WO1998047293A1

WO1998047293A1 - Transmission of mpeg-2 encoded video in atm networks

Info

Publication number: WO1998047293A1
Application number: PCT/SE1998/000648
Authority: WO
Inventors: Per Tholin; Gunnar Bahlenberg; Erik Dillner; Magnus Johansson; Lis-Marie Ljunggren; Harry Tonvall; Per-Ola Wester
Original assignee: Telia Ab (Publ)
Priority date: 1997-04-16
Filing date: 1998-04-08
Publication date: 1998-10-22
Also published as: EE9900491A; SE519917C2; NO994939L; EP0986915A1; NO994939D0; SE9701415D0; SE9701415L

Abstract

A telecommunications transmission system adapted for the transmission of MPEG-2 data streams with a variable bit rate includes an encoder and decoder. The encoder includes a system layer multiplexer adapted to empty a video encoder buffer by reading data blocks therefrom, at a bit rate which is constant within a block. The encoder provides a way of realising VBR in an MPEG-2 system layer multiplexer which has been adapted for ATM, and represents a particularly simple implementation of a VBR MPEG-2 encoder having good jitter control. The encoder eliminates the need for a special 'cell shaping' function in order to fulfil a traffic contract. To recover bit timing in the MPEG-2 decoder, data is buffered in a reception buffer. By continuously reading the deviation of the buffer level, i.e. the level to which the buffer is filled with data, from its nominal value, in this buffer, the deviation of the emptying rate of the buffer, from the nominal bit rate, can be controlled with a phase lock loop (PLL), or similar arrangement.

Description

TRANSMISSION OF MPEG-2 ENCODED VIDEO IN ATM NETWORKS

The present invention relates to variable bit rate MPEG-2 encoders and decoders, particulariy encoders and decoders adapted for use with ATM, together with methods of encoding and decoding MPEG-2 data having a variable bit rate and systems for transmitting MPEG-2 data with a variable bit rate

Systems for the transmission of MPEG-2 encoded video data over ATM networks are known, see for example, "The AMPEC System A Broadband

Multimedia Terminal Supporting High Definition Audiovisual Communications" -Bo Burman, Session 1 4 of the 7th World Telecommunications Forum, Geneva 3 - 11

October 1995

MPEG-2, or ISO/IEC IS 13818, which is the official name for the MPEG-2 standard, is a key standard for high quality audiovisual communication and distribution A number of other standards/recommendations employ MPEG-2 for coding and multiplexing, for example, the new ITU-T Rec H 310 which sets out a terminal specification for broadband audiovisual communication over ATM

At the present time, standards such as H 310 and ATM Forum's Video on Demand specification, etc , only support MPEG-2 streams which have a constant bit rate, CBR However, both MPEG-2 and ATM have the potential to operate with variable bit rates, VBR VBRs can be used, with advantage, for encoding video signals It is worthy of note that the new DVD discs, which will in due course replace CDs, will have the ability to store MPEG-2 coded streams with VBR Future versions of H 310, and other standard specifications, are expected to include VBR-support

VBR, which must comply with a traffic contract, is not easily implemented in a MPEG-2 encoder There are three major problems which must be overcome if a VBR MPEG-2 encoder is to be realised, namely reading the video buffer;

guaranteeing that the traffic characteristic is not changed by subsequent network adaptations, e.g. the ATM Adaption Layer; and

avoiding the introduction of self induced jitter.

The present invention provides, inter alia, a way of realising VBR in a

MPEG-2 system layer multiplexer which has been adapted for ATM. This can be used to provide a particularly simple implementation of a VBR MPEG-2 encoder having good jitter control and, at the same time, eliminates the need for a special "cell shaping" function in order to fulfil a traffic contract.

