MXPA99010708A - Spread of video data in video currents progressively renova - Google Patents

Abstract

A method is provided for replacing compressed video from a first source with video from a second source at a predefined insertion point in a video data stream. If the video from the second source is progressively renewed video, it decompresses sufficiently before the insertion point to allow the recovery and recompression of a first video structure (in order of coding) from the second source, to be inserted in the second source. insertion point The first recovered video structure is then recompressed as an intra-coded structure (I). Then the insertion of the video from the second source at the insertion point begins using the structure I. The insertion point can indicate, for example, the beginning of a commercial cut in the video stream, in which case the Second source provides a commercial advertisement for its insertion in the commercial court. When the commercial advertisement is stored in the memory, it can be stored with the first structure of it (in order of coding) as a structure

Description

SPREAD OF VIDEO DATA IN PROGRESSIVELY RENEWED VIDEO CURRENTS BACKGROUND OF THE INVENTION The present invention relates to digital television and, more particularly to a method for the insertion of video or current switching in progressively renewed video streams. The insertion of commercial advertisements, insertion of promotional items, the direction of studies, the changing cameras, editing tapes and the like, are basic operations in the production and transmission of television. In current analog systems, the switching takes place in the vertical interval of the video signal. With the advent of digital television, as implemented, for example, in the rules of the Group of Experts of Moving Image (MPEG), switching or "splicing" with compressed bit streams is very complicated. In order to facilitate such operations, the MPEG-2 standard defines "splice points" where point signals (eg, a commercial) can be replaced by another signal (eg, a main program) in a video stream.

