WO1999030503A9

WO1999030503A9 - Method and apparatus for encoding and slicing digital video bit streams

Info

Publication number: WO1999030503A9
Application number: PCT/IB1998/001922
Authority: WO
Inventors: Hayder Radha; Mahesh Balakrishnan
Original assignee: Koninkl Philips Electronics Nv; Philips Svenska Ab
Priority date: 1997-12-09
Filing date: 1998-12-03
Publication date: 2005-03-17
Also published as: WO1999030503A3; EP0958703A2; JP4242459B2; DE69836416D1; WO1999030503A2; JP2001511335A; DE69836416T2; EP0958703B1; US6049569A

Abstract

A method of encoding a digital video bit stream which includes a plurality of successively-encoded pictures. The invention discloses the steps of constructing seamless In Points and seamless Out Points while using the normal bitrate, without causing a decoder buffer underflow or overflow in an MPEG-2 compliant receiver. The constraints that the decoder delay ddf must adhere to (ddf is the amount of time by which the first byte of the digital video bit stream after the seamless In or Out Point must stay in a decoder buffer before it is decoded), is expressed by respective equations.

Description

Method and apparatus for encoding and slicing digital video bit streams.

BACKGROUND OF THE INVENTION

The present invention relates generally to the encoding of digital video signals, and morre particularly, to a novel approach to the construction of seamless splice points which minimizes the constraints on the upper and lower bounds of the bit rate which can be used for encoding a digital video bit stream (such as an MPEG-2 compliant video bit stream), to thereby significantly improve the picture quality relative to the presently known technology. Switching of compressed video streams (i.e., "digital video bit streams") is expected to be an important operation for future digital television production and distribution services. However, switching of digital video bit streams presents technical challenges that do not exist in current analog-based TV systems. One way of dealing with these challenges is to decompress the desired streams, perform the switching operation in baseband, and then re-compress the resulting switched signal. This approach, however, could introduce significant degradations, especially for broadcast quality streams. Also, performing switching in digital baseband may require costly solutions, especially when multiple programs have to be switched simultaneously. Another approach to compressed video signal switching is to create seamlessly spliceable streams, which makes it feasible to switch from one compressed video stream to another without the need to go to baseband. The creation of seamlessly spliceable streams requires the specification and standardization of seamless splicing parameters (such as the ones outlined in the MPEG-2 system document). These splicing parameters are used to construct Seamless Splice Points. A Seamless Splice Point (SSP) is a location in a bitstream that provides an opportunity to splice. When switching is desired, a switching device can switch from one SSP of one stream to another SSP of another stream without causing any decoder buffer underflow or overflow in an MPEG-2 compliant receiver. It is important to note that switching from an arbitrary point (i.e. not an SSP) in an MPEG-2 compliant stream to another arbitrary point of an MPEG-2 compliant stream can (and most likely will) cause an underflow or an overflow event at the decoder buffer. Thus, such "arbitrary point switching" is not "seamless". In MPEG-2, the presence of a splice point is indicated (the spice point is "constructed") by a splice flag field and splice_countdown fields in the bitstream. Some properties of the stream at the splicing point are defined by two seamless splicing parameters: splice_decoding_delay (SDD) and maximum_splice_rate (MSR). In accordance with MPEG-2, when constructing an SSP, the following constraints must be adhered to: The amount of time (dd_f ) that the first byte of the MPEG-2 compliant stream after the SSP has to stay in the decoder buffer before it is decoded is defined as follows: dd_f = SDD To prevent a decoder buffer overflow event, it is necessary to adhere to the following constraints when constructing an SSP: The bitrate that must be assumed (r_a ) when creating the pictures prior to an SSP has to be MSR: r_a = MSR The bitrate that must be used (r_u ) after an SSP has to be less than or equal to MSR: r_u < MSR.

The SMPTE committee on MPEG-2 splicing built on the MPEG-2 seamless splicing work and started a new effort to standardize splicing constraints and tools, in general, and specific SDD and MSR values, in particular. The SMPTE committee defines two types of splice points: In Point (IP) and Out Point (OP). According to the SMPTE draft standard, an In Point is "a point in a PID stream where a splicing device may enter". (A PID stream can be a video stream, audio stream, etc.). Similarly, an Out Point is defined as "a point in a PID stream where a splicing device may exit".

Also, according to SMPTE teachings, every seamless video In Point must meet all the constraints of a seamless video Out Point (the terms seamless video In Point and seamless video Out Point will hereinafter be referred to simply as "seamless In Point" and "seamless Out Point"). However, a seamless In Point must have additional constraints. The main (video related) constraint is that the first picture after a seamless In Point must be an Intra-coded (I) picture. Therefore, based on the SMPTE committee teachings, all of the SDD and MSR related constraints (outlined above) must be adhered to when constructing both seamless In Points and seamless Out Points.