If a MPEG-2 stream has been transferred by an ATM-network which introduces significant delay variation, an MPEG2 decoder, used for decoding the bit stream, must include a unit which recovers the transmitted bit rate. The demands made by variable bit rate (VBR) on timing recovery differ from those made by constant bit rate (CBR). In particular, it should be noted that a VBR traffic class is expected to introduce higher delay variation, or cell jitter, than is typically introduced by a CBR traffic class, because statistical multiplexing requires comparatively large buffers in the ATM-switches.

To recover bit timing in an MPEG2 decoder, data is buffered in a reception buffer. By continuously reading the deviation of the buffer level, i.e. the level to which the buffer is filled with data, from its nominal value, in this buffer, the deviation of the emptying rate of the buffer, from the nominal bit rate, can be controlled with a phase lock loop (PLL), or similar arrangement. This is the principle employed in the "Adaptive clock" technique, which is specified in ITU-T Rec 1.363 (AAL1). The "Adaptive Clock" principle is, however, only able to handle CBR. In a system with variable bit rate, timing recovery is harder to achieve because:

the nominal buffer level must be allowed to vary in time as the bit rate varies in order to keep the delay constant in the receiver; and the nominal bit rate is not known.

The present invention can be used in encoders, or terminals, for audiovisual communication, or distribution, based on MPEG-2 over ATM. Although this technique is primarily intended for use with MPEG-2 decoders which manage variable bit rate, for example, ATM-adapted MPEG-2 decoders operating with

VBR. The technique is not, however, limited to ATM transmission technology, but can be used for any transmission technology which supports VBR.

Proposals have been made in ETSI NA5 for an AAL2 which manages the dejittering of a VBR data stream. These proposals are, however, significantly more complicated than the technique herein described for dejittering a VBR data stream, for example, they require the use of special time stamps which substantially increase the complexity of any practical implementation.

According to a first aspect of the present invention, there is provided a telecommunications transmission system adapted for the transmission of MPEG-2 data streams, over a transmission network, with a variable bit rate including an encoder and decoder, characterised in that said encoder has data reading means, for reading data blocks from a video buffer, a bit rate within a block being constant and time slots between blocks having a variable duration.

Said decoder may have a reception buffer for receiving incoming MPEG-2 data, timing recovery means arranged to extract a current bit rate from a MPEG-2 protocol system layer, calculation means for calculating a nominal buffer level and recovery means for adaptively recovering deviations from said nominal bit rate.

According to a second aspect of the present invention, there is provided a telecommunications transmission system adapted for the transmission of MPEG- 2 data streams, over a transmission network, with a variable bit rate including an encoder and decoder, characterised in that said decoder has a reception buffer for receiving incoming MPEG-2 data, timing recovery means arranged to extract a current bit rate from a MPEG-2 protocol system layer, calculation means for calculating a nominal buffer level for said reception buffer and recovery means for adaptively recovering deviations from said nominal bit rate

Said encoder may have data reading means, for reading data blocks from a video buffer, a bit rate within a block being constant and time slots between blocks having a variable duration

Said transmission network may be an ATM transmission network and said encoder and decoder may be adapted to operate with ATM

According to a third aspect of the present invention, there is provided an encoder adapted to encode MPEG-2 data for transmission over an ATM network with a vanable bit rate, characterised in that data reading means are provided for reading data blocks from a video buffer, a bit rate within a block being constant and time slots between blocks having a variable duration

According to a fourth aspect of the present invention, there is provided an encoder adapted to encode MPEG-2 data for transmission over an ATM network with a variable bit rate, characterised in that said encoder includes a system layer multiplexer adapted to empty a video encoder buffer by reading data blocks therefrom, at a bit rate which is constant within a block

Said blocks may have a size B, where B = B₀ - H, where B₀ is equal to a SAR-PDU payload and H is the number of header bytes contained in a block, and in that H = 0 for most, but not all, of said blocks

Said system layer multiplexer may be adapted to control time intervals between said blocks in response to signals transmitted thereto by a regulator means

Said system layer multiplexer may generate an AAL payload flow which meets criteria specified by a traffic contract

