is also provided, as is any combination, expansion of the effects {towards higher delay time values} and feedback of the signal {depending on the delay elements T _OR , T _OL , T _R , T _L , the intensity of which can be regulated and variably assigned to the delay element}.

There may be an extension for point V) from I) to IV) {instead of above from 'I) to III)'} conceivable, but this is not worth mentioning sound-generating or other effect built up {im continuous text as sound image synthesis designated}. Here, variable and time-varying Comb filter structures {e.g. B. quarter-third width Drops and resonances} below frequency 1.6 kHz {1 / 0.63 milliseconds} generated, which once this range {<1.6 kHz} regarding the directional / specific linear distortions {e.g. B. a third width Level increase} and the related Leaves directional perception unaffected. On the other hand, those in the frequency domain below 1.6 kHz mainly occurring tonal Fundamental waves of musical instruments and noises {vs. harmonics of a spectrum} are influenced, which is why sounds without Beat can be processed so far that, except the well-known flanging and phasing effects, chorus effects are also possible can.  

Another general note: In order to indicate the variety of possibilities according to the invention, in the example shown the monophonic, digital original sound signal was passed to the various delay elements {here ( 12 ) and ( 14 )} undistorted, ie without sound image synthesis, which is why only the primary binaural { digital; see signal flow after ( 21 ) and ( 23 )} sound signal is distorted. In principle, there is no restriction in the invention in this regard, so that, for. B. the monophonic original sound signal can initially be subject to a tone color synthesis {the reason for this follows from 'Notes'; see above}.

The mixing of the digital code of T _OL ( 16 ) with that of T _R ( 18 ) and R _R ( 17 ) can be achieved by means of data linkage ( 24 ). The same applies to the code of T _OR ( 22 ), linked with T _L ( 19 ) and R _L ( 20 ).

The code modified at this point into stereophonic, digital audio signal data is then assigned a variable "middle" directional characteristic of the outer ear by two controllable ( 7 ) digital filters {Z _OL ( 21 ), (Z _OR ( 23 )}.

The realization of the following "fictitious" multiple, two-channel mixture ( 25 ) of the various digital partial signals according to Z _OL ( 21 ) and Z _OR ( 23 ) with the spatial reflections k _m1 × T₁ ( 12 ), Z₁ ( 13 ), up to k _mn × In principle, T _n ( 14 ), Z _n ( 15 ) is again realized by means of data linkage or concatenation {note: in practice by the CU ( 40 ), see above}. All R, T and Z elements are "fictitiously" controlled by the central unit ( 40 ) or supplied with data records from ( 26 ) via the bidirectional bus ( 7 ).

At the input of a pair of digital-analog converters ( 27 ) and ( 31 ) there is a code that changes over time, the analog, binaural sound signals that are generated finally with respect to their stereo output level via a controllable post-amplifier ( 28 ) and ( 32 ) in a suitable voltage level to be brought. The data input for the general setting or regulation of the amplifiers {in addition to ( 28 ) and ( 32 ) and also ( 2 )} is carried out alphanumerically via a keypad or joystick ( 41 ), whereupon the central unit ( 40 ) receives a digital code via ( 7 ) an analog-digital converter ( 34 ) conducts. The digital-to-analog converter ( 34 ) then generates the control signal for the controllable post-amplifiers ( 28 ) and ( 32 ). These take over {see, inter alia, the points P 1 ff.} Instead of the preamplifier ( 2 ), optionally also the loudness-related, broadband audio signal compression control.

Parallel to the analog audio signal outputs left ( 35 ) and right ( 37 ), the binaural code for digital signal processing or signal use {after the connection points ( 25 )} is concatenated with directional information, amplified {( 30 ), driver} to a third { digital} audio signal output ( 36 ). In addition, this direction information, in the form of a direction code generated by ( 40 ), is clocked in a frequency-varying manner depending on the speed of movement of a "fictitious" listener or sound event and is program-controlled via ( 7 ) to a further driver ( 33 ) on a separate, specially designed output ( 38 ) brought. The direction information created by ( 40 ) describes in a corresponding resolution {e.g. B. in 10 degree steps} the current time-variable, binaural distortion state of the elements Z _OL ( 21 ) and Z _OR ( 23 ) as a function of the three-dimensionally-spatially simulated direction of incidence of a hearing event related to a "fictitious" listener.