As a consequence, the presently known technology imposes the following constraints on the construction of both seamless video In Points and Out Points: dd_f = SDD r_a = MSR r_u <>1SR The constraints to assume a defined bitrate r_a before the splicing point and restrict the bitrate r_u used after the splicing point make it very difficult to maintain good picture quality when constructing seamless splice points. This is especially true when the normal bit rate (r_n) of the stream where seamless splice points must be inserted is very different from MSR.

FIG. 1 graphically depicts these constraints in accordance with the teachings of Applicant's co-pending U.S. Patent Application Serial Number 08/829,124. In the Figure, t_sp-Δ is the time when the encoder bounds start changing, and where Δ is the end-to- end delay (i.e. encoder buffer plus decoder buffer delays), t_e is the time of encoding the "First- picture", t_sp is the time when the first byte of the "First-Picture" enters the decoder buffer, and t_Sp+SDD is the time of decoding the "First-Picture". The Figure illustrates an example of the lower encoder buffer bound 1 and upper encoder buffer bound 2 that must be adhered to when constructing a seamless In Point or seamless Out Point, when the normal bit rate (r_n) of the stream where seamless splice points must be inserted is smaller than MSR (block 10) or greater than MSR (block 11). The Figure also shows the "normal" lower (3) and upper (4) encoder bounds, i.e., the encoder bounds that must normally be adhered to by the encoder when there is no need to construct an SSP, As illustrated in this example, the encoder must adhere to stringent buffer constraints when constructing an SSP. Reference numeral 12 denotes that the constraint dd_f = SDD is the same as in the prior art.

A method for constructing seamless In and Out Points in such a manner as to eliminate some of these stringent constraints is proposed in co-pending U.S. Patent Application titled Video Buffer for Seamless Splicing of MPEG Streams, filed on November 25, 1997, by Hayder Radha, one of the present co-inventors, and assigned to the present assignee, the disclosure of which is also herein incorporated by reference. An example of how this method is used to construct seamless In and Out Points is graphically depicted in FIG. 2, which illustrates the lower (1) and upper (2) bound constraints for Out Points (OP) and In Points (IP). The stringent constraints for Points that are both In Points and Out Points (IOP) are not eliminated by using this approach. This is denoted by block 13 that represents the prior art strategy. As can be seen in FIG. 2, the constraints for Out Points where r_n>MSR (block 14) and In Points where r_n<MSR (block 15) are now less stringent. In some seamless In and Out Point cases indicated by reference numerals 10 (Out Point where r_n<MSR) and 11 (In Point where r_n>MSR), the stringent constraints are not eliminated by using this approach. As in FIG. 1, the constraint dd_f = SDD is the same as in the prior art (cf. blocks 12).

Based on the above and foregoing, it can be appreciated by those skilled in the pertinent art that there presently exists a need in the art for a novel method for constructing seamless Out Points and seamless In Points which enables the encoder to construct seamless Out Points and seamless In Points while using and assuming bit rates equal to the normal bit rates, thereby eliminating major seamless splicing constraints which can not be eliminated using the present technology, and consequently significantly improving the picture quality around the seamless In Points and seamless Out Points relative to the present technology. The present invention fulfills this need in the art.

SUMMARY OF THE INVENTION

Accordingly, one objective of the present invention is to enable encoders to construct seamless In and Out Points while assuming and using their normal bitrates. The notion "normal bitrate" means the bitrate the encoder would "normally" use without worrying about the operation of inserting splice points. The invention aims to enable encoders to operate in their (or as close as possible to their) normal way, and therefore minimize their constraints when they need to put splice points in the stream. This objective can be expressed as follows: r_a = r_n, and