Said regulator means may select time intervals between said blocks in dependence on an input signal indicative of said video data buffer's contents

Said system layer multiplexer may be adapted to locate PCR fields early in said blocks and thereby maintain alignment between transport packets and data blocks

Said bit rate may be substantially equal to one eighth of a network bit clock over which said MPEG-2 data is to be transmitted

Said encoder may include a programmable counter adapted to count down from a value, corresponding to the separation of two consecutive cells, to zero, and, in a latter part of a counting cycle, a cell may be transmitted

Said regulator means may be adapted to measure network load and bandwidth requirements for a given application and to determine the separation of two consecutive cells in dependence thereon

Said regulator means may be adapted to negotiate available bandwidth with a network over which said MPEG-2 data is to be transmitted

ATM cells may be transmitted in accordance with AAL1 , said programmable counter may be adapted to count from 52 to zero, a header may be transmitted when said counter has values between 52 and 47 and MPEG-2 data may be transmitted when said counter has values between 46 and zero

Flag means may be provided and arranged to set a flag, when said counter is zero, and reset said flag prior to reading out cells

According to a fifth aspect of the present invention, there is provided a method of encoding MPEG-2 data for transmission over an ATM network with a variable bit rate, characterised by reading data blocks from a video buffer, a bit rate within a block being constant and time slots between blocks having a variable duration A video encoder buffer may be emptied by reading data blocks therefrom at a bit rate which is constant within a block.

Said blocks may have a size B, where B = B₀ - H, where B₀ is equal to a SAR-PDU payload and H is the number of header bytes contained in a block, and H may be zero for most, but not all, of said blocks.

Time intervals between said blocks may be controlled in response to signals transmitted by a regulator means.

An AAL payload flow may be generated which meets criteria specified by a traffic contract.

Time intervals between said blocks may be selected in dependence on said video data buffer's contents.

PCR fields may be located early in said blocks thereby maintaining alignment between transport packets and data blocks.

Said bit rate may be set substantially equal to one eighth of a network bit clock for a network over which said MPEG-2 data is to be transmitted.

Network load and bandwidth requirements may be measured for a given application and the separation of two consecutive cells may be determined in dependence thereon.

Available bandwidth may be negotiated with a network over which said MPEG-2 data is to be transmitted.

ATM cells may be transmitted in accordance with AAL1.

According to a sixth aspect of the present invention, there is provided a decoder, adapted to decode MPEG-2 data transmitted with a variable bit rate, characterised in that said decoder has a reception buffer for receiving incoming MPEG-2 data, timing recovery means arranged to extract a current bit rate from a MPEG-2 protocol system layer, calculation means for calculating a nominal buffer level and recovery means for adaptively recovering deviations from said nominal bit rate.

Said MPEG-2 data may have been transmitted over an ATM network and said decoder may be adapted to operate with ATM.

Said deviations, from a nominal bit rate, of the rate of emptying said buffer may be controlled by a phase locked loop.

Said nominal buffer level may be permitted to vary in time as said variable bit rate varies.

Said calculation means may operate on a bit rate immediately preceding said current bit rate, and said current bit rate, to calculate a nominal buffer level.

Said nominal buffer level may be used as a reference for adaptively recovering deviations from a current nominal bit rate by means of said phase locked loop.

Said nominal buffer level may vary, in a period where said nominal buffer level changes from an old to a new value, in the form of a ramp, either up, or down, between periods where said nominal buffer level is substantially constant.

Bit rate information may be extracted from outgoing data from said buffer and said bit rate information may be used to determine when said nominal bit rate is changed

According to a seventh aspect of the present invention, there is provided a method of decoding MPEG-2 data for transmission with a variable bit rate, characterised by:

- extracting a current bit rate from a MPEG-2 protocol system layer, calculating a nominal buffer level for a reception buffer adapted to receive incoming MPEG-2 data; and

adaptively recovering deviations from said nominal bit rate.

