Binaural recording
A binaural recording is a recording of acoustic signals with microphones that will create a natural listening experience with precise localization direction when playing only through headphones. When recording in a dummy head stereophony dummy head is frequently used.
- 4.1 Without artificial head
- 4.2 Binaural dummy head recordings
- 4.3 Further binaural microphone system alternatives
- 7.1 measurements
- 7.2 Radio Plays
- 7.3 music productions
Stereo and binaural
The term " binaural " was formerly often equated with stereo.
General stereo sounds will be mixed alone speaker systems when listening; hence the name " speaker stereophonic ". The properties used by humans for localization as the head or ear molds are not considered to be legal. This is because each ear of the listener is the ear signals itself in natural hearing when played through the stereo speakers in the stereo triangle.
Binaural recordings are " stereo " recordings with special recording technique that should be but typically correctly reproduced only with headphones; hence the name "headphones stereophonic ". Binaural recordings - which replace the ligated by headphone reproduction natural ear signals - are the best way to reproduce a realistic sense of spatial hearing.
Purists want only the images of artificial heads with auricles called " binaural ". This has however not enforced.
Basics of Hearing
Definition of an auditory event
A sound event is only at a hearing event when the sound waves have penetrated the ear, lie in the brain as an irritant and processed by him to an auditory perception. One can sound and auditory event also are not directly comparable because the stimulus shape is changed in the middle and inner ear. This means that the ear the signal is distorted to some extent. These charms are feelings that are different from person to person and depend on frequency, duration, and the sound pressure level of the sound event. Sensations are not measurable, but perceptually statistically detectable by listening studies.
Localization of sound events
Man is able to allocate its perceived auditory events specific directions. Humans have perception mechanisms for half the horizontal plane and the median plane. In a head- based coordinate system, the direction is defined horizontally forwardly as a 0 ° axis.
Horizontal plane
As soon as the sound source is not directly on the 0 ° axis, it comes to interaural time differences ( ITD) and interaural level differences ( ILD). Travel time differences ( time differences ) can be evaluated by the human ear already at a size of 10 microseconds to the direction of localization. This corresponds to a localization sharpness in the horizontal plane of about one degree. Up to a time difference of 0.63 ms (ie, 630 microseconds ) increases the lateral localization is roughly proportional to the propagation time difference. A time difference of 0.63 ms corresponds to a path difference of the sound of 21.5 cm. This, also, " Hornbostel - Wertheimer constant" called size corresponds to the distance difference in sound incidence from 90 ° or 270 ° direction of incidence at an average distance between the two ears on the human head.
By interaural differences in the sound signal can be located in the horizontal plane. However, this can not be distinguished from noise from the front and rear horizontal plane, since this signal interaural differences are equal. A distinction is three frequency ranges; also compare duplex theory of John William Strutt, 3rd Baron Rayleigh, JW Strutt in 1907.
- Below 800 Hz, the determination of interaural time differences ITD, especially from the analysis of phase differences between the ear signals. Interaural level differences play no role here.
- Above 1600 Hz, the localization is based on the evaluation of interaural level differences ILD, and the analysis of interaural group delay differences ( differences in delay of the signal envelope )
- In the range between 800 Hz and 1600 Hz, the effective range of the effects involved overlaps. With increasing frequency, the angular range can be evaluated in the interaural phase differences smaller. For the size of the interaural level differences increases.
Interaural level differences caused by shadowing by the head. Interaural level differences are strongly frequency dependent. Frequencies with wavelengths on the order of the obstacle are barely bent. Low frequencies below about 300 Hz do not form a sound shadow and thus not quotable for the localization of the direction of level differences. However well these low frequencies with the phase differences a special influence on the perception of space and enclosure feeling.
Median plane
In the median plane, there is almost no run- time differences and level differences between both ears. For localization of a sound event here, the acoustic properties of the outer ear can be utilized.
The different elevations and depressions of the outer ear, together with the ear canal, an acoustic resonator, which is, depending on whether the sound is received from the front, from above or from behind, different excited. This results in directional minima and maxima in the frequency response of the ear sensitivity. The sound thus modified noise is isolated from the hearing in the median plane and corrects phonetically.
The frequency ranges, which are excited by specific directions of incidence are also referred to as direction-determining bands. Since the shape of the outer ear for each person is different, everyone has slightly different " frequency responses " for front, top and rear.
Taken with the artificial head sound events that lie in the median plane, are often difficult to locate during the subsequent interception. So there is often front-rear exchanges or in-head localization in headphone reproduction. One reason is that the dummy head does not produce the correct frequency structure because its spatial filter effect does not coincide with the individual Head Related Transfer Function of the listener.
Nature, can not be too narrow- auditory events for this type of localization. It is generally at a dummy head recording does not make sense to process the signals measured with an equalizer. By manipulation of the frequency response and determining the direction of the frequency bands are changed otherwise may, thus resulting in a greater localization defects.
The localization of focus is at unknown signals at around 17 degrees, with known signals at 9 degrees. These values are for looking to the future. The farther a signal emerges from the forward direction, the worse the localization accuracy.