where r_a is the assumed bit rate, r_n is the normal bit rate, and r_u is the bit rate actually used. To achieve this and other objectives, the present invention encompasses, in one of its aspects, a method of encoding a digital video bit stream which includes a plurality of successively-encoded pictures which includes the steps of constructing a seamless In Point in accordance with a constraint expressed by the following equation (1): (1) ddf__ιp < minimum {DD_tnrjp , (B_d/r_n_,_P) } , where dd_f_,_p represents an amount of time by which a first byte of the digital video bit stream after the seamless In Point must stay in a decoder buffer before it is decoded; B_d is a size of the decoder buffer; r_{n ιp} is a bit rate at which the digital video bit stream is normally encoded at the seamless In Point; and, DD_th_{r ιp} represents a prescribed minimum decoding delay threshold for seamless In Points, where DD_thr__ιp is defined by the following equation (2): (2) DD_{thr ιp} = ((MSR/r_{n p}) * SDD) + T_p_, * (1 - (MSR/r_njp)), where T_P is a time needed for displaying a picture which starts being displayed at a time of decoding of a last picture prior to the seamless In Point; SDD is a prescribed splice decoding delay; and, MSR is a prescribed maximum splice rate, and then inserting the seamless In Point at a first point in the digital video bit stream at which another digital video bit stream can be spliced into the digital video bit stream. The decoder can derive the decoding delays dd_f__ιp and dd_f__op, for example, from the vbv_delay field, the PCR field (program clock reference), or the DTS field (display time stamp) of the MPEG bitstream. The present invention encompasses, in another of its aspects, a method of encoding a digital video bit stream which includes a plurality of successively-encoded pictures which includes the steps of constructing a seamless Out Point in accordance with a constraint expressed by the following equation (3): (3) DD_thr__op < ddf op < (B_d/r_n__op), where dd_f__op represents an amount of time by which a first byte of the digital video bit stream after the seamless Out Point must stay in a decoder buffer before it is decoded; r_n__op is a bit rate at which the digital video bit stream is normally encoded at the seamless Out Point; and, DD_thr__op represents a prescribed minimum decoding delay threshold for seamless Out Points, where DD_th_r_o_P is defined by the following equation (4): (4) DD_{thr op} = ((MSR/r_{n op}) * SDD) - T_{p o} * ((MSR/r_n__op) - 1 ), where T_{p 0} is a time needed for displaying a picture which starts being displayed at a time of decoding of a last picture prior to the seamless Out Point; and then inserting the seamless Out Point at a first point in the digital video bit stream at which it can be spliced into another digital video bit stream. The present invention encompasses, in another of its aspects, a method for splicing a first digital video bit stream at an Out Point into a second digital video bit stream at an In Point, which includes the steps of measuring dd_f__op, measuring dd_f__ιp, computing a difference W equal to dd_f__op - dd_f__ιp, switching from the Out Point, and, after switching from the Out Point, waiting an amount of time equal to W, and then switching into the In Point.

In yet another of its aspects, the present invention encompasses a method for splicing a first digital video bit stream having at an Out Point into a second digital video bit stream at an In Point, which includes the steps of measuring dd_f__op, measuring ddf_,_p, computing a difference W equal to dd_f__o - dd_fjp, switching from the Out Point, streaming the second digital video bit stream at a first rate R_w < r_{n op} for a first time period having a duration equal to W, and, then streaming the second digital video bit stream for a second time period having a duration dd'f__ιp at a second rate R_{sp n} defined by the following equation: R_sp_,_n = ((r_n__ip * dd _f_,p) - (Rw * W))/dd'_f__ip, where dd'_f_,_P = dd_fjp - T_p_₀-

In yet another of its aspects, the present invention encompasses a method for splicing a first digital video bit stream at an Out Point into a second digital video bit stream at an In Point, which includes the steps of measuring dd_f__op, measuring dd_f__ιp, computing a difference W equal to dd_f__op - dd_fjp, computing a stream bit rate R_sp__op according to the following equation (1): ( 1 ) Rsp_o_P = (ddf__ιp/ddf_op) * r_njp; determining if R_{sp 0}p < r_n__op; if it is determined that R_sp_o_P ≤ i"n_o_P, then switching from the Out Point and streaming the second digital video bit stream at the rate R_sp__op for the duration of ddf__op; and, if it is determined that R_sp_₀p > i"_n_o_P, then switching from the Out Point and performing a selected one of the following two possible procedures: (i) waiting an amount of time equal to W, and then switching into the In

Point; or, (ii) (A) streaming the second digital video bit stream at a first rate R_w < r_{n op} for a first time period having a duration equal to W; and, (B) then streaming the second digital video bit stream for a second time period having a duration ddY _P at a second rate R_sp_i_π defined by the following equation (2): (2) R_spJn = ((r_nJp * ddV_,_p) - (Rw * W))/ddV _P, where dd'_{f ιp} = dd_fjp - T_{p 0}.

In still other of its aspects, the present invention encompasses a video signal which includes a digital video bit stream encoded in accordance with any one or more of the encoding methods of the present invention; a video signal which includes first and second digital video bit streams spliced together in accordance with any of the splicing methods of the present invention; and, an encoder for implementing any one of the encoding methods of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other objects, features, and advantages of the present invention will become apparent from the following detailed description taken in conjunction with the accompanying drawings, in which:

FIG. 1 is combination flow chart and graphical depiction of a method used by an encoder in constructing a seamless splice (In or Out) point, in accordance with the teachings of co-pending U.S. Application Serial No. 08/829,124; FIG. 2 is a combination flow chart and graphical depiction of a method used by an encoder in constructing a seamless splice points, in accordance with the teachings of co-pending U.S. Application titled Video Buffer for Seamless Splicing of MPEG Streams filed November 25, 1997, and, FIG. 3 is a combination flow chart and graphical depiction of a method used by an encoder in constructing a seamless splice points, in accordance with the preferred embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

A flow chart of the methodology of constructing seamless In and Out Points in accordance with the present invention is depicted in FIG. 3, as well as graphs which illustrate exemplary uses of the present invention. Like reference numerals in this Figure denote like blocks in FIGS. 1 and 2. In accordance with the conventional approach, when constructing a seamless Out Point, when the normal encoder bit rate r_{n op} is smaller than MSR, the encoder must assume the higher MSR rate, i.e., r_a = MSR > r_n__op.

