EP1318502A2 - Method for coding audio - Google Patents

Abstract

The method involves compressing several audio signals with a conventional compression technique by combining sound sources similar to a sound field and selecting compression factors so that the maximum permitted data rate is not exceeded. Information is added to the combined sound sources regarding the type of source and the object position with respect to the observer <??>AN Independent claim is also included for the following: an arrangement for implementing the inventive method.

Description

The invention relates to a method for audio coding according to the features the preamble of claims 1 and 14.

When playing movie sound and music signals, the To reconcile the image of the sound with the position of origin. With pure music signals, a good approximation is already carried out Achieved use of stereo speakers. If you want to Music pieces achieve a different spatial impression, for example one Jazz clubs, these two speakers are hardly enough. Especially at More channels are desirable because there are also acoustic effects come from behind, should be played.

One of the best-known methods for this is the so-called "Dolby Pro Logic" method, which is mainly used in film material to influence the localization of the sound. Speakers are usually mapped onto the center channel and noises can only come from the rear speakers.
There is also a whole class of methods used to simulate room acoustics. Commonly used names of such methods are "Hall", "Stadium", "Jazz", "Club" etc. With these methods optimized for music signals, it is not desirable to hear voice signals (vocals) only from the center speaker or a music signal only output from the rear speakers, which is possible using the "Dolby Pro Logic" method.
The Dolby Pro Logic follow-up process, which was called Dolby Pro Logic II, has a mode for music that takes these differences into account, in addition to the film mode.

A method for coding speech is known from EP 0 481 374 B1. Here a discrete reshaping of a speech window is carried out in order to to get a discrete spectrum of coefficients. An approximate one Envelopes of the discrete spectrum are found in any of a variety of Subbands calculated and for digitally coding the defined Envelopes of each subband used. Within subbands each scaled coefficient is divided into a number of bits with at least one converted by a variety of quantizers of different bit lengths. The quantizer used for each subband becomes for each speech window by calculating the allocation of bits as a number of bits greater than or determined to be zero, depending on a power density estimate for the subband and a distortion error estimate for the Language window.

EP 0 587 733 B1 describes a signal analysis system for filtering one or input signals representing several signals. There are Input buffer means for grouping the input samples into time domain signal sample blocks intended. The input samples are Analysis window weighted samples. In addition, means of analysis are Generation of spectral information in response to the time domain signal sample blocks available; where the spectral information Spectral coefficients, which are essentially an even-numbered stacked Time domain aliasing transformation applied to the Time domain signal sample blocks. With the spectral coefficients are essentially coefficients of a modified discrete Cosine transformation or coefficients of a modified discrete sine transformation. The analysis means comprise forward-forward transformation means to generate modified sample blocks and Forward transforming means for generating frequency domain transform coefficients.

From EP 0 664 943 B1 is a coding device for adaptive processing of Audio signals for coding, transmission or storage and Recover known, the noise level with the Signal amplitude level fluctuates. It is a processing facility available, which responds to input signals such that they either a first and a second signal or the sum and difference of the first and outputs second signal. The first and second signals correspond to the two matrix-coded audio signals of a four to two audio signal matrix, the Processing device also generates a control signal which indicates whether the first and second signal or the sum and difference of the first and second signal is output.

EP 0 519 055 B1 describes a decoder consisting of a receiving means to receive a plurality of information channels formatted delivery channels, Deformatting means for generating in response to the receiving means, a deformed representation depending on each delivery channel, and Synthesis means for generating output signals depending on the deformed representations, known. Between the means of deforming and the synthesis means are arranged distribution means which on the Deformers address and one or more intermediate signals generate, at least one intermediate signal by combining the Information is generated from two or more of the deformatted representations. The synthesis means generate a respective output signal in response to each of the intermediate signals.

From EP 0 520 068 B1 is an encoder for coding two or more Audio channels known. The encoder has a subband device Generation of subband signals, a mixing device for creating a or more composite signals, and means for generating Control information for a corresponding composite signal. In addition, the encoder has an encoder for generating encoded Information by assigning bits to the one or more composite signals. It is still a formatter for assembling the coded information and the control information an output signal available.