B) Simulation example

In a completely arbitrarily chosen, however practical use case, is also with the device previously shown in section A) a simulation of head-related movable Outlined listening events. Typing Data on general room acoustics and on the multidimensional complex movement takes place menu driven in real time {e.g. B. Today commonly used mouse, joystick or The like}. Specifically

• a) a sic in a theater hall,
• b) on a stage {from the audience considered},
• c) artist moving from left to right
• d) at the same time at a distance of two meters a sound source on an approximate hurls a circular path,
• e) this overall sound image itself still hears
• f) simulated in a binaural space-related manner
• g) synthesized, so that the time changing, three-dimensional acoustic Hall impression as well as the spatial Spin movement of the sound source remains, but not  
• h) their head-related stereo base width {0.63 millisecond time delay between the left and right ear} and
• i) the stereo base broadening associated with h) as well as dynamic sound distortion, which is a temporal variance of the Stereo base width triggers.

In other words, it is for a medium Stereo headphones sounded person the impression created as if in this, easily by natural standards deviating {see points g) ff .; Synthesization}, "acoustic" situation would be.

In order to meet the requirements of points a) to i), all the data required for this must first be collected and entered into an EDP system, as shown in FIG. 1 {requirement}, which essentially comprises the following:

• X₁) The maximum loudness dynamic range of the limited fluctuating, monophonic Input signal {here from 70 to 90 sone},
• X₂) the middle, to be observed, stereophonic Output voltage level {here 0.5 volt},
• X₃) a scale representation of the Room dimensions and design,
• X₄) the staggered sequence of movements of the artist on the stage in the hall {here combined in a straight-line rotating manner; The speed of the linear movement in meters per second}, the frequency of rotation of the sound source is twice as high as that of the moving artist, takes the value one revolution per second and runs in reverse to the direction of rotation of the artist; see Fig. 3,
• X₅) the spin or rotational movement of the Ball sound source at a distance of two meters,
• X₆) by a stereo base broadening characteristic constant, interaural Time delay value {here 50 milliseconds; see above III} as well
• X₇) two temporally up and down modulated, however the left channel inversely to the right delayed behavior, mixed, delayed additional signals {see above V2) combined with VI); the Modulation frequency of the delay elements is equal to one sixteenth of the rotation frequency the thrown sound source}.

The basic state or basic place of residence {on Time zero} of those available in this example Sound source and the artist {which one Be assigned to listeners} be neglected, what with a general simulation would not be allowed.

To the example above, especially with regard to points X₃ to X₅, some detailed explanations: The spatial data of the theater hall serve, including the related information about the partial reflectivity of the individual hall parts {ia the acoustic behavior of geometrically different walls, objects and their Degree of sound absorption three-dimensionally resolved}, for a matrix-like assembly of the n _m directional {z₁ to Z _n } varying k _m -multiple room reflections {k _m1 × T₁ to k _mn × T _n }, which are also called Hall. Acoustic data on the absorption or reflectivity of any objects and on the directionally variable linear distortions on a "middle" test person's ear are simpler to assume than integrated in the EDP system, whereby software-controlled immediate connection {or "managing"} this data for the simulation according to the invention is made possible in real time.

According to the descriptions of the resolving power of the directional characteristics in various levels in accordance with section A), point d) and X₄ from above, the rotational movement of the spherical sound source in the horizontal plane around the artist in the example results in an exchange frequency of 36 (corresponds to the "fictitious" listener) Records per second. The rotational movement takes place at the same distance {two meters}, which is why influences of human distance perception {in the form of the spatial reflection pattern and the distance law, i.e. sum of all partial volumes and spectral diffractions depending on the sound source distance} in the formation of binaural, linear, direction-dependent distortions { Z _OL , Z _OR ; see Fig. 1} are negligible.