In accordance with one aspect of the present invention, an encoder can assume its normal bit rate r_{n op} when constructing a seamless Out Point, by selecting a value for the decoding delay ddf__op (which is larger than SDD) such that dd_f__op meets the constraint expressed by the following equation (1): (1 ) DD,hr_op < ddf op < (B_d / r_{π op}), where B_d is the decoder buffer size, and DD_th_r__op is a decoding delay threshold for a seamless Out Point which has a value defined by the following equation (2): (2) DD._hr. op = ((MSR/r„_„p) * SDD ) T_{p o} * ((MSR/r_n__op) - 1 ), where T_{p o} is the time needed for displaying the picture which starts displaying at the time of decoding of the last picture prior to the seamless Out Point.

Thus, in accordance with the present invention, an encoder can construct a seamless Out Point when the normal encoder bit rate r_{n o} is smaller than MSR, by selecting a value for the decoding delay dd_f__op which meets the constraints specified in equations (1) and (2) above. This enables the encoder to assume its normal bit rate r_{n op} and not the higher bit rate MSR, thereby significantly improving the quality of pictures constructed around a seamless Out Point.

A particular value for dd_f__op which always meet the constraints expressed in equations (1) and (2) is defined by the following equation (3): (3) dd_{f 0p} = (MSR/r_{n op})* SDD.

Therefore, in accordance with the present invention, an encoder can construct a seamless Out Point when the normal rate r_{n op} is smaller than MSR, by selecting the particular value for the decoding delay ddf__op defined by equation (3) . This enables the encoder to assume its normal bit rate r_{n op} and not the higher bit rate MSR, thereby significantly improving the quality of pictures constructed around a seamless Out Point. The step of selecting dd_f__op defined by equation (3) is denoted 16 in FIG. 3. The corresponding buffer boundary constraints are shown by the block 17. If MSR is not greater than r_n__op, the encoder has the option for the decoder delay ddf__op to select a value > SDD (denoted 26). The corresponding buffer boundary constraints are shown by the block 27.

Since equation (3) always satisfies the constraints expressed by equations (1) and (2), and further, since MSR and SDD are always selected such that (MSR*SDD < B_d), then defining a particular value for dd_f__op by using equation (3) provides the following enumerated benefits:

1. It ensures the construction of a seamless Out Point without the need to evaluate the relationship among the quantities B_d, r_{n op}, MSR, and SDD.

2. It eliminates the need for knowing the value of T_{p o} when evaluating the decoding delay threshold DD_th_r__o as expressed in equation (2). This could be useful for video sequences that have variable picture-display times.

3. It requires less buffering at the splicer, as will be developed hereinafter. 4. It is easier to specify in an industry-wide standard (e.g. SMPTE or MPEG).

In accordance with the conventional approach, when constructing a seamless In Point when the normal encoder bit rate r_n_,_p is larger than MSR, the encoder must use the lower MSR rate, i.e., r_u =MSR< r_{n ιp} for an amount of time equal to SDD after the seamless In Point.

In accordance with the present invention, an encoder can use its normal bit rate r_n__ιp when constructing a seamless In Point, by selecting a value for the decoding delay dd_fJP which is smaller than SDD, such that dd_f_,_p meets the constraint expressed by the following equation (4): (4) dd_f__ip < minimum {DD_{thr p} , (B_d / r_njp) } , where B is the decoder buffer size, and DD_tr,r__ip is a decoding delay threshold for an In Point which has a value defined by the following equation (5): (5) DD,hr_jp = ((MSR/r_njp) * SDD ) + T_p_, * ( 1 - (MSR/r_njp) , where T_p_, is the time needed for displaying the picture which starts displaying at the time of decoding of the last picture prior to the seamless In Point.

Thus, in accordance with the present invention, an encoder can construct a seamless In Point when the normal encoder bit rate r_n_,_p is larger than MSR, by selecting a value for the decoding delay ddf__ιp which satisfies the constraints expressed in equations (4) and (5). This enables the encoder to use its normal bit rate r_njp and not the lower bit rate MSR, thereby significantly improving the quality of pictures constructed around a seamless In Point.