A phase locked loop may be used to control said deviations, from a nominal value, of the buffer level of said reception buffer.

A nominal buffer level may be calculated from an immediately preceding bit rate and said current bit rate.

Bit rate information may be extracted from outgoing data from said buffer and said bit rate information may be used to determine when said nominal bit rate is changed.

According to a eighth aspect of the present invention, there is provided a system for transmitting MPEG-2 data with a variable bit rate, including an encoder, a decoder and an ATM network characterised in that:

said encoder is an encoder as described in any preceding paragraph; and/or

said encoder operates according to a method as described in any preceding paragraph; and/or

said decoder is an decoder as described in any preceding paragraph; and/or

- said decoder operates according to the method as described in any preceding paragraph.

According to a ninth aspect of the present invention, there is provided a MPEG-2 terminal for transmitting/receiving MPEG-2 data with a variable bit rate, including an encoder, a decoder and an ATM network characterised in that

said encoder is an encoder as described in any preceding paragraph, and/or

said encoder operates according to a method as described in any preceding paragraph, and/or

said decoder is an decoder as described in any preceding paragraph, and/or

said decoder operates according to the method as described in any preceding paragraph

Embodiments of the invention will now be described, by way of example, with reference to the accompanying drawings, in which

Figure 1 illustrates a process of reading data, block-by-block, from a video buffer with a variable bit rate

Figure 2 illustrates the variation with time of a programmable counter's contents with constant block and cell distance

Figure 3 illustrates the variation with time of a programmable counter's contents with variable block and cell distance.

ATM for broadband transmission has the ability to support traffic sources with variable data rates, VBR Special traffic classes are defined for VBR ATM transmission, namely

real time traffic, and

other data traffic without special real time demands The MPEG-2 video encoding technique supports VBR. Thus, it is possible to construct a MPEG-2 compliant encoder which has the advantages conferred by

VBR, such as, a more uniform picture quality and shorter delay. It should also be noted that MPEG-2 specifications which define packeting, multiplexing and synchronisation, include support for VBR.

An encoder operating with VBR requires a control mechanism to continuously calculate a current data rate on the basis of a number of different input signals. Ideally, the data rate is a continuously variable function of time and can be different for each bit produced by an encoder. This ideal solution would be difficult to implement in a real encoder since the bit clock, which empties the video buffer, would, in principle, be required to vary continuously.

However, a VBR implementation of a MPEG-2 encoder can be substantially simplified by using a fixed clock frequency, selected in dependence on the highest data rate to be handled by the encoder, instead of a continuously varying transmission clock Data is then read out of the video buffer, in blocks, at the fixed clock frequency The variation in bit rate is achieved by varying the interval between data blocks In other words, the instantaneous rate will be burst-like with varying lengths of time slot between bursts, in which the bit rate is zero This process is illustrated in Figure 1 It can be seen that the bit rate can be varied from a maximum value, on the right of Figure 1 , where the bursts are continuous, i e no intervening time slots between data blocks, to a much lower bit rate on the left of Figure 1 , in which there are substantial time slots between data blocks By averaging over time from the start of one data block to the start of the next block, it can be seen that a constant data rate is achieved within a data block, which can be varied from data block to data block. Provided the size of data blocks is kept comparatively small, e g equivalent to the size of an ATM cell, there are no restrictions imposed by the use of a step-wise block-by-block variation in bit rate compared with a continuous vaπation in bit rate. In practice, the use of a step-wise vaπation in bit rate, based on data blocks corresponding to the size of ATM cells, fits very well with the implementation of VBR in a MPEG-2 encoder adapted for use with ATM MPEG-2 supports VBR in the form of a "piecewise constant rate". This means that the byte stream has a constant byte rate between either:

two consecutive PCR fields (the transport stream case); or

two consecutive program-mux-rates (the program stream case),

see IS 13818-1 for more information.