In - head localization
The in-head localization is perceived as unpleasant effect which can occur especially in the headphone reproduction, but also in loudspeaker reproduction. The auditory events will no longer be located outside of the head, but in my head.
The brain compares the incoming ear signals known interaural signal differences and known spectral structures. If the signals have interaural time and level differences that are unknown in this combination or if the signals have a spectral characteristic which has never been seen with one's own outer ear, it can lead to in-head localization. For speakers this can happen in reverse polarity and when you put them in an angle greater than 90 degrees.
Such a localization may be simulated experimentally by. For a stereo signal, the two channels are identical, the amplitude values of the two reflected channels (inverted ), and the resulting sound signal to headphones
Diffuse-field equalization and free -field equalization
In the early days of this admission process all art heads were freifeldentzerrt. Early 80s developed the Institute for Radio Technology with the company the Neumann KU80 KU81 for whose only difference was the equalization of microphones. The KU80 was not suitable to reproduce the artificial head recordings through loudspeakers, which is why the free-field equalization used at that time was called into question.
Measurements in a free sound field (free field ) are performed without reflective acoustic boundary surfaces. These conditions are obtained with good approximation in an anechoic chamber. The resulting free field frequency response is only valid for a certain direction of sound incidence. As a flat frequency response is desired, one has to equalize the signal with the aid of a filter. When they are used in an artificial head pressure microphones, the diffuse-field frequency response is very different from the free field frequency response. This is because that when the pressure microphone for direct sound from the 0 ° direction meeting at a diaphragm diameter of approximately 18 mm, a level increase of 6 dB at 10 kHz occurs in the transmission level. This is caused by the sound waves, whose wavelength corresponds to the diaphragm diameter or smaller. They are reflected by the diaphragm and the sound pressure to the diaphragm thus doubled. In the diffuse field which leads to a waste height, as frequencies are diffracted with a smaller wavelength no longer the microphone capsule. This applies, however, only frequencies from the side or rear direction of sound incidence. For sound waves in a particular altitude range incident from the front of the membrane, that is, from the vicinity of the front-to- sound incidence direction, it comes to a level increase of 6 dB. Since the dummy head but is not intended for inclusion in the near field, but rather has a greater distance to the sound source, the diffuse-field frequency response plays a much larger role. The diffuse field is characterized by uniformly incident sound components from all directions. Are in the diffuse sound field not only a direction of sound incidence, as is the case in a free sound field.
When playing through headphones a diffuse-field headphones should also be used. The headphones with a flat Diffusfeldübertragungsmaß provide optimum sound neutrality. It would be better, headphones with a special IRT equalization.
Binaural recording technique
Without artificial head
In the simplest binaural recording method requires two microphones pointing sideways away from each other and at a distance ( microphone basis) of about 17 cm to 22 cm are to each other. Popular are the mysterious 17.5 cm. This ear distance and make the placement approximated the position of the ears ( ear canals ) of an average person represents a sound absorbent or reflective separator, such as a football or a metal plate is placed between the microphones to simulate approximately a head. A well-known version of this arrangement is the Jecklin disc OSS, an absorbing, 30 -cm-disc between two microphones with omnidirectional microphone at a base of 16.5 cm. The microphones are thereby turned slightly outward. In the scripts of Jecklin one that has recently become the enlarged disc diameter of 35 cm and the now even more than doubled microphone base of 36 cm.
This has not been generally noticed, because it still takes the obsolete 30 -cm disc dimensions and to small 16.5 cm microphone base.
Another head-like arrangement is the sphere microphone.
Binaural dummy head recordings
More sophisticated techniques consist of accurate head replicas by means of an artificial head. A typical binaural recording unit has two studio condenser omnidirectional microphones, which are inserted in the ear canal of the artificial head. Here are the compiled in psychoacoustic research communities head-related transfer functions HRTF simulated.
The first stereo dummy head with a reproduction of the human ear canal was built in 1933. The art heads KU -81 and KU -100 of the company Neumann are the most commonly used binaural microphones today. The Kemar system ( registered name: KEMAR ) is another alternative. The more involved HMS system of Head Acoustics ( registered name: HEAD acoustics ) from Duke near Aachen has an automatic frequency response setting and gives a better all-round impression. It is mainly used in the acoustic measurement technology.
The first broadcast on German radio play in artificial head Stereofonie was for radio exhibition in Berlin in 1973, the RIAS / BR / WDR production "Demolition " (The demolished man ) from the novel by Alfred Bester.
More binaural microphone system alternatives
There are also alternatives such as the ear plug with microphone that work similarly.
This " earbuds with microphone ", also known as "original head - microphone '(OMC ) is known, look like ordinary headphones, as they are known from a Walkman. They are provided with Elektretmikrofonkapseln omnidirectional. The microphones are fitted with 3.5 mm jack plugs and are thus suitable for mobile use, as you can record with standard MiniDisc players or DAT recorders.