A particular value for dd_f_,_p which always meets the constraints expressed in equations (4) and (5) is defined by the following equation (6): (6) ddf__ιp = (MSR/r_n_,_p) * SDD. Thus, in accordance with the present invention, an encoder can construct a seamless In Point when the normal encoder bit rate r_{n ιp} is larger than MSR, by selecting the particular value for the decoding delay dd_f__ιp defined by equation (6). This enables the encoder to use its normal bit rate r_n_,_p and not the smaller bit rate MSR, thereby significantly improving the quality of pictures constructed around a seamless In Point. The step of selecting ddf_,_p defined by equation (6) is denoted 18 in FIG. 3. The corresponding buffer boundary constraints are shown by the block 19. If MSR is greater than r_n__φ, the encoder has the option for the decoder delay dd_f__ιp to select a value < SDD (denoted 28). The corresponding buffer boundary constraints are shown by the block 29. Since equation (6) always satisfies the constraints expressed by equations (4) and (5), and further, since MSR and SDD are always selected such that (MSR*SDD < B_d), then equation (6) provides the following enumerated benefits: 1) It ensures the construction of a seamless In Point without the need to evaluate the relationship among the quantities of B_d, r_{n ιp}, MSR, and SDD. 2) It eliminates the need for knowing the value of T_p_, when evaluating the decoding delay threshold DD_thr_,_p as expressed in equation (5). This could be useful for video sequences that have variable picture-display times.

3) It requires less buffering at the splicer, as will be developed hereinafter.

4) It is easier to specify in an industry-wide standard (e.g. SMPTE or MPEG).

A splicer operating on digital video bit streams that include Seamless Splice Points (SSP,) constructed in accordance with the present invention can use the following procedures to achieve seamless splicing (i.e., without causing the decoder buffer underflow or overflow): 1) The decoding delay (ddf__op) of the seamless Out Point is measured.

2) The decoding delay (dd_f__ιp) of the seamless In Point is measured.

3) The difference (W) between the two decoding delays is computed in accordance with the following equation (7): (7) W = ddf_o_P - dd_fJP

4) After switching from the seamless Out Point, the splicer can use any one of the following procedures a, b, or c (the desired procedure could be selected by the user or automatically under software control): a) Wait an amount of time equal to W before switching into the seamless In Point. For example, the splicer can send NULL packets for the duration of W. b) Use the following steps: i.) Stream the In Point stream (e.g., first digital video bit stream)at a rate (Rw) less than or equal to the Out Point rate (i.e. r_{n op}) for the duration of W, as expressed by the following equation (8): (8) Rw < r_n__op ii.) Then, stream the In Point stream for the duration dd'_f_,_p at a rate R_sp__m defined by the following equation (9): (9) R_sp_,n = ( (r_n_,_P *dd _fJP) - (Rw * W) )/dd'.f_,_p, where dd'_f_,_p = dd_f__ip - T_{p 0}. c) Use the following steps: i.) Compute the bitrate R_sp_o_P in accordance with the following equation (10): ( 10) R_sp_op = (ddf_,_p/ddf_op)* r_njp ii.) If Rs_P__oP ≤ r_n_o_P, Then: stream the In Point stream for the duration ddf__op at the rate R_{sp op}- Else: use either procedure (a) or (b) above.

It should be appreciated by those skilled in the pertinent art that seamless splicing is possible between a Seamless Splice Point (SSP) constructed in accordance with other known techniques and a Seamless Splice Point (SSP,) constructed in accordance with the present invention. In a similar vein, it should also be appreciated that the above- described methods of splicing in accordance with the present invention for splicing between an SSP, and another SSP„ can also be used to splice from an SSP to an SSP„ from an SSP, to an SSP, and from an SSP to another SSP. (In the last case, W will be zero). Further, the splicing technique of the present invention can be used to splice between non-seamless In and Out Points, or between a non-SSP and an SSP (or SSPj), although, in these cases, it should be recognized that the splice will not be seamless unless the following constraints are satisfied: dd_f__op ddf ip , and r_n__op > r_n_i_P. Further, the size of the buffer (B_sp) needed in the splicer to implement the present invention can be computed from the maximum value W_maχ that the duration W can have, in accordance with the following equation (11):

where R_max is the maximum bit rate the splicer can receive. If equations (3) and (6) are used to construct seamless In and Out

Points in accordance with the present invention, then W_max can be expressed by the following equation (12): W max = W max 1 = ((MSR / r_{min op} ) - (MSR / r_{max φ} )) * SDD

(12) = MSR * SDD * ((l / r_min__op) - (l / r_maxJp)) < B_d * ((l / r_τnin__op) - (l / r_maxJp)). However, if Equations (l)/(2) and Equations (4)/(5) are used to construct seamless In and Out Points in accordance with the present invention, then W_max can be expressed by the following equation (13):

