Data processing: speech signal processing – linguistics – language – Audio signal bandwidth compression or expansion
Reexamination Certificate
2000-07-14
2004-03-23
McFadden, Susan (Department: 2654)
Data processing: speech signal processing, linguistics, language
Audio signal bandwidth compression or expansion
C381S332000
Reexamination Certificate
active
06711546
ABSTRACT:
The invention relates to a coding scheme for digital audio signals which is especially suitable for wireless transmission applications.
BACKGROUND OF THE INVENTION
Data compression techniques have been developed which reduce the bandwidth for transmission of digital audio signals. Examples of widespread conventional digital audio coding schemes are parametric coding (e.g. linear predictive coding), waveform coding (e.g. NICAM), and transform coding (e.g. MPEG-4). These techniques generally involve numerically complex decoding algorithms, especially so for techniques applicable to so-called multimedia audio, that is audio that may include music or background sounds, not only speech. To implement these known algorithms, a receiver terminal needs to be equipped with a complex decoder that increases the power consumption of the terminal.
A simple compression technique used for voice communications over a wireless link takes advantage of the fact that greater than 50% of human speech comprises silence. During the silent periods a special code is transmitted called the “silence descriptor” which serves to reduce the average power consumption of a mobile telephone unit. However, this technique is not especially useful for music, since music is generally characterised by long periods of uninterrupted sound, and hence data transmission. Multimedia audio compression techniques applicable to digital audio signals with a music content, such as MPEG-4, are generally numerically quite complex.
One multimedia audio application of interest is a hand-held video terminal, for example a video telephone, a hand-held television, or a hand-held video disk player. With such a terminal, the user may wish to view a video track while listening to the music sound track. Viewing will take place by holding the terminal in front of the user at arms length. The proximity between terminal and the ear used in a pure audio terminal such as a telephone is lost. The audio may then be played through a speaker in the terminal or more likely will be transmitted to an ear piece or to headphones for private listening. In the latter case, it is likely to be impractical for commercial and other reasons to transmit the audio in conventional compressed form from the hand-held video terminal, as this would require complex decompression electronics to be housed in the ear piece or headphones.
It is therefore an aim of the invention to provide a coding scheme which allows data compression of a digital audio signal, but which is not associated with complex digital signal processing, especially at the receiver.
SUMMARY OF THE INVENTION
The coding scheme of the invention is a differential coding scheme that exploits an inherent property of unary digital coding, namely that all data elements within a unary number carry equal weight. Unary digital coding is a simple tally scheme in which the information content of each sample of a signal can be expressed merely as a positive integer or tally number.
A coding method according to one aspect of the invention is applied to a digital audio signal comprising a succession of absolute amplitude values associated with respective sample periods of the digital audio signal. The digital audio signal may be a conventional digital audio signal expressed in conventional binary format, for example in a signal which is pulse code modulated, with the amplitude value for each sample being represented by a group of “n” data bits for an n-bit digital audio signal.
Data compression is implemented by calculating the numerical difference between adjacent ones of the absolute amplitude values, thereby to convert the succession of absolute amplitude values into a corresponding succession of difference amplitude values. In the compressed digital audio signal, the difference amplitude values are expressed by a signed number of unary data bits. A sign bit is used since the difference amplitude value represents change in sound amplitude and thus may be negative or positive. In an embodiment of the invention, the difference amplitude values are represented by a sign bit and a plurality of data bits, and conveyed by a signal in which the data bits are grouped together into a single block, the width of which block varies with difference amplitude value. This may be considered to be a form of pulse width modulation in that the block width varies, although the width of each individual pulse in the signal is fixed. Unary coding of the data bits of the compressed signal allows the bit rate to be reduced in comparison to binary coding, since the data bits of each sample can be grouped together into a single block, which is thus amenable to low bit rate transmission, for example through pulse width modulation optionally in combination with a rate conversion to a fraction of the sample frequency.
Further aspects of the invention relate to a corresponding decoding method and apparatuses for implementing the coding and decoding methods, as well as loudspeaker and audio recording devices incorporating such apparatuses.
The bandwidth required for transmitting each sample of a digital audio signal in unary form can thus be reduced from the full dynamic range of the signal (e.g. 2
16
) to a small fraction of the dynamic range represented by the numerical difference between the tally numbers of adjacent samples in the digital audio signal. Data compression can be achieved in this way because of the strongly continuous nature of most audio signals, which results in the average numerical difference between the tally numbers of adjacent unary samples being orders of magnitude smaller than a mid-point value in the dynamic range (e.g. 2
15
).
The differential unary coding scheme thus takes advantage of the time-continuous properties of music, or other content-bearing audio signals, to reduce the data transmission rates without the need to apply any computationally intensive algorithms using large amounts of digital signal processing. By contrast, existing coding schemes developed for conventional digital audio, such as those mentioned further above, are based on algorithms that require considerable digital signal processing power to implement, both at the transmitter and receiver.
The differential unary coding scheme for data transmission is fully compatible with digital loudspeakers based on unary principles in that a digital audio signal coded in this way is convertible into a drive signal for a unary digital loudspeaker with very simple electronics, the invention now being further explained, by way of example only, with reference to a unary digital loudspeaker.
A unary loudspeaker comprises an array of acoustic output transducers or speaker elements, as described in PCT/GB96/00736. To reproduce a signal with n-bit digital audio quality, it will generally be necessary for 2
n
−1 or 2
n−1
speaker elements to be provided. Generally, a very large number of speaker elements will be required for reasonable reproduction quality. For example, 16-bit quality will necessitate provision of something of the order of 2
16
or 2
15
speaker elements.
A first basic principle of a unary loudspeaker is that all of the speaker elements are identical, since all carry equal significance. Moreover, for this reason, there is no information contained in the unary signal that defines the specific speaker element that needs to be activated. Addressing in the conventional sense thus does not arise and is not required.
A second basic principle of a unary loudspeaker is that the audio output volume is dictated by the number of speaker elements that are activated at a given point in time. The greater the number of active speaker elements the greater the instantaneous amplitude of the sound pressure wave. Based on the fact that music is generally a time-continuous signal, one can assume that the sound pressure wave at any given point in time is dependent on the sound pressure wave in the previous time period. For example, the sound of a guitar comprises a fluctuation in a sound pressure wave which continues over a period of time as the string
Brady III W. James
McFadden Susan
Petersen Bret J.
LandOfFree
Unary coding scheme for digital audio signals does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Unary coding scheme for digital audio signals, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Unary coding scheme for digital audio signals will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3200849