One can get from the disadvantage that one is always bound to a human head, which is certainly never held very still, make an advantage. For radio plays choreographies are possible, which would not produce a conventional artificial head without further notice. Thus, one can not only make movements around the dummy head, but deliberately also include the dummy head in the piece. Many bootlegs of live concerts are produced with the help of these microphones, since the recording earbuds are very unobtrusive to carry and can provide useful results for headphone playback.
Playback
A binaural sound recording thus produced can be optimally learn " space " only when using headphones. Playback through speakers does not give that impression, but just a little " hollow" sounding stereo effect. Trying to diffusfeldentzerrtem artificial head to explain artificial head recordings for loudspeaker reproduction compatible, has not been accepted due to sonic weaknesses.
Listening with headphones results in a listening experience that can surpass the spatiality of the usual speaker sound image, as it allows a more precise picture of the binaural sound waves. Although the right-left localization is performed safely, identifying the top - down position of audio signals is difficult. The localization of a frontal sound event causes problems, since it seems moved at a certain angle in height and elevation ( audio engineering). Typically it is that actually existing front signals are often heard from behind rare reversed so that you honestly can not speak of a front / rear - interchange.
Loudspeaker reproduction
The loudspeaker reproduction is one of the biggest problems of artificial head recordings. When playing back the signals should be taken to ensure that both signals, so the right side and the left side, completely separate arrive at the two ears. That is precisely: The left ear may only receive signals from the left stereo channel and the right signals only the right channel. In headphone reproduction is of course the case, but at a normal stereo speaker placement is not. Furthermore, just adds that the respective reverberation times of the listening room be added to the already recorded. Against this you can not do much though.
For the problem of channel separation, there is a solution proposed by the Heinrich Hertz Institute in Berlin and the Institute of Technical Acoustics at the Technical University of Berlin. Here, as can be seen in the sketch, four speakers have been placed. The respective opposite Speakers ' working together '. That is, the rear speakers are each phase- rotated with a filter processed (height reduction) and each running a time delay so that a cancellation is caused. It is necessary at this speaker installation, ensure the correct listening position.
The second proposed solution is derived from the 3 Physics Institute in Göttingen. Here each ear is sonicated with a speaker. One uses here but not the usual stereo speaker placement, but the speakers are positioned next to the handset. It comes to level and time differences at the ears, and the frequency response is changed. All frequencies are in the range of head size and smaller are reflected at the top and no longer bent. The signals arriving at the other ear still be phase shifted through the resulting time differences and thereby possibly wiped out. This method also requires a firm adherence to the listening position and, moreover, is sensitive to head movements. If a listener does not comply with the optimal listening position and sits outside, he still gets a stereo - like sound transmission installation, but must be reckoned with colorations.
Applications
The dummy head is used everywhere in principle, where the aim is to get a more natural recording of what a listener would get to the point. Therefore, the artificial head is used for measurements, radio plays and, although rare, when recording music used. In these various fields of application, various artificial head models with different properties are used.
Measurements
Often when measuring ordinary measurement microphones are used, however, are not really good for many applications, as they do not take into account just the spatial impression, so also the human ear, when evaluating the acoustic signals. Studies involving human characteristics with are, can only be realized with the aid of the artificial head. For measurements in the industry of artificial head is often used as a kind of " dummy ", located in the immediate vicinity of explosions, accidents, noisy machines or the like is. The artificial heads for measurements are usually also associated with particular computer interface in order to evaluate the acoustic signals. Measurements also psychoacoustic parameters, such an expectation of the listener include are, it is usually evaluated in the context of listening studies.
Radio plays
Especially in the 70s of the artificial head was also used in creative radio plays. But even today you can still find occasional radio plays that were recorded with the artificial head, especially in the public service broadcasters. When radio plays you can clearly hear that the mere statement of the artificial head with entering into the action of the play and the fact emerging mood, which is also deliberately used.
Music productions
Only in very few musical productions, the artificial head technique. This is not due to the associated costs, but on the non-existent speaker compatibility. Artificial head recordings sound queasy discolored over loudspeakers, because the dummy head 's spectral differences even added his own ear signals, and the double and crosswise.
Basic distinction is made on a number of individual tracks and pure stereo recording when recording music from the usual in the rock and pop field "take -by -take process ." This procedure is also possible with artificial head recordings. You have before you start recording, so all the instruments are distributed around the artificial head, as they should be placed later in the sound. You can then record as usual succession on a multi -track recorder, the individual instruments. During later mixdown is just take care that you should not edit fully with an equalizer, the individual signals, as this affects the direction determining frequency bands that are important for the localization of the respective instrument. A Editing with the pan pot should be avoided. Limitations in the possibilities for post-processing are the main reason that the procedure could not become widely with today studio productions. Audio productions are now an art product that is intended to produce appealing perception on a home system, which does not necessarily match the original event. But the artificial head Stereofonie remains the only option, the sound field of the receiving space in all three dimensions of space, really spatially to reproduce, to this day next to Ambisonics and WFS Holofonie.