(13) Wmax ⁼ W_maχ_2 = B_d/r_min_o_P By comparing Equations (12) and (13), it is clear that W_maχ_₂ is always larger than W_maxj. Therefore, as mentioned previously, using Equations (3) and (6) to construct seamless In and Out Points reduces the size of the buffer needed in the splicer when compared with the buffer size needed when the more general thresholds of Equations (l)/(2) and (4)/(5) are used for constructing the seamless In and Out Points. Although the present invention has been described in detail hereinabove, it should be clearly understood that many variations and/or modifications of the basic inventive concepts taught herein which may appear to those skilled in the pertinent art will still fall within the spirit and scope of the present invention as defined in the appended claims. Further it should be appreciated by those skilled in the art that the implementation of the encoding and splicing methods of the present invention described hereinabove in hardware and/or software is routine and straightforward. For example, simple programming of programmanble logic circuitry of an encoder or splicer can be used to implement the methods of the present invention.

Claims

CLAIMS:

1. A method of encoding a digital video bit stream which includes a plurality of successively-encoded pictures, including the steps of: constructing a seamless In Point in accordance with a constraint expressed by the following equation (1): ( 1 ) ddf_,_p < minimum {DD_thr_i_P , (B_d/r_n__ιp) } , where dd_f_,_p represents an amount of time by which a first byte of the digital video bit stream after the seamless In Point must stay in a decoder buffer before it is decoded; B_d is a size of the decoder buffer; r_{n ιp} is a bit rate at which the digital video bit stream is normally encoded at the seamless In Point; and, DD,_{hr p} represents a prescribed minimum decoding delay threshold for seamless In Points, where DD_thr_,_P is defined by the following equation (2): (2) DD_{thr ιp} = ((MSR/r_{n ιp}) * SDD) + T_p_, * (1 - (MSR/r„__ιp)), where T_p_, is a time needed for displaying a picture which starts being displayed at a time of decoding of a last picture prior to the seamless In Point; SDD is a prescribed splice decoding delay; and, MSR is a prescribed maximum splice rate; and, inserting the seamless In Point at a point in the digital video bit stream at which another digital video bit stream can be spliced into the digital video bit stream.

2. The method as set forth in Claim 1, further including the steps of: determining if MSR > r_n_,_p; selecting ddf _P to have a value < SDD if it is determined that MSR is greater than r_n_,_p; and, selecting ddf_,_p to have a value defined by the following equation (3) if it is determined that MSR is not greater than r_{n ιp}: (3) dd_f_,_p = (MSR/r_n_,_p) * SDD.

3. The method as set forth in Claim 1, wherein r_n_,_p > MSR.

4. The method as set forth in Claim 1, wherein dd_f ,_p is selected to have a value defined by the following equation (3): (3) dd_{f p} = (MSR/r_n_,_p) * SDD.

5. The method as set forth in Claim 1, wherein MSR and SDD are selected to have values that satisfy a constraint expressed by the following equation (4): (4) MSR*SDD < B_d.

6. The method as set forth in Claim 1, wherein SDD and MSR are selected to have specific values in accordance with a coding standard.

7. A method of encoding a digital video bit stream which includes a plurality of successively-encoded pictures, including the steps of: constructing a seamless Out Point in accordance with a constraint expressed by the following equation (1): ( 1 ) DD_thr_o_P < ddf_o_P < (B_d/r_{n op}), where dd_f__op represents an amount of time by which a first byte of the digital video bit stream after the seamless Out Point must stay in a decoder buffer before it is decoded; B_d is a size of the decoder buffer; r_{n op} is a bit rate at which the digital video bit stream is normally encoded at the seamless Out Point; and, DD_th_r__op represents a prescribed minimum decoding delay threshold for seamless Out Points, where DD_th_r__o is defined by the following equation (2): (2) DD_tnr__op = ((MSR/r_n__op) * SDD) - T_{p o} * ((MSR/r_n__op) - 1), where T_{p o} is a time needed for displaying a picture which starts being displayed at a time of decoding of a last picture prior to the seamless Out Point; SDD is a prescribed splice decoding delay; and, MSR is a prescribed maximum splice rate; and, inserting the seamless Out Point at a point in the digital video bit stream at which it can be spliced into another digital video bit stream.

8. The method as set forth in Claim 7, further including the steps of: determining if MSR > r_{n op}; selecting dd_f__op to have a value > SDD if it is determined that MSR is not greater than r_{n o} , within a constraint that ddf__op * r_{n op} < B_d; and, selecting dd_f__op to have a value defined by the following equation (3) if it is determined that MSR is greater than r_{n op}: (3) dd_f__op = (MSR/r_n__op) * SDD.

9. The method as set forth in Claim 7, wherein r_{n op} < MSR.

10. The method as set forth in Claim 7, wherein dd_f__op is selected to have a value defined by the following equation (3): (3) dd_{f o}p = (MSR/r_{n op}) * SDD.

