GB2125255A - Digital data coding - Google Patents

Abstract

In a near-instantaneous companding system operating on a stream of 14-bit words in which blocks of 32 words are examined to determine the maximum word value in the block, and 32 10-bit words are formed for each block from the input words subject to a scaling factor dependent upon the value of the maximum sample in the block and these words are output with the scaling factor which takes one of five values, auxiliary digital data can be transmitted using the time period occupied by selected blocks which are discarded. The auxiliary data is transmitted with a dummy scaling factor to identify it. When the transmitted signal is an audio signal, the selected block can correspond to quiet or silent passages in the signal. At the receiver these blocks can be replaced in the audio signal by a repeated mid-range sample or a low- level synthetic noise signal, or by a selected one of several possible substitute signals identified by a code sent in the auxiliary data or by the use of a corresponding number of different dummy codes.

Description

SPECIFICATION Digital data coding This invention relates to digital data coding, or more particularly to a companding system of the type known as near-instantaneous companding, and is particularly suitable for use with digital audio signals.

A paper by C. R. Caine, A. R. English and J. W.

H. O'Clarey entitled "NlCAM 3: nearinstantaneous companded digital transmission system for high quality sound programmes" published in The Radio and Electronic Engineer Vol. 50, No. 10 pages 519-530, October 1980, describes near-instantaneous companding and in particular a system referred to as NICAM 3 (Near Instantaneously Companded Audio Multiplex, Mark 3). Other systems are also referred to and further described in the references to that paper.

In near-instantaneous companding, blocks of samples are taken and examined to discover the maximum sample value in the whole block. The coding system for the whole block is then set according to the amplitude of that maximum sample. For example, with audio signals having 14 bits per sample, 32 samples may be taken together to form one block. This corresponds to 1 ms at a 32 kHz sampling rate. The samples are all transmitted with 10 bit accuracy, but are first subjected to scaling by a factor dependent upon the value of the maximum sample in the block.

The possible factors are 20(=1), 21(=2), 22(=4), 23(=8), and 24(=16), corresponding to a 'shift' of 0, 1, 2, 3 or 4 bits. If the value of the maximum sample in the block of 32 samples is greater than half the maximum possible sample value which can be encoded, then the scaling factor is 2 , i.e.

the samples are unchanged. The first ten bits of each sample are thus transmitted. If the value of the maximum sample in the block lies between one quarter and a half of the maximum possible value, then the scaling factor is 2'. In this instance the first bit of each sample in the block is discarded and the next ten bits are transmitted, i.e. bits 2 to 11. For progressively smaller values of the maximum sample the bits transmitted may be bits 3 to 12, bits 4 to 13, or finally bits 5 to 14.

The bits discarded from bit positions 1, 2, 3 and 4 will, of course, all be zeros.

For transmission therefore each input block of 32 14-bit samples becomes a block of 32 10bit samlpes plus a scale factor which can take one of five values, i.e. requiring three bits. In this way an input block requiring 448 bits is transmitted by 323 bits, a reductin of about 28%. Each block has 5 possible scaling factors, so if three blocks are taken together there are a total of 53 or 125 possible states. This can be encoded by just seven bits. However, an error in the range code or scaling factor code word could cause a very noticeable impairment, and hence it may be desirable to allocate an additional four bits per coding word to enable single errors in the sevenbit range code word to be corrected. The scale factors for the three blocks would then be transmitted with eleven bits.

We have appreciated that in either event there are spare scaling factor codes, and with use of these, it is possible to make yet more effective use of the transmission channel through which the data is being transmitted.

The invention is defined in the appended claims to which reference may now be made.

In a preferred embodiment of the invention, an auxiliary signal can be transmitted in an audio signal using the time occupied by selected blocks.

These blocks are indicated by allocating to them a dummy scaling factor. The format of the information transmitted during these blocks is unimportant, i.e. it is not necessary to maintain the structure jm-bit words during these blocks.

The blocks selected for this treatment are preferably blocks which correspond to quiet passages to the extent that the whole block can be replaced in the audio output of the decoder by a repeated mid-range sample or a low-level synthetic noise signal without affecting the subjective sound quality. It may be desirable to provide a code signifying which of a plurality of low-level signals is the most appropriate to use as substitute, and this can be done using a very small portion of the data block released.