In detail now for the present Example of the following acoustic relationships based on the room geometry or design of a real theater hall, but shown in a much simpler way, are discussed, the i. a. per technical drawing can be entered into a computer system. The theater hall is a stage with a back room, a  fanned auditorium and semi-circular, curved, three-story Ranks marked. The roof is also semicircular, but resembles a vault. The stage is a ten meter high, heavy curtain attached to the side. The back room is four meters rectangular viewed backwards limited so that a seen open to the auditorium Room measuring 16 × 6 × 4 meters {Width, depth and height; equals to Stage decoration} arises. With the relative dimensions of 25 × 25 × 20 meters {Latitude, Longitude, and Height} then begins directly after the stage decoration the auditorium. With fully occupied parquet and fully occupied Ranks result in the sonicated state relatively dry acoustics caused by a three-dimensional reverberation field {average reverberation time for example 1.5 seconds}. The spectral components of the Hall field are above approx. 1 kHz with 12 dB per octave from {below 40 Hz no reverb} accepted. Directional Half-part components are not meant to be here be worked out as this as well as the implementation of room acoustic data for so-called {digital} room simulators, already described numerous in the literature and was applied. The same will in practice {i.e. on the market Devices} from a wide range of acoustical conditions, what in its diversity for the invention applies, but secondly in detail the reasonable Would exceed the scope of this example.  

The room acoustical data discussed above, required in point X₃, with the following descriptions for points X₄ and X₅ results in a complete, additionally sonically synthesized, room sound simulation with a "fictitious" mobile listener according to FIG. 2:

A moving simulated sound source (2) and for a specific time t _x applicable binaural localization of the sound source (2) of a "fictitious" listener 1 corresponding to a dancing on the stage performers, found on the dashed lines drawn web (4) from the left to the right instead. Doing so

• Y₁) the spherical sound source ( 2 ) {with its radiation characteristic},
• Y₂) thrown the sound source ( 2 ) counterclockwise ( 6 ),
• Y₃) the sound incidence angle α, which occurs relatively to the artist, and the associated binaural, direction-dependent, linear distortions {corresponds to the mode of action of the elements Z _OL and Z _OR ; see Fig. 1},
• Y₄) the rotating movement clockwise ( 5 ) of the artist ( 1 ) himself and
• Y₅) one of the statically distributed, direction-dependent reflection subcomponents arriving at the artist ( 1 ) {( 3 ); corresponds to one of the elements Z₁, k _m1 × T₁ to Z _n , k _mn × T _n ; see Fig. 1} shown generally vectorially.

Accordingly, an acoustic situation is presented to a passive head-related listener {at time t _x } which comprises a primary listening event diagonally left-front {horizontal plane} as well as, in part, a secondary listening event which corresponds to one of the statistically distributed many theater hall reflections T _n ( 3 ) from the direction oblique-right-front-top {median plane}. The "minus" at T _n denotes the noise emission in the selected vector representation, since without the "minus" there would otherwise be a noise emission.

As a special feature of the invention any monophonic, then binaural-spatial, variable stereo base width widened, upgraded sound signal synthesized. The generation of spatial reflections remains thereby {only here in the example} from the Sound image synthesis unaffected.

These mainly simulation components and processes described imaginarily in points Y₁ to Y₅ are, as already indicated, ear-related in a device according to the invention according to FIG. 1, that is, built up binaurally three-dimensionally. It is recommended to play back {sound signals} via compatible headphones {for details see above}.

Claims (25)

1. A method for hearing-related audio signal processing for stereo headphone playback, characterized in that

• a) that starting from any sound signal (original signal), the stereo sound image width of which is set variably, and the intensity of which is fed back to the original signal,
• b) that the two signal components of the stereo sound signal generated according to a) are mixed with one another with delay, feedback and in terms of their intensity,
• c) that the signal components of the stereo signal prepared according to b) pass through filters which can be set separately from one another, in order to simulate, if appropriate, time-dependent variable two-ear outer ear transmission functions,
• d) that the signal components of the stereo signal prepared according to b) are delayed in a variable manner depending on the intensity, distorted and mixed depending on the frequency in order to emulate spatial reflections which can be localized in a three-dimensional and directional manner, and
• e) that the stereo signals prepared according to c) and d) are mixed with one another in a variable manner.