11. The method as set forth in Claim 7, wherein MSR and SDD are selected to have values that satisfy a constraint expressed by the following equation (4): (4) MSR*SDD < B_d.

12. A method for splicing a first digital video bit stream at an Out Point into a second digital video bit stream at an In Point, the method including the steps of: measuring ddf__op; measuring dd_fJP; computing a difference W equal to ddf__op - ddπ_p; switching from the Out Point; and, after switching from the Out Point, waiting an amount of time equal to W, and then switching into the In Point.

13. The method as set forth in Claim 12, wherein the step of waiting is carried out by sending null packets for the duration of the waiting time W.

14. The method as set forth in Claim 12, wherein the In Point is a seamless In Point and the Out Point is a non-seamless Out Point, whereby the switching is seamless if the following two constraints are satisfied: a) ddf op > dd_f p ; and, θ) Tn_o_P ^■■* ϊn_i_P-

15. The method as set forth in Claim 12, wherein the In Point is a non- seamless In Point and the Out Point is a non-seamless Out Point, whereby the switching is seamless if the following two constraints are satisfied:

o) r_n_o_P -> r_njp.

16. The method as set forth in Claim 12, wherein the In Point is a non- seamless In Point and the Out Point is a seamless Out Point, whereby the switching is seamless if the following two constraints are satisfied: a) ddf op > ddf ,_p; and,

17. A method for splicing a first digital video bit stream having at an Out Point into a second digital video bit stream at an In Point, the method including the steps of: measuring ddf__op; measuring dd_f ,_p; computing a difference W equal to dd_f__op - ddf ,_p; switching from the Out Point; streaming the second digital video bit stream at a first rate R_w r_n__op for a first time period having a duration equal to W; and, then streaming the second digital video bit stream for a second time period having a duration dd'f ,_P at a second rate R_sp ,_n defined by the following equation: sp n = ((Γ„_,_P * ddVj_P) - (Rw * W))/dd'_f_,_P, where dd'_{f ιp} = dd_fjp - T_{p 0}.

18. A method for splicing a first digital video bit stream at an Out Point into a second digital video bit stream at an In Point, the method including the steps of: measuring dd_f__op; measuring ddf ,_p; computing a difference W equal to dd_f__op - dd_f _,_p; computing a stream bit rate R_sp__oP according to the following equation (1): ( 1 ) R_sp_o_P = (ddf ,p/ddf_op) * r_n ,_p; determining if R_sp__op < r_n__op; if it is determined that R_sp__op ≤ r_n__oP, then switching from the Out Point and streaming the second digital video bit stream at the rate R_sp__op for the duration of ddf__op; if it is determined that R_sp__op r_n_o_P, then switching from the Out Point and performing a selected one of the following two possible procedures: (i) waiting an amount of time equal to W, and then switching into the In Point; or, (ii) (A) streaming the second digital video bit stream at a first rate R_w < r_{n op} for a first time period having a duration equal to W; and, (B) then streaming the second digital video bit stream for a second time period having a duration ddγ _ιp at a second rate R_{sp n} defined by the following equation (2): (2) R_{sp ιn} = ((r_{n ιp} * ddVjp) - (Rw * W))/dd'f_jP, where ddV ,_P = dd ,_p - T_{p 0}.

19. A video signal comprising a digital video bit stream having a seamless In Point constructed in accordance with a constraint expressed by the following equation (1): (1) dd_fjp < minimum {DD_{thr ιp} , (B_d/r_n ,_p) } , where dd_{f P} represents an amount of time by which a first byte of the digital video bit stream after the seamless In Point must stay in a decoder buffer before it is decoded; B_d is a size of the decoder buffer; r_{n ιp} is a bit rate at which the digital video bit stream is normally encoded at the seamless In Point; and, DD_thr_ι_P represents a prescribed minimum decoding delay threshold for seamless In Points, where DD _rjP is defined by the following equation (2): (2) DD_{thr ιp} = ((MSR/r_{n i}p) * SDD) + T_Pj * (1 - (MSR/r„__ιp), where T_Pj is a time needed for displaying a picture which starts being displayed at a time of decoding of a last picture prior to the seamless In Point; SDD is a prescribed splice decoding delay; and, MSR is a prescribed maximum splice rate.

20. A video signal comprising a digital video bit stream having a seamless Out Point constructed in accordance with a constraint expressed by the following equation (1): ( 1 ) DD_thr_op < ddf op < (B_d/r_n__op), where dd_f__op represents an amount of time by which a first byte of the digital video bit stream after the seamless Out Point must stay in a decoder buffer before it is decoded; B is a size of the decoder buffer; r_{n op} is a bit rate at which the digital video bit stream is normally encoded at the seamless Out Point; and, DD_th_r_o represents a prescribed minimum decoding delay threshold for seamless Out Points, where DD_tnr_o_P is defined by the following equation (2): (2) DD._hr.o_p = ((MSR/r„_op) * SDD) - T_{p o} * ((MSR r_{n op}) - 1 ), where T_{p 0} is a time needed for displaying a picture which starts being displayed at a time of decoding of a last picture prior to the seamless Out Point; SDD is a prescribed splice decoding delay; and, MSR is a prescribed maximum splice rate.