A speech encoder is known from EP 0 208 712 B1. This speech encoder contains a Fourier transformer for executing a discrete one Fourier transform of an incoming speech signal for generation of a discrete transform spectrum of coefficients, one Standardization device for modifying the transformation spectrum for Generation of a standardized, flatter spectrum and for coding one Function that modifies the discrete spectrum. Besides, one is Device for coding at least a part of the spectrum is available. The normalization device has a device (44) for defining the approximated envelopes of the discrete spectrum in each of several Subbands of coefficients and for coding the defined envelope of each subband of coefficients and scaling devices each spectrum coefficient relative to the defined envelope of the relevant subband of coefficients.

As another well known method to improve this problem Dolby Digital (also known as AC3) and DTS introduced which play 6 different channels allow. Still a right and a left rear separate channel possible, this was still not enough for a perfect reproduction, Therefore, these methods were expanded again by matrixing so that a additional rear center channel is possible. With DTS-ES DISCRETE then dispensed with the matrixing and a real seventh audio channel transfer. THX-EX now generates 8 channels and AAC are even over 100 separate audio channels possible.

ISO / IEC JTC 1 / SC 29 / WG11 (MPEG 4 Structured Audio) is a procedure known which for the synthetic generation of music signals the Offers the artificially created musical instruments a Assign position information and one or more effects. It is also possible to integrate samples in this way.

A disadvantage of MPEG 4 Structured Audio is that it is for real-time transmission over a channel with a limited data rate is not suitable because there is no upper one Limitation for the data rate and the computing power is provided. Also the Limitation to samples, which are usually only short pieces of sound correspond and that already corresponds to the playback device at the time need to be completely known when their play starts poses a problem Synchronization is also difficult, should start with the playback at a certain point so it may be necessary to open the file to be calculated completely from the start. An application in film material also contradicts the lack of compression of the samples, which thereby needlessly a lot of data storage. Specifying the angle of the Sound sources alone are also not enough to always ensure a good allocation of the Sound sources to reach the image material, because the image material through different projection techniques different sizes and distances can have to the observer. If a speaker can be seen on the left edge of the picture, so its language should also come from the left speakers in a cinema are reproduced in a representation over a 5 meter distance Televisions with 72 cm screen diagonals make more sense, however Output language on the middle speaker, because then there too the speaker appears to be.

A disadvantage of all known inventions with regard to Audio data compression is that due to the fixation of the positions of the Sound sources have a high number of audio channels to emulate the desired sound field is required. This causes a high one Amount of data, both for storage media and for transmission is undesirable. There is also the possibility of replicating Room acoustics limited by fixing the speaker positions.

For example, Dolby Digital and DTS do not have ceiling speakers provided that could emulate sounds coming from above.

The object of the invention is to provide a method which is efficient Coding of multiple audio signals taking into account the associated sound field.

The present invention solves this problem by the features of Claims 1 and 14. Advantageous refinements and developments of Invention are in the dependent claims, the accompanying description indicated in addition to Figure 1.

The present invention solves the problem in that the one Sound sources belonging to the sound field first individually with a conventional one Compression methods, e.g. AC3, MP3, AAC, WMA, etc. be compressed. If a maximum allowed is exceeded Data rate becomes a summary of sound sources or a stronger one Compression performed to reduce the data rate. Each of these Sound sources are information about their type as well as a Position information added, indicating where the object is in relation to the viewer. Information about the properties of the the room to be reproduced and the current horizontal and vertical Viewing angles are transmitted. Based on the position information and information about the sources, the information about the properties of the room to be reproduced, the current horizontal and vertical Viewing angles as well as the size and position of the image of the Playback device are then postprocessing and images of the individual sources to the existing speakers or a headphone signal carried out.

If one looks at natural sound fields, one finds that the human Heard this analyzed in such a way that a division into different Sound sources and their characteristics take place and also the Property of the room in which the acoustic reproduction takes place, is evaluated. The distribution of the sound sources is primarily done by Evaluation of the spectral composition and the position of the Sound source. Runtime differences and the directional The frequency sensitivity curve of the hearing has only a minor influence on the division. A good approximation of many sound fields can can be achieved if two voice channels and two effect channels are used To be available. It allows communication between two people as well various ambient noises (e.g. passing cars) recorded and be replicated.