When an ATM network carries a cell stream with VBR, the cell stream transfer is such that the cell interval is permitted to vary. A traffic contract is established when a VC (Virtual Channel) is established, such a traffic contract can be modelled by a "leaky bucket", see for example the ATM Forum's description of VBR. To comply with the traffic contract, the encoder must supervise the following three parameters:

peak cell rate;

sustained cell rate; and

burst size.

These three parameters must not be allowed to exceed the maximum values specified in the traffic contract if the network is to guarantee transmission of the cell stream.

The present invention provides VBR in a MPEG-2 encoder through the action of the system layer multiplexer which empties the video encoder buffer block-by-block at a bit rate R. The block size B is selected according to the size of the AAL payload that is currently applicable. For most blocks B is a constant B₀, but for certain blocks the value of B must be reduced, i.e. B < B₀. It is at these points that other information is multiplexed into the bit stream, for example, PES- headers etc.. In other words: B = B₀ - H where H indicates the number of header bytes that are multiplexed into the data stream

If AAL1 is used for VBR, then B₀ is selected to be equal to 47 bytes, since this is the SAR-PDU (Segmentation and Reassembly Protocol Data Unit) payload size in the simplest case

The interval between two consecutive blocks is controlled by a multiplexer which, in turn, gets information from a regulator The multiplexer therefore becomes the "rate master" in the system and other elements of the system adapt their processing rate to that set by the multiplexer Thus, the multiplexer can, while the video buffer is being emptied, create a cell flow, strictly speaking an AAL payload flow, which meets the parameter values specified in the traffic contract No following "cell shaping" is needed provided that sequential steps in the AAL are implemented as pipe-lines with fixed delays

R is preferably selected to have a comparatively high value because this gives a low self induced jitter in the case where a plurality of different channels are multiplexed together, each with its own VCI

If the control mechanism, used to select the data rate, uses the degree of fill of the video buffer as an input signal, allowance must be made for rapid changes of the signal duπng the short period when a data block is read out of the video buffer The gradient in buffer filling, as a function of time increases as the selected value of R increases This does not cause any practical difficulties, but it does have implications for the design of the regulator

Internally, in the encoder, the multiplexer and the AAL, etc , data is transmitted in the form of blocks which always have an internal bit rate of R This need not cause jitter problems Provided alignment between transport packets and data blocks is maintained, i e so that PCR fields always occur early in the blocks the self induced jitter amplitude can be kept to an acceptably low level By using the method outlined above, the multiplexer generates a "piece- wise constant cell rate", which is referred to as PCCR in this patent specification Implementation of the PCCR control mechanism in the multiplexer will now be considered

The control mechanism for PCCR operates with a clock frequency which is preferably, but not necessarily, selected as the bit clock of the network divided by eight, i.e. the clock rate corresponds to the network "byte clock" This rate is used to select the number of network clock cycles that will separate each cell This mechanism includes a programmable counter which counts from the value "cell distance", see Figure 2, to zero The "cell distance" is the distance between two consecutive cell starts. During the latter part of the counting cycle the cell is placed on the channel for transmission

By using a programmable counter, the cell distance value can be changed and this in turn enables the distance between cells to be varied The value of the cell distance is decided in a separate regulator which can measure the load on the network and bandwidth requirements of the application to be used On the basis of this data, the regulator negotiates, with the network, the bandwidth that will be made available - traffic contract The case of variable block and cell distance is illustrated in Figure 3

The number of bits the counter contains depends on how low a bit rate, i.e the size of the maximal distance between cells, can be achieved For instance, if the maximal channel rate is 155Mbιt/s and the encoder needs to utilize 1 5 Mbit/s, then the distance will be

cell distance = (cell size x 155)/1 5 expressed as a number of network clock cycles (byte oriented transmission)

