Data processing: speech signal processing – linguistics – language – Audio signal bandwidth compression or expansion
Reexamination Certificate
1998-10-16
2001-04-10
Hudspeth, David (Department: 2641)
Data processing: speech signal processing, linguistics, language
Audio signal bandwidth compression or expansion
C704S201000
Reexamination Certificate
active
06216107
ABSTRACT:
FIELD OF THE INVENTION
The present invention is directed toward an enhanced half-rate encoding and receiving apparatus and method and, more particularly, toward a half-rate encoding and receiving apparatus and method for transmitting and receiving half-rate encoded speech at a rate normally associated with full-rate encoding.
BACKGROUND OF THE INVENTION
In typical U.S. digital cellular telephone systems, analog voice signals are sampled and converted to a digital bit stream. In order to save bandwidth and therefore provide economic advantage, the digital bit stream is compressed by a source encoder before transmission across a radio channel. Generally, cellular systems provide for either full-rate or half-rate encoding of voice signals.
Full-rate source encoding in a typical GSM (Global System for Mobile Communications) system utilizes an LPC (Linear Prediction Coding) encoder with long-term prediction and regular pulse excitation. The output of the LPC encoder is generally 260 bits every 20 milliseconds, to give an encoding rate of 13 kbits/s. The 13 kbits/s signal output by the LPC encoder is input to a signal expander, for example, a channel encoder, which adds redundancy to the bit stream and thereby increases its rate to 22.8 kbits/s. The purpose of the redundancy is to minimize the consequence of bit errors that are often induced by a noisy radio channel.
Half-rate GSM source encoding utilizes a VSELP (Vector Sum Excited Linear Prediction) encoder whose output is generally 112 bits every 20 milliseconds, to give an encoding rate of 5.6 kbits/s. The 5.6 kbits/s signal output by the VSELP encoder is expanded by a channel encoder, which adds redundancy to the bit stream and outputs 228 bits every 20 milliseconds, thus increasing the bit rate from 5.6 kbits/s to 11.4 kbits/s. Again, the purpose of the added redundancy is to minimize transmission errors.
Half-rate encoding effectively doubles the capacity of a cellular system. Accordingly, established cellular standards provide half-rate capability as an economic benefit (twice as many revenue generating users). However, there is some penalty to be paid in return, namely, lower bit rate source encoding, e.g., half-rate encoding, goes hand-in-hand with reduced audio fidelity.
The above comparison between full-rate source encoding and half-rate source encoding assumes that the RF (Radio Frequency) channel carrying the encoded signals does not introduce transmission errors beyond the correction capability of the receiver receiving such signals. In general, this is a good assumption because cellular systems are typically designed to provide a relatively high SNR (Signal-to-Noise Ratio), and therefore a relatively low BER (Bit-Error-Rate). However, under conditions of low SNR, a low-bit-rate encoded signal, whether half-rate or full-rate, can suffer severely degraded audio quality.
This breakdown in performance is important in practice, as commercial applications are being developed where a radio system, conforming to an established cellular-radio air-interface standard, fails to provide the relatively high SNR on which the half-rate/full-rate trade is premised, or which fails to provide adequate SNR for the effective operation of either full or half-rate encoding. Such commercial applications include, but are not limited to, satellite communication systems with a cellular-standard air-interface which inherently has low link margins, as well as extended-range cellular systems that provide marine telephone service or telephone service capable of penetrating an office building, or the like. Moreover, cellular systems often experience episodes of significantly reduced SNR caused by channel fading and shadowing, which may result in degraded audio quality.
The present invention is directed toward overcoming one or more of the above-mentioned problems.
SUMMARY OF THE INVENTION
In a transmitter for transmitting communication signals across a radio channel, an improved encoder is provided including a half-rate encoder receiving a digitized speech signal and generating a compressed bit stream at half-rate, and a signal expander receiving the compressed bit stream and generating an expanded bit stream at full-rate for transmission across a radio channel.
In one form of the improved encoder, the full-rate is approximately 2× the half-rate.
In another form of the improved encoder, the signal expander includes a repeater repeating each bit in the compressed bit stream to generate the expanded bit stream.
In another form of the improved encoder, the signal expander includes a repeater repeating the compressed bit stream to generate the expanded bit stream.
In another form of the improved encoder, the compressed bit stream includes bits classified as one of critical, important and unimportant. The signal expander includes a plurality of encoders additionally encoding the compressed bit stream according to bit classification to generate the expanded bit stream.
An improved receiver, according to a first embodiment, is provided for receiving a digitized speech signal transmitted at a full-rate across a radio channel in a wireless communication system, the digitized speech signal selected from the group consisting of (a) a full-rate encoded digitized speech signal including a stream of binary bits, and (b) a half-rate encoded digitized speech signal including a stream of binary bits expanded for transmission at the full-rate by repeating each bit in the binary bit stream, the improved receiver including a full-rate equalizer, a half-rate equalizer, a switch initially routing the received digitized speech signal to the full-rate equalizer, wherein the full-rate equalizer demodulates the received digitized speech signal producing a full-rate demodulated signal and dibits of information corresponding to the full-rate demodulated signal, and an analyzer analyzing the dibits of information, the analyzer controlling the switch to route the received digitized speech signal to one of the full-rate and half-rate equalizers based upon said analysis.
In one form of the improved receiver, the analyzer includes an XNOR gate receiving the dibits of information, the XNOR gate outputting a logical one if the bits of the dibit are the same and a logical zero if the bits of the dibit are different, a counter receiving the output of the XNOR gate, the counter counting the occurrences of logical ones at the output of the XNOR gate, and a threshold detector connected to the counter, the threshold detector controlling the switch to route the received digitized speech signal to the half-rate equalizer if the number of logical ones counted by the counter exceeds a threshold value.
In another form of the improved receiver, the dibits of information include dibits of soft information, each soft dibit including soft values. The analyzer includes a multiplier multiplying the soft values of the soft dibits together, the multiplier outputting a positive value if the soft values of the soft dibit are of like polarity and negative value if the soft values of the soft dibit are of different polarity, a summer receiving and summing the output of the multiplier, and a threshold detector connected to the summer, the threshold detector controlling the switch to route the received digitized speech signal to the half-rate equalizer if the summed value exceeds a positive threshold value.
In another form of the improved receiver, the dibits of information include dibits of soft information plottable on a differential constellation having real and imaginary axes, each plotted soft dibit representing a complex value of a differential symbol. The analyzer includes a rotator rotating the differential symbols by &pgr;/4, the rotated differential symbols having components on the real and imaginary axes, a summer summing magnitudes of the rotated differential symbol components on the real and imaginary axes and calculating a ratio of real axis summed magnitudes versus imaginary axis summed magnitudes, and a threshold detector connected to the summer, the threshold detector controlling the
Fulghum Tracy
Khayrallah Ali S.
Koilpillai R. David
Rydbeck Nils
Azad Abul K.
Ericsson Inc.
Hudspeth David
Wood Phillips VanSanten Clark & Mortimer
LandOfFree
High-performance half-rate encoding apparatus and method for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with High-performance half-rate encoding apparatus and method for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and High-performance half-rate encoding apparatus and method for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2507607