According to the invention, the audio data generated by different sound sources are encoded in such a way that sound sources are combined that have the same characteristics for the hearing and then this with position information, information about the type of channel (vocals, background noise, speech ... ) and information about possible effects (reverb, dynamic compressor). This information is transmitted continuously. During playback, the type of channel can be used for post-processing in which this signal is decorrelated several times for background noise and output to the loudspeakers. An individual post-processing algorithm is assigned to each signal type. An advantageous implementation results if, in addition to each sound source, information about its radiation characteristics, such as spherical, kidney-shaped, etc., is also transmitted. A further advantageous implementation results if additional information about the desired room characteristics (e.g. bathroom, cathedral, etc.), parameters (e.g. reverberation time) or directly algorithmic descriptions to achieve these sound impressions as well as properties of the individual sound sources (such as language or music or effect) are also transmitted. The room geometry and the room absorption properties are to be transferred as parameters for the description of the room characteristics. A living room with curtains and carpets absorbs the sound much more than a bathroom or a church.
If a playback device is not powerful enough to evaluate this data, it ignores this additional data during playback.
A further advantageous embodiment of the invention results if properties are additionally transmitted via the viewing angle, ie that the angles of the sound source and these viewing angles can be used to directly determine whether the associated object is visible in the image. This is the case if the horizontal angle of the object is within the horizontal viewing angle range and the vertical angle of the object is within the vertical viewing angle range. The viewing angle, like the position of the objects, is related to the viewer of the original scene. The viewing angle can change continuously, which is why it is advantageously transmitted to each individual image. Based on the angle of the object, the viewing angle and the position and size of the projected image, a virtual angle is then determined based on the listening position. In order that there is no abrupt jump when leaving objects from the viewing angle range, a non-linear mapping is advantageously used there. For example, the angular range shrunk on small screens can be compensated for by stretching the remaining angular range. An advantageous embodiment results if the strength of the compression / extension can be adjusted by the user.

• eine wesentlich realistischere Audiodarbietung erreicht;
• die Ausnutzung der vorhandenen Kanäle wesentlich verbessert, somit wird bei DVDs Kapazität und bei Broadcast-Verfahren Bandbreite gespart;
• bei Hinzufügung der Klangeffekte im Endgerät bei der Codierung der unverfälschten Originalsignale eine wesentlich geringere Datenrate benötigt, dies gibt eine zusätzliche Datenratenersparnis;
• die Kanalabbildung wesentlich verbessert, ein überfliegender Hubschrauber kann mit nur einem Kanal perfekt nachgebildet werden;
• das System abwärtskompatibel gehalten, die Wiedergabe anderer Standards kann durch Annahme der Positionen der Klangquellen auf die dabei üblichen Aufstellungspositionen der Lautsprecher erfolgen; es muss die vom Kunden gewählte Lautsprecherkonstellation nicht modifiziert werden;
• die Aufstellungspositionen der Lautsprecher nicht vorgegeben und kann beliebig sein, die Lautsprecher müssen dem System nur bekannt gemacht werden; das System berechnet dann, auf welchen Lautsprechern es bestimmte Kanäle am geeignetesten ausgibt; so ist es möglich, die räumlichen Gegebenheiten perfekt auszunutzen;
• eine bessere Anpassung an verschiedene Bildformate bei der Wiedergabe erreicht, Klangquellen zu im Bildbereich gehörenden Objekten können klanglich auch dorthin fokussiert werden.

Außerdem ist die Anzahl an aufgestellten Lautsprechern theoretisch unbegrenzt, sie ist nur begrenzt durch die Anzahl an Verstärkerkanälen und durch die Rechenleistung des Systems. Ein Nutzer kann nachträglich ohne großen Aufwand und Probleme zwei seitliche Lautsprecher nachrüsten oder einen Deckenlautsprecher entfernen.
Schließlich ist die zur Übertragung der Zusatzinformationen benötigte Datenmenge gegenüber der Datenmenge zur Codierung der einzelnen Audiokanäle vernachlässigbar gering.Through the invention

• achieved a much more realistic audio performance;
• the utilization of the existing channels is significantly improved, thus saving capacity for DVDs and bandwidth for broadcast processes;
• when adding the sound effects in the terminal when coding the unadulterated original signals requires a much lower data rate, this gives an additional data rate saving;
• the canal image significantly improved, a flying helicopter can be perfectly replicated with just one canal;
• the system is kept downward compatible, the reproduction of other standards can be done by assuming the positions of the sound sources on the usual placement positions of the speakers; the speaker constellation chosen by the customer does not have to be modified;
• the installation positions of the loudspeakers are not specified and can be arbitrary, the loudspeakers only have to be made known to the system; the system then calculates on which speakers it most suitably outputs certain channels; this makes it possible to make perfect use of the spatial conditions;
• a better adaptation to different image formats achieved during playback, sound sources for objects belonging to the image area can also be focused there.