When this method is used, it is advantageous to adapt the reading out of a cell to the counter This implies that if ATM cells are transmitted in accordance with AAL1, then the header will be transmitted when the counter has values between 52 and 47, inclusive, so that data can be transmitted between counter values of 46 to 0, inclusive, (byte oriented transmission) It should be noted that it is not permitted to load the counter with values smaller than the cell size, because this will inevitably result in faulty reading out By setting a flag when the counter is zero, which is reset prior to commencement of reading out cells, it is possible to provide protection against faulty operation If the counter value, when the flag is reset, is not passed, the cell reading out process is aborted

To recover bit timing in an MPEG2 decoder, data is buffered in a reception buffer By continuously reading the deviation of the level, from its nominal value, in this buffer, the deviation of the emptying rate of the buffer, from the nominal bit rate, can be controlled with a phase lock loop (PLL), or similar arrangement This is the principle employed in the "Adaptive clock" technique, which in specified in ITU-T Rec I 363 (AAL1) The "Adaptive Clock" principle is, however, only able to handle CBR In a system with variable bit rate, timing recovery is harder to achieve because

- the nominal buffer level must be allowed to vary in time as the bit rate vanes in order to keep the delay constant in the receiver, and

the nominal bit rate is not known

A buffer in the receiver is filled with incoming MPEG-2 data Before being entered into the buffer, the piece-by-piece constant bit rate Is extracted from incoming data by the system layer of the MPEG2 protocol This may be executed, for example, by means of the parameter programjnux_rate, see ISO/IEC IS13818-1 The nominal buffer level is calculated on the basis of this information and the preceding bit rate The nominal buffer level will typically adopt the form of a ramp up, or down which changes to a constant level when all data which was transmitted with the old rate has left the buffer

The calculated nominal level can then be used as reference for the adaptive recovery of the deviation from the nominal bit rate by means of a PLL, or similar arrangement To determine when the nominal bit rate should be changed, the bit rate information from outgoing data is also extracted This technique can be regarded as an extension of 'Adaptive Clock", which enables it to function with VBR, as well as CBR. An implementation of this technique in an ATM-adapted MPEG-2 decoder can, therefore, utilize the same hardware as that used for an AAL1 Adaptive Clock. This means that separate AAL1 implementations for VBR and CBR are not needed.

Claims

1 A telecommunications transmission system adapted for the transmission of MPEG-2 data streams, over a transmission network, with a variable bit rate including an encoder and decoder, characterised in that said encoder has data reading means, for reading data blocks from a video buffer, a bit rate within a block being constant and time slots between blocks having a variable duration

2 A telecommunications transmission system adapted for the transmission of MPEG-2 data streams, over a transmission network, with a variable bit rate including an encoder and decoder, characterised in that said decoder has a reception buffer for receiving incoming MPEG-2 data, timing recovery means arranged to extract a current bit rate from a MPEG-2 protocol system layer, calculation means for calculating a nominal buffer level for said reception buffer and recovery means for adaptively recovering deviations from said nominal bit rate

3 A telecommunications transmission system, as claimed in claim 1 , characterised in that said decoder has a reception buffer for receiving incoming MPEG-2 data, timing recovery means arranged to extract a current bit rate from a MPEG-2 protocol system layer, calculation means for calculating a nominal buffer level for said reception buffer and recovery means for adaptively recovering deviations from said nominal bit rate

4 A telecommunications transmission system, as claimed in claim 2, characterised in that said encoder has data reading means, for reading data blocks from a video buffer, a bit rate within a block being constant and time slots between blocks having a variable duration

5 A telecommunications transmission system, as claimed in any previous claim, characterised in that said transmission network is an ATM transmission network and in that said encoder and decoder are adapted to operate with ATM

6 An encoder adapted to encode MPEG-2 data for transmission over an ATM network with a variable bit rate, characterised in that data reading means are provided for reading data blocks from a video buffer, a bit rate within a block being constant and time slots between blocks having a variable duration

7 An encoder adapted to encode MPEG-2 data for transmission over an ATM network with a variable bit rate, characterised in that said encoder includes a system layer multiplexer adapted to empty a video encoder buffer by reading data blocks therefrom, at a bit rate which is constant within a block