21. An encoder for encoding a digital video bit stream which includes a plurality of successively-encoded pictures, including: means for constructing a seamless In Point in accordance with a constraint expressed by the following equation (1): (1) ddf _ip < minimum {DD,_hr_ι_P , (B_d/r_{n ιp}) } , where ddf ,_p represents an amount of time by which a first byte of the digital video bit stream after the seamless In Point must stay in a decoder buffer before it is decoded; B_d is a size of the decoder buffer; r_{n ip} is a bit rate at which the digital video bit stream is normally encoded at the seamless In Point; and, DD _nrjP represents a prescribed minimum decoding delay threshold for seamless In Points, where DD_thrjP is defined by the following equation (2): (2) DD_{thr ιp} = ((MSR r„__ιp) * SDD) + T_P * (1 - (MSR/r_n ,_p), where T_p , is a time needed for displaying a picture which starts being displayed at a time of decoding of a last picture prior to the seamless In Point; SDD is a prescribed splice decoding delay; and, MSR is a prescribed maximum splice rate; and, means for inserting the seamless In Point at a first point in the digital video bit stream at which another digital video bit stream can be spliced into the digital video bit stream.

22. An encoder for encoding a digital video bit stream which includes a plurality of successively-encoded pictures, including: means for constructing a seamless Out Point in accordance with a constraint expressed by the following equation (1): (1 ) DDthr φ < ddf__op < (Bd r„_op), where dd_f__op represents an amount of time by which a first byte of the digital video bit stream after the seamless Out Point must stay in a decoder buffer before it is decoded; r_n__op is a bit rate at which the digital video bit stream is normally encoded at the seamless Out Point; and, DD,_hr_op represents a prescribed minimum decoding delay threshold for seamless Out Points, where DD_thr_o_P is defined by the following equation (2): (2) DD_{thr op} = ((MSR/r„_op) * SDD) - T_{p o} * ((MSR r_n__op) - 1 ), where T_{p o} is a time needed for displaying a picture which starts being displayed at a time of decoding of a last picture prior to the seamless Out Point; and, means for inserting the seamless Out Point at a second point in the digital video bit stream at which it can be spliced into another digital video bit stream.

23. A splicer for splicing a first digital video bit stream at an Out Point into a second digital video bit stream at an In Point, the splicer including: means for measuring dd_f__op and for measuring dd_f _,_p; means for computing a difference W equal to dd_f__op - dd_f ,_p; and, means for switching from the seamless Out Point, and, after switching from the

Out Point, waiting an amount of time equal to W, and then switching into the In Point.

24. A splicer for splicing a first digital video bit stream at an Out Point into a second digital video bit stream at an In Point, the splicer including: means for measuring dd_f__op and for measuring ddf ,_p; means for computing a difference W equal to dd_f__op - dd_{f ιp}; means for switching from the Out Point; and, means for streaming the second digital video bit stream at a first rate R_w < r_n__op for a first time period having a duration equal to W, and, for then streaming the second digital video bit stream for a second time period having a duration dd ,_P at a second rate R_spjn defined by the following equation: R_sp_,n = ((r_n p * ddVj_P) - (Rw * W))/dd'_fj_P, where ddV ι_P = ddf ,_p - T_{p 0}-

25. A splicer for splicing a first digital video bit stream at an Out Point into a second digital video bit stream at an In Point, the splicer including: means for measuring dd_f__op and for measuring dd_{f ιp}; means for computing a difference W equal to dd_f__op - dd_f ,_p, and for computing a stream bit rate R_sp_o_P according to the following equation (1): (1) R_s _o_P = (ddf _ip/ddf_o_P) * r_njp; means for determining if R_sp__o < r_n__op; means for switching from the Out Point; means for streaming the second digital video bit stream at the rate R_sp__oP for the duration of ddf__o , if it is determined that R_sp__op ≤ r_{n op}; and, means for performing a selected one of the following two possible procedures: (1) waiting an amount of time equal to W, and then switching into the In Point; or, (2) (A) streaming the second digital video bit stream at a first rate R_w < r_{n o} f°^r a first time period having a duration equal to W; and, (B) then streaming the second digital video bit stream for a second time period having a duration ddV,_p at a second rate R_{sp ι}, defined by the following equation (2): (2) R_sp _,„ = ((r„__ιp * ddV_jP) - (Rw * W))/ddV_jP, where dd' _ιp = dd_{f ιp} - T_p0, if it is determined that R_{sp op} r_n__op-