Data processing: speech signal processing – linguistics – language – Audio signal bandwidth compression or expansion – With content reduction encoding
Reexamination Certificate
2000-06-26
2003-02-25
Knepper, David D. (Department: 2641)
Data processing: speech signal processing, linguistics, language
Audio signal bandwidth compression or expansion
With content reduction encoding
C704S203000, C704S205000
Reexamination Certificate
active
06526384
ABSTRACT:
BACKGROUND OF THE INVENTION
The invention is directed to a process and a device for generating a bit rate scalable audio data stream. The invention is applicable in the field of data communications, particularly in the field of audio data communications.
A known problem in the field of data communications consists in that the data signal to be transferred is present in a data source with a high data rate, for example, at 64 kbit per second, but the data channel available for transfer or processing can only transfer the data to be transferred at a lower data rate, for example, at 32 kbit per second. In this case, the data must first be decoded at the higher data rate and then coded again at the lower data rate. This brings about a high expenditure on hardware technology and computing, especially because the data rate in modern data networks is not constant but variable and is adapted to the particular load situation of the data network. Compared to this, it would be more beneficial if a bit rate scalable data stream were to be supplied, of which only a part of the available data bits is transferred depending on the data rate offered by the transfer channel. Corresponding processes for generating bit rate scalable audio data streams are being undertaken across the world at the present time, in particular within the framework of efforts made towards standardization, for example, within the framework of the MPEG 4 (Moving Picture Expert Group) standardization. In particular, the CODECs (COder/DECoder) developed within the framework of the MPEG4 standardization must guarantee the functionality of the bit rate scalability.
A process for the generation of a bit rate scalable audio data stream in which the audio data stream is compressed in a core codec accompanied by determination of parameters is known from WO 97 159 83A. The coding is enhanced in subsequent enhancement stages. The enhancement stages are controlled in dependence on the core parameters.
The invention is therefore based on the problem of providing a process and a device for generating a bit rate scalable audio data stream which can be used in a versatile manner, which ensure a good transfer quality even when the available transfer rate is low, and which achieve a high degree of flexibility with respect to the available transfer rate in an economical manner.
In accordance with the present invention a process for generating a bit ratio scalable audio data stream includes: compression of the audio data stream in a core codec accompanied by the determination of core parameters; and enhancement of the coding in at least one downstream enhancement stage, wherein the enhancement in the enhancement stage is controlled by the core parameters, wherein the audio data stream is frequency-transformed, wherein a synthesized audio signal produced by the core codec is likewise frequency-transformed, and wherein the frequency-transformed synthesized audio signal is combined with the frequency-transformed audio data stream.
For a process for generating a bit rate scalable audio data stream with the step of compression of the audio input data stream in a core codec accompanied by determination of core parameters and the step of enhancement of the coding in at least one downstream enhancement stage, the problem is solved in that the enhancement in the enhancement stage is controlled by means of the core parameters. For the process according to the invention, the core codec forms the core assembly and codes the arriving input data stream at a low bit rate of, for example, 2, 4 or 6 kbit per second. The core codec is followed by any number of improvement or enhancement stages, as they are called, which code at a data rate of 1, 2, 3 or 4 kbit per second depending on application. An advantage of this process consists in that omission of any one enhancement stage has no effect on the other parts of the bit stream. It is an essential condition that the provided transfer system guarantees at least the bit rate of the core codec. The core codec parameterizes the incoming audio signal and determines, for example, parameters like pitch, voiced/unvoiced sounds, or the volume of sound. A core codec according to ITU-T G.723.1 (ITU, International Telecommunication Union), for example, can be used. It is particularly advantageous in the process according to the invention that the core parameters determined by the core codec control the subsequent enhancement stages because a considerable increase in the efficiency of the enhancement stages is achieved in this way.
In a particular embodiment type of the invention, the process is characterized in that a vector coding is effected in the enhancement stage and in that the core parameters control the selection of code books. This is advantageous because the code books for vector coding used in periodic audio segments are different from those used in non-periodic audio segments. Also, the energy parameters of the core codec are used directly for the coding of the signal energy (volume of sound), which results in a considerable bit rate saving. The use of core parameters is possible because they have to be transferred to the receiver in any case.
In a particular embodiment form, the process is characterized by the steps of transforming the audio input data stream, transforming a synthesized audio signal generated by the core codec, and a combination of the transformed synthesized audio signal with the transformed audio data stream. In this way, the difference between the audio signal compressed by the core codec and the original signal is determined advantageously in an economical manner and with high precision. In the simplest case, the combination can be a subtraction; however, more complex operations such as adapting the core spectrum for better matching with the original spectrum can also be involved. In this latter particular embodiment of the invention, the combination parameters for the adaptation must be transferred to the receiver.
In a particular embodiment of the invention, the process is characterized in that the core codec divides the input signal into at least two subframes; in that the transformation is a frequency transformation running synchronous to the subframe of the core codec; in that, by means of the frequency transformation, a transformation is effected per subframe which generates a spectrum vector in each instance; in that each spectrum vector is divided into at least two partial vectors corresponding to two partial bands; and in that each enhancement stage enhances one of these partial bands. The utilized frequency transformation and the dividing into partial bands has the advantage that the process according to the invention not only enables a high efficiency for the bit rate scaling according to objective criteria, but also takes into consideration subjective criteria such as the acoustic and physiological boundary conditions of human hearing. A resource allocation unit determines which partial band is to be enhanced. As has already been mentioned, this determination can be effected using a psycho-acoustic model which determines which frequency bands are subjectively important, or by measurements of signal-to-noise ratios, for example.
In a particular embodiment type of the invention, the process is characterized in that, for each enhancement stage, one set of parameters and one address for the enhanced partial band are transferred. Since the allocation unit enhances the partial bands in the order of their importance, this embodiment of the invention has the advantage that the enhanced bits are stored in the bit stream in the order of their importance. Since every stage has been provided with an address, allocation can be effected in the receiver without problems and in a manner that is reliably correct. Scaling is now made possible without problems and without any additional expenditure merely by suppressing a corresponding number of enhancement stages, beginning with the last, least significant stage. It is further advantageous that this scaling can be effected at any point on
Mueller Joerg-Martin
Waechter Bertram
Knepper David D.
Siemens AG
Striker Michael J.
LandOfFree
Method and device for limiting a stream of audio data with a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for limiting a stream of audio data with a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for limiting a stream of audio data with a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3144624