8 An encoder, as claimed in claim 7, characterised in that said blocks have a size B, where B = B₀ - H, where B₀ is equal to a SAR-PDU payload and H is the number of header bytes contained in a block, and in that H = 0 for most, but not all, of said blocks

9 An encoder as claimed in either claim 7, or 8, characterised in that said system layer multiplexer is adapted to control time intervals between said blocks in response to signals transmitted thereto by a regulator means

10 An encoder, as claimed in claim 9, characterised in that said system layer multiplexer generates an AAL payload flow which meets criteria specified by a traffic contract

11 An encoder, as claimed in either claim 9, or 10, characterised in that said regulator means selects time intervals between said blocks in dependence on an input signal indicative of said video data buffer's contents

12 An encoder, as claimed in any of claims 7 to 11 , characterised in that said system layer multiplexer is adapted to locate PCR fields early in said blocks and thereby maintain alignment between transport packets and data blocks

13 An encoder as, claimed in any of claims 7 to 12, characterised in that said bit rate is substantially equal to one eighth of a network bit clock over which said MPEG-2 data is to be transmitted

14 An encoder as, claimed in any of claims 7 to 13, characterised in that said encoder includes a programmable counter adapted to count down from a value, corresponding to the separation of two consecutive cells, to zero, and in that, in a latter part of a counting cycle, a cell is transmitted.

15 An encoder, as claimed in any of claims 9 to 14, characterised in that said regulator means is adapted to measure network load and bandwidth requirements for a given application and to determine the separation of two consecutive cells in dependence thereon

16 An encoder, as claimed in claim 15, characterised in that said regulator means is adapted to negotiate available bandwidth with a network over which said MPEG-2 data is to be transmitted

17 An encoder, as claimed in any of claims 9 to 16, characterised in that ATM cells are transmitted in accordance with AAL1 , in that said programmable counter is adapted to count from 52 to zero, in that a header is transmitted when said counter has values between 52 and 47, and in that MPEG-2 data is transmitted when said counter has values between 46 and zero.

18 An encoder, as claimed in claim 17, characterised in that flag means are provided and arranged to set a flag, when said counter is zero, and reset said flag prior to reading out cells

19 A method of encoding MPEG-2 data for transmission over an ATM network with a vanable bit rate, characterised by reading data blocks from a video buffer, a bit rate within a block being constant and time slots between blocks having a variable duration

20 A method of encoding MPEG-2 data for transmission over an ATM network with a variable bit rate, characterised by emptying a video encoder buffer by reading data blocks therefrom at a bit rate which is constant within a block

21 A method, as claimed in claim 20, characterised in that said blocks have a size B, where B = B₀ - H, where B₀ is equal to a SAR-PDU payload and H is the number of header bytes contained in a block, and in that H = 0 for most, but not all, of said blocks

22 A method, as claimed in either claim 20, or 21 , characterised by controlling time intervals between said blocks in response to signals transmitted by a regulator means

23 A method, as claimed in claim 22, characterised by generating an AAL payload flow which meets criteria specified by a traffic contract

24 A method, as claimed in either claim 22, or 23, characterised by selecting time intervals between said blocks in dependence on said video data buffer's contents

25 A method, as claimed in any of claims 20 to 24, characterised by locating PCR fields early in said blocks and thereby maintaining alignment between transport packets and data blocks

26 A method, as claimed in any of claims 20 to 25, characterised by setting said bit rate substantially equal to one eighth of a network bit clock for a network over which said MPEG-2 data is to be transmitted

27 A method, as claimed in any of claims 22 to 26, characterised by measu╧Çng network load and bandwidth requirements for a given application and determining the separation of two consecutive cells in dependence thereon

28 A method, as claimed in claim 27, characterised by negotiating available bandwidth with a network over which said MPEG-2 data is to be transmitted