In addition, the number of speakers installed is theoretically unlimited, it is only limited by the number of amplifier channels and the computing power of the system. A user can retrofit two side speakers or remove a ceiling speaker without much effort and problems.
Finally, the amount of data required to transmit the additional information is negligibly small compared to the amount of data for coding the individual audio channels.

The invention is described below on the basis of a specific exemplary embodiment shown in FIG.
The exemplary embodiment shows a device according to the invention. The device according to the invention has the audio signal inputs A1 to An. Information on the position and properties of the corresponding audio signal is transmitted to the n audio signals via the inputs I1 to In. The horizontal and vertical viewing angles and the property of the room to be simulated are fed in via the SF input. The incoming audio signals are compressed in the devices K1 to Kn in the data rate to the signals C1 to Cn. The data streams C, I and SF are then mixed together in the device MX to form a single data stream US. The data stream US can then either be transmitted in broadcasting format or can also be temporarily stored on a storage medium. For output, this data stream is then sent to a device DX, which regenerates the data streams C1 to Cn and generates ISF1 to ISFn. ISFi corresponds to the data stream li, to which SF is added. The compressed data streams C1 to Cn are decompressed in the devices D1 to Dn. The individual decompressed data streams are then each given to the devices R1 to Rn together with their associated data stream ISFi and the signal BP. The signal BP corresponds to information about the image format, the position of the playback device and the parameters selected by the customer for playback. The decoded audio signals are then mapped in the devices Ri to the sum signals S1 to Sk, taking into account the parameters present in BP and ISFi. The number of sum signals k corresponds to the number of loudspeakers installed by the customer. The sum signals S generated by the individual devices Ri are summed and then output to the loudspeakers L1 to Lk.

LIST OF REFERENCE NUMBERS

A

Audio signal input

BP

Information about picture format, position of the sound source and customer parameters

C

Compressed audio signals

D

Device for decompressing individual audio signals

DX

Device for splitting the data stream

I

Input for information about an audio signal

ISF

Information about an audio signal and information about the Viewing angle of the sound source

K

Device for compressing individual audio signals

L

Speaker output

MX

Device for mixing individual data streams

R

Device for mapping an audio signal onto the Loudspeaker signals

S

Sum channels for the speakers

SF

Information about the viewing angle of the sound sources

US

transmission path

Claims (14)

Method for coding audio signals, wherein a plurality of audio signals are compressed using a compression method known per se,
characterized in that
Similar sound sources belonging to a sound field are summarized and the compression factors are selected so that the maximum permissible data rate is not exceeded,
For each of these summarized sound sources, information about the type of the source and position information, which indicates where the object is in relation to the viewer, is added. Method according to claim 1,
characterized in that
Information about the properties of the room to be simulated as well as information about the current horizontal and vertical viewing angles are inserted into the generated data stream. Method according to claim 1 or 2,
characterized in that
The playback device evaluates the size and position of the image projection and, based on it, the properties of the room to be reproduced, the information about the current horizontal and vertical viewing angles as well as the position information and information about the types and abstract properties of the sources, the sound sources are mapped onto the existing loudspeakers becomes. Method according to one of claims 1 to 3,
characterized in that
for coding, the AC3 method, the DTS method or the MP3 method, or the AAC method, the WMA method or a similar method is used. Method according to one of claims 1 to 4,
characterized in that
the information about the type of source is given in the form of algorithms or parameters for given algorithms. Method according to one of claims 1 to 4,
characterized in that
the information about the properties of the room to be simulated is given in the form of algorithms or parameters for specified algorithms. Method according to one of claims 1 to 6,
characterized in that
playback via headphones or speakers. Method according to one of claims 1 to 7,
characterized in that
each sound source can be assigned information about its radiation characteristics. A method according to claim 8,
characterized in that
the radiation properties are spherical, conical, flat or kidney-shaped. Method according to one of claims 1 to 9,
characterized in that
in the playback device, the properties of the listening room, ie the room in which the loudspeakers are set up, are adjustable and / or can be determined by measurements and these are taken into account during playback. A method according to claim 10,
characterized in that
the properties of the listening room are determined by the geometry of the room and the nature of the walls and floor. Method according to one of claims 1 to 11,
characterized in that
any number of speakers can be used for playback. Method according to one of claims 1 to 9,
characterized in that
the speaker positions are made known to the playback device so that the best possible mapping of the individual channels onto the speakers is carried out. Device for performing the method according to one or more of the preceding claims.

Images