The system may be used to send auxiliary data as and when the opportunity arises during quiet passages of a conventional sound signal. The rate of auxiliary data transmission will then be unpredictable (except on a long-term statistical basis). Alternatively, the digitised sound signal may be controlled in order to exploit the technique as a way of giving sound immediate priority in a dedicated channel but only when needed. It could, for example, be used to send a spoken commentary or narrative to a series of graphics sent in the auxiliary data channel. In this instance the system might spend most of its time in the auxiliary mode, with a dummy scale factor being transmitted, but as soon as sound was received the system would revert to 'normal' operation to transmit the sound as blocks of m-bit words.

An example of the invention will now be described with reference to the drawings, in which: Figure 1 is a block circuit diagram indicating the additions required to a near-instantaneous companding system in accordance with this invention, and Figure 2 shows at (a) the format of a block containing audio data and at (b) the format of a block containing auxiliary data.

For a full description of a near-instantaneous companding system reference should be made to the paper mentioned above. The following description is, therefore, confined to those parts of the circuit which need modification in accordance with this invention.

At the encoder 10, an audio input signal is received at an input 12. The signal at input 12 is encoded into a data stream comprising blocks of 32 words each having 10 bits, together with a scaling factor, which can take one of five values, for each block. Each 10 bit word is derived from a 14 bit word in the NICAM encoder by applying the scaling factor which is dependent upon the value of the maximum sample in the block as described above. There are five possible scaling factors for each block, or 53 possible states for 3 blocks A, B and C. These states are encoded into an eleven bit word having seven data bits and four error-protection bits.

The signal thus may be as shown in Figure 2 at (a). A scaling factor code word precedes each set of 3 blocks and gives the scaling factor for each block A, B and C. The blocks follow sequentially and each comprise 32 1 O-bit words.

Reverting to Figure 1, a sample group examiner 14 determines which is the maximum sample in a block and the amount of scaling which is required for that block. The scale factor is applied to a scale factor coder 1 6. The examiner 14 also determines whether a condition exists in which all the blocks in the audio signal are below a predetermined reference value, and if so it applies a dummy code to the scale factor coder. Upon receipt of this dummy code, the coder 16 operates a switch 18 to allow data from an auxiliary data input 20 into the data channel, instead of the encoded audio signal. In these circumstances the signal may be as shown at (b) in Figure 2, in which it is assumed that block A relates to a quiet period and can be omitted and substituted by the auxiliary data. Auxiliary data need not be in the form of 1 O-bit words.However, a small additional code may be transmitted, e.g.

at the beginning of each omitted block, to signify which of a plurality of substitute signals might preferentially be used to replace the omitted block at the decoder.

The whole signal is transmitted through a data channel 25 the precise form of which is immaterial. In Figure 1 the scaling factor code is shown as being transmitted separately from the main data but this is purely arbitrary. Any suitable multiplexing system may be used for example to combine the outputs of several encoded audio signals for transmission together.

The encoder 30 receives the output of the channel 25 and detects the scaling factor code words. The scaling factors detected by a scaling factor decoder 32 for the various blocks of audio samples are applied to those blocks to provide a decoded output, with the aid of a sample bit redistribution circuit 34 and a diagrammaticallyshown switch 35.

When a dummy scale factor is located, the output of the data transmission channel 25 is applied for the associated block to an auxilairy data output 36. During this block it may be desirable to apply the output of a low-level sound generator 38 to the audio output 40. The low level sound generator 38 may generate various possible outputs, in which case the substitute signal code transmitted at the beginning of the omitted block indicates which of the possible substitute signals is to be used.

In this way the audio signal applied to input 12 of the coder reappears at the output 40 of the decoder, subject only to a small amount of degradation introduced by the reduction of the number of bits for transmission, and with quiet passages in the signal omitted or replaced by the output of the low-level sound generator 38. The auxiliary data received at input 20 of the encoder appears at the output 36 of the decoder. It is, therefore, possible to transmit auxiliary data in addition to the audio signal without significantly degrading the audio signal and without an increase in the bit rate.

As described above only a single dummy scale factor has been referred to. It is, however, possible to have more than one dummy scale factor available. The different dummy scale factors can be used to identify respective different ones of several auxiliary data sources and outputs, or to indicate the preferred form of masking noise in which case it would not be necessary to transmit this information separately as part of the data.