29 A method, as claimed in any of claims 22 to 27, characterised by transmitting ATM cells in accordance with AAL1

30 A decoder, adapted to decode MPEG-2 data transmitted with a variable bit rate, characterised in that said decoder has a reception buffer for receiving incoming MPEG-2 data, timing recovery means arranged to extract a current bit rate from a MPEG-2 protocol system layer, calculation means for calculating a nominal buffer level for said reception buffer and recovery means for adaptively recovering deviations from said nominal bit rate

31 A decoder, as claimed in claim 30, characterised in that said MPEG-2 data has been transmitted over an ATM network and said decoder is adapted to operate with ATM

32 A decoder, as claimed in either claim 30, or 31 , characterised in that said ; deviations, from a nominal bit rate, are controlled by a phase locked loop

33 A decoder, as claimed in any of claims 30 to 32, characterised in that said nominal buffer level is permitted to vary in time as said variable bit rate varies

34 A decoder, as claimed in any of claims 30 to 33, characterised in that said calculation means operates on a bit rate immediately preceding said current bit rate, and said current bit rate, to calculate a nominal buffer level

35 A decoder, as claimed in any of claims 32 to 34, characterised in that said nominal buffer level is used as a reference for adaptively recovering deviations from a current nominal bit rate by means of said phase locked loop

36 A decoder, as claimed in any of claims 30 to 35, characterised in that said nominal buffer level varies, in a period where said nominal buffer level changes from an old to a new value, in the form of a ramp, either up, or down, between periods where said nominal buffer level is substantially constant

37 A decoder, as claimed in any of claims 30 to 36, characterised in that bit rate information is extracted from outgoing data from said buffer and said bit rate ^ information is used to determine when said nominal bit rate is changed

38 A method of decoding MPEG-2 data for transmission with a variable bit rate, characterised by

extracting a current bit rate from a MPEG-2 protocol system layer,

calculating a nominal buffer level for a reception buffer adapted to receive incoming MPEG-2 data, and

- adaptively recovering deviations from said nominal bit rate

39 A method, as claimed in claim 38, characterised in that said MPEG-2 data has been transmitted over an ATM network

40 A method, as claimed in either claim 38, or 39, characterised by using a phase locked loop to control said deviations, from a nominal value, of the buffer level of said reception buffer

41 A method, as claimed in any of claims 38 to 40, characterised by permitting said nominal buffer level to vary in time as said variable bit rate varies

42 A method, as claimed in any of claims 38 to 41 , characterised by calculating a nominal buffer level from an immediately preceding bit rate and said current bit rate

43 A method, as claimed in any of claims 40 to 42, characte╧Çsed by using said nominal buffer level as a reference for adaptively recovering deviations from a current nominal bit rate by means of said phase locked loop

44 A method, as claimed in any of claims 38 to 43, characterised by said nominal buffer level varying, in a period where said nominal buffer level changes from an old to a new value, in the form of a ramp, either up, or down

45 A method, as claimed in any of claims 38 to 44, characterised by extracting bit rate information from outgoing data from said buffer and using said bit rate information to determine when said nominal bit rate is changed

46. A system for transmitting MPEG-2 data with a variable bit rate, including an encoder, a decoder and an ATM network characterised in that:

said encoder is the encoder as claimed in any one of claims 6 to 18; and/or

j - said encoder operates according to the method as claimed in any one of claims 19 to 29; and/or

said decoder is the decoder as claimed in any one of claims 30 to 37; and/or

said decoder operates according to the method as claimed in any 0 one of claims 38 to 45.

47. A MPEG-2 terminal for transmitting/receiving MPEG-2 data with a variable bit rate, including an encoder, a decoder and an ATM network characterised in that:

said encoder is the encoder as claimed in any one of claims 6 to 5 18; and/or

said encoder operates according to the method as claimed in any one of claims 19 to 29; and/or

said decoder is the decoder as claimed in any one of claims 30 to 37; and/or

0 - said decoder operates according to the method as claimed in any one of claims 38 to 45.