2. The method according to claim 1, d. g. that a stereo signal component by component, separate Analog-digital conversion of sound signals in digital, separate codes. 3. The method according to any one of claims 1 or 2, characterized by a before the analog-digital Switchable adjustable, broadband, analog audio signal amplification. 4. The method according to any one of claims 2 or 3, d. g. that per stereo channel by bit shift and caching the data of the respective digital stereo signal components a preparation this signal to the head-related or a Simulation of binaural and / or interaural Delays, as well as finely structured linear distortion according to one of the two ears Polar pattern, is made. 5. The method according to any one of claims 2 to 4, characterized by a permanent Storage of suitable digital Records over a three-dimensional spatial resolved outer tube directional characteristic of a viewed anatomically, selected medium subjects.   6. The method according to claim 5, d. g. that records by means of a clocking variable in frequency accessed from a permanent storage medium which are used for the simulation of in Require real-time moving hearing events Information about the three-dimensional spatial resolved outer ear polar pattern of the selected subjects. 7. The method according to claim 6, d. g. that, from a multiple existing two-channel, but monophonic digital audio signal, the formation of a two-channel, binaurally different base tone signal, enables and, based on the value the interaural audio signal delay of the selected subject, by means of a variable, cyclically changeable, interaural digital audio signal delay is made. 8. The method according to any one of claims 4 or 7, d. g. that one after the Generation of the binaural base signal following separated-two-channel, in its intensity and Signal delay time independent, variable, variable, cross-over mixing of the two digital audio signal components is made.   9. The method according to any one of claims 1, 4, 7 or 8, d. g. that for sound image processing and / or Synthesis, per delay element, a variable in intensity, generated on digital signal processing level, Feedback is made. 10. The method according to any one of claims 1 to 6, d. G., that it can be determined without restriction which one three-dimensional spatial trajectories the digitally simulated listening event for the obey headphone-related sonicated person should. 11. The method according to claim 10, dg that it is freely selectable whether

• a) a one-dimensional-complex movement, which describes a three-dimensionally moving sound source located in an acoustic room to a motionless fictional listener, who is arbitrarily three-dimensional in this room or vice versa, a motionless sound source to a moving listener or
• b) a two-dimensionally complex movement, which describes a sound source moving three-dimensionally in an acoustic space and a fictitious listener moving three-dimensionally, is simulated.

12. The method according to claim 1, 6, 10 or 11, characterized by a separately transmitted digital signal contained in a code encrypts all geometric information about a time-changing Directional perception of a fictitiously moving Locating listener listening events represents. 13. The method according to any one of claims 11 or 12, d. g. that for the relative change of Directional perception of hearing events one, the full scope of this change, inclusive of the complete data sets and the simulated one- or two-dimensional complex Movement type corresponding resolution of the Information about the directional perception containing codes is realized. 14. The method according to any one of claims 1 or 13, characterized by a variable implementation of hearing events related to distance perception. 15. The method according to claim 14, d. g. that the variable realization of hearing events, with regard to distance perception, by means of selective digital filter and acoustic, multi-dimensionally distributed, varying Room reflections due to digital audio signal delays according to one of claims 4, 7 or 8 based.   16. The method according to claim 3, 4, 7 to 11 and 15, d. g. that the varying spatial reflections created from the digital base tone signal and in an acoustically realistic, influenceable relationship to the head-related Base tone signal can be added digitally. 17. The method according to any one of claims 6, 10, 15 or 16, i.e. g. that the generated vary Room reflections often copied in several separate single signal processing paths with associated three-dimensional spatial outer ear Directional characteristics are provided. 18. The method according to any one of claims 15 to 17, d. g. that to simulate three-dimensional spatial reflection patterns of any fictional room configurations and the lining of which is assumed. 19. The method according to any one of claims 1 to 18, d. g. that for ear-related sound signal compression a loudness-dependent, in its effectiveness variably adjustable and / or yourself automatic level correction of the original sound signal.   20. The method according to claim 19, d. G., that the loudness-dependent influence of the sound signal by means of two fixed entered numerical values which are the maximum and minimum limits of Loudness dynamic range, adapted to the respective properties of the sound signal, represent is determined. 21. The method according to any one of claims 6, 10 to 20, characterized by an interpolative Processing of digital data, which one smooth transition from one to the next directional simulating pair of binaural data sets realized without background noise. 22. The method according to any one of claims 1 to 21, characterized by a separate multi-channel Digital-to-analog conversion of the digital present, edited, binaural Tone signal. 23. The method according to claim 22, d. g. that the digital-to-analog converted multi-channel Sound signal broadband and adjustable as desired is amplified.   24. The method according to any one of claims 3 to 23, d. g. that all variable or adjustable control variables of the simulation by means of alphanumeric data entry or graphic Screen entry can be made. 25. The method according to claim 24, characterized through computer-aided control of the entire data processing and I / O operations.