Thus it is seen that in the examples described at least one additional 'scale factor' is defined and used to indicat that the 1 ms sampled sound passage is sufficiently quiet that it can be replaced by a repeated mid-range sample or a low-level synthetic noise signal without affecting the subjective sound quality. If it is necessary to signal which of a variety of low-level signals is most appropriate, this can be done either by using an appropriate number of different dummy codes, or by using a very small portion of the auxiliary data made available, this information does not require the same degree of protection as the scale factor itself.

Claims (17)

Claims

1. A coding method for coding digital signals, comprising receiving an input signal comprising a stream of multi-bit words having n bits per word, examining a block of a plurality of words in the input data stream to determine the maximum word value in the block and outputting a coded signal for each block comprising j m-bit words (where m is less than n) together with a scaling factor which can take (n-m+l) values, the output m-bit words for each block being formed from the input words subject to a scaling factor dependent upon the value of the maximum sample in the block, characterised by receiving an auxiliary digital input signal, selectively outputting the coded signal or the auxiliary input signal for each block and transmitting a dummy scaling factor for the block when the auxiliary input signal is outputted in substitution for the coded signal.

2. A coding method according to claim 1, including detecting when all the words in a block of the main input signal are below a predetermined reference value, and in response thereto outputting the auxiluary input signal.

3. A method according to claim 1 or 2, in which when the auxiliary input signal is outputted, during a part of the duration of the omitted block of the coded signal a code is transmitted signifying which of a plurality of substitute signals might preferentially be used to replace the omitted block.

4. A method according to claim 1, 2 or 3, in which the scaling factors for a plurality of blocks are encoded as a composite range code word.

5. Coding apparatus for coding digital signals, comprising an input for receiving an input signal comprising a stream of multi-bit words having n bits per word, range coding means for examining a block of a plurality ofj words in the input data stream to determine the maximum word value in the block for outputting a coded signal for each block comprisingj m-bit words (where m is less than n) together with a scaling factor which can take (n-m+ 1) values, the output m-bit words for each block being formed from the input words subject to a scaling factor dependent upon the value of the maximum sample in the block and an output, characterised by an auxiliary input for receiving an auxiliary digital input signal, controllable selection means for selectively applying the coded signal or the auxiliary input signal to the output for each block and means for transmitting a dummy scaling factor for the block when the auxiliary input signal is applied to the output in substitution for the coded signal.

6. Coding apparatus according to claim 5, including means for'detecting when all the words in a block of the main input signal are of a value below a predetermined reference value, and in response thereto operating the selection means to apply the auxiliary input signal to the output.

7. Coding apparatus according to claim 5 or 6, including means operative when the auxiliary input signal is applied to the output for transmitting during a part of the duration of the omitted block of the coded signal a code signifying which of a plurality of substitute signals might preferentialy be used to replace the omitted block.

8. Coding apparatus according to claim 5, 6 or 7, in which the scaling factors for a plurality of blocks are encoded as a composite range code word.

9. A decoding method for decoding digital signals encoded by the method of claim 1, comprising receiving a coded signal comprisingj m-bit words for each of a plurality of blocks together with a scaling factor appropriate for each block, applying the scaling factor to each word in the block, and applying the resultant decoded signal to an output, characterised by detecting the presence of a dummy scaling factor, and in response thereto applying data received during the associated block to an auxiliary output.

10. A decoding method according to claim 9, in which when the received data is applied to the auxiliary output, a substitute signal is applied to the main output.

1 A decoding method according to claim 10, in which the received data contains a code signifying which of a plurality of predetermined substitute signals is to be applied to the main output.

12. Decoding apparatus for decoding digital signals encoded by the method of claim 1, comprising an input for receiving a coded signal comprisingj m-bit words for each of a plurality of blocks together with a scaling factor appropriate for each block, means for applying the scaling factor to each word in the block, and an output connected to receive the resultant decoded signal, characterised by means for detecting the presence of a dummy scaling factor and in response thereto applying data received during the associated block to an auxiliary output.

13. Decoding apparatus according to claim 12, including means for applying a substitute signal to the main output when the received data is applied to the auxiliary output.

14. Decoding apparatus according to claim 13, including means responsive to a code in the received data signifying which of a plurality of predetermined substitute signals is to be applied to the main output

1 5. A method of coding digital signals substantially as herein described with reference to the drawings.

16. Apparatus for coding digital signals substantially as herein described with reference to the drawings.

17. A method of decoding digital signals substantially as herein described with reference to the drawings.

1 8. Apparatus for decoding digital signals substantially as herein described with reference to the drawings.