The Society of Engineers of Television and Image in Motion (SMPTE) is currently working on a preliminary draft of a standard for splicing MPEG-2 transport streams. See, for example, Proposed Standard of SMPTE PT20.02 / 010"Junction Points for MPEG-2 Transport Currents", Second Draft, July 1997, incorporated herein by reference. There are several reasons why MPEG video splicing is more difficult than uncompressed video switching. In order to explain this, it is necessary to understand some basic aspects about the MPEG transport stream. MPEG gets a lot of its compression capacity by sending only the changes between the different video structures. A first type of structure, known as a prediction or "P" structure, is used by the decoder to predict the structure from a previous "P" structure or from an internally-coded structure "I" in the video stream. The current merely carries "precise tuning" information to correct errors of an approximate prediction. A Structure I is compressed without predicting movement. Therefore, a complete structure can be reconstructed from a structure I without reference to any other structure. In this way, the errors in the predictions of the previous structure will be eliminated once a structure I arrives and is decoded. The bidirectional structures "B" are like the structures P, except that the prediction is made not only from of the previous structure I or P, but also from the following. For purposes of splicing, this means that although it is possible to leave an old stream after any current po I, it is not possible to leave the stream after a structure B since it would cut the inverse prediction reference structure for that structure B. From the point of view of the bitstream, this means that a first video stream can only be abandoned after a P oly structure has passed and all the B structures (if any) that follow immediately. The problems are more complicated when the bitstream of the compressed MPEG is renewed progressively. As was thought in the U.S. Patent. commonly assigned 5,057,916, by Krause et al. , entitled "Method and Apparatus for Renewing Sequential Video Images Compensated by Motion", in the progressively renewed streams the sequential video images are renewed by region at the same time. A different region in each of the plurality of video images communicates without compression during a renewal cycle. In this way, an image area defined by the region is progressively renewed by the non-compressed regions during the renewal cycle. Compression of video images can be controlled to prevent data contained in regions not yet compressed during a current refresh cycle from corrupting the data contained in regions that have been renewed during the current refresh cycle. For the insertion of commercial advertisements or for switching purposes, the compressed video of a first source is replaced with compressed video of a second source at a predefined insertion point in the video data stream. If the video of either or both sources, First and second, it is progressively renewed, then progressively renewed related artifacts will appear at one or both boundaries between the first and second video sources. It would be advantageous to provide a method for providing the splicing of progressively renewed video streams that reduces the remarkable artifacts. The present invention provides such a method. SUMMARY OF THE INVENTION According to the present invention, there is provided a method for replacing compressed video from a first source with compressed video from a second source, for example, at a predefined insertion point in a video data stream. The determination is made whether the video coming from the second source is progressively renewed video. The determination can be a real time determination (for example, based on the examination of the second source) or it can be the result of a prior knowledge of the second source. If the video coming from the second source is progressively renewed, then it decompresses sufficiently before the moment of replacement of the first video source for allow the recovery and recompression of a first replacement video structure from said second source before the replacement begins. It is noted that for purposes of the present disclosure, the "first replacement video structure" is the first structure in the order of coding the replacement video stream, which may differ from the order of deployment, as is well known in The matter. The first recovered video structure is recomputed as an intra-codified structure (I) Then, the insertion of the video from the second source (for example, at the predefined insertion point) begins by using structure I. Since structure I comprises all the information necessary to reconstruct a complete data structure and not it is progressively renewed, it allows a perfect commutation of the video data coming from the first source to the video data coming from the second source. The insertion point can indicate, for example, the beginning of a commercial cut in the video stream. In this case, the second source provides a commercial advertisement for its insertion in the commercial court. At the end of the commercial cut, the video coming from the first source starts again at a second insertion point in the data stream. As with the first splice, it is necessary to determine if the video coming from the first source is progressively renewed video. If so, that video is decompressed sufficiently before the second insertion point to allow the recovery and recompression of a first video structure (in the encoding order) from the first source, to be inserted at the second insertion point . The first recovered video structure, from the first source, is then recompressed to produce a structure I. The insertion of the video from the first source at the second insertion point begins with the use of structure I produced from the first source. A method is also provided for replacing the compressed video from a first source with compressed video, progressively renewed, substitute (e.g., at a predefined insertion point) in a video data stream. The substitute video (for example, a commercial advertisement) is stored with a first structure of the same, coded as an internally encoded structure (I). The insertion of the substitute video is started (for example, at the predefined insertion point) with the use of structure I as the first inserted structure. When the compressed video, coming from the first source, is progressively renewed and is about to start again at a second insertion point in the data stream that follows the first insertion point (for example, at the end of a commercial cut), the video from the first source must be processed to avoid artifacts related to the progressive renewal. In particular, the video of the first source is decompressed sufficiently before the second insertion point to allow the recovery and recompression of a first video structure from the first source to be inserted at the second insertion point. The first recovered video structure, from the first source, is recompressed to produce a structure I. The insertion of the video from the first source begins at the second insertion point with the use of structure I produced from the first source. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a diagram illustrating data streams, first and second, SI, S2, respectively, together with a third data stream S3 that compresses the SI current into which a portion of the current has been inserted. current S2; and Figure 2 is a diagram illustrating a data stream S4 created by changing the data stream SI to the data stream S2. DETAILED DESCRIPTION OF THE INVENTION The present invention provides a method for inserting commercial advertisements or other video segments into progressively renewed data streams and for changing from one bit stream to another when progressively renewing any of the bit streams. Figure 1 illustrates a first compressed digital video stream 10 (SI current). The video stream 10 has a first insertion point ti in which a commercial cut begins. The commercial cut ends at the insertion point t2. It is noted that the insertion points may or may not include points predetermined in the video stream. Since it is more typical to insert a video segment such as a commercial advertisement at a predefined time, it is expected that the predetermined insertion points will be used more frequently with the present invention. However, according to the invention, the use of predetermined insertion points is not required. For example, a local television studio may wish to insert a local segment into a main video stream at an arbitrary time, in which case the invention can be implemented without the use of a predetermined insertion point for the replacement of the main video with the video replacement. As indicated in figure 1, a commercial advertisement or other video segment Sic is inserted in the commercial section defined by the insertion points ti and t2. In particular, the inserted Sic video starts at the insertion point ti and ends at the insertion point t2, where the main video program transported by SI starts again at the end of the commercial cut. A similar video stream 12 (stream S2) comprises a main program S2 in which an S2c commercial advertisement is inserted. HE notes that the commercial advertisements Sic and S2c can be part of the SI and S2 currents, respectively, in which case they do not need to be inserted separately into the mainstream. However, in the case that commercial advertisements (or other video segments that are inserted) are independent video segments, they should be inserted into the main stream at the appropriate insertion point tl and terminated at the insertion point t2 when they return. to start the mainstream. The video stream 14 (current S3) represents a classic case of video splicing, in which part of the SI current is replaced by S2c. As indicated above, S2c may be part of the S2 stream or it may be an independent commercial advertisement or another independent video segment. If SI and / or S2 are progressively renewed video data streams, then artifacts related to progressive renewal will occur at one or both of SI and S2c limits. Such artifacts related to renovation can last as long as a second or more, depending on the speed of renovation. Therefore, when a commercial advertisement is inserted or When switching by one, one would see the artifacts each time the commutation was made. To avoid such artifacts while the commercial is switched by switching on a main data stream (i.e., by shifting the main current out and inserting the commercial advertisement by a cut), the commercial advertisement may be compressed in structure mode I, P, B However, when the commercial announcement ends and the progressively renewed main stream is re-inserted by switching, the artifacts of the progressive renewal will still appear. It is noted that the video structures in the stream illustrated in Figure 1 are in order of coding, which is different from the structure deployment order. When the structures are in coding order, the first structure of the segment inserted by switching will be a structure I. For purposes of the present description, the terms "structure" and "video structure" are used in a generic sense to refer to a portion of a video signal or a stream from which other portions are reconstructed by the use of , for example, a structure I, a structure P or a structure B. In this way, although typically a "structure" will be a "complete" video structure, no attempt is made to limit the meaning thereof to a complete video structure for purposes of the present description, since that implementations can exist or be developed where they are used, for example, substructures, fields, subfields or other processing. With respect to video processing at a field level instead of a structure level, it should be appreciated that fields I, fields P and fields B may be used. In such cases, these fields are equivalent to, and attempt to be covered by, the broad terms of "structure" and "video structure" as used in the processing of video signal structure. The present invention overcomes the problem of the artifacts related to the progressive renewal at the beginning of the processing of a structure to be inserted before the moment of insertion. First, the video to be inserted is decompressed sufficiently before the insertion point to allow the recovery and recompression of a first structure to be inserted. As noted above, the "first structure to be inserted "is the first structure in the order of coding and not necessarily the first structure in the order of deployment.Once this structure has been decompressed, it is recompressed as an intra-codified structure (I). - encoded are able to be reconstructed without referring to any other structure and therefore are not subjected to progressive renewal.After the first recovered structure is recompressed as a structure I, it is used as the first structure inserted at the insertion point. a first example, SI can be a compressed video stream, progressively renewed while S2 is not progressively renewed.If a sequence is not progressively renewed then it is a renewed structure I. In this case, the artifacts of the progressive renewal will appear in the commutation from S2 to SI, but not when going from SI to S2, therefore, a switchover can be made from SI to S 2 in order to insert S2c in a conventional manner, as it is done for the cases of the renewed structure I. However, the technique of the present invention should be used to avoid artifacts from the progressive renewal when commuting again from S2c to SI. According to the invention, when S2c is inserted, the decoding of SI begins at some time dt before the insertion point t2. In other words, the decoding of SI will start at the moment (t2-dt). After the recovery of a first structure (in order of coding) of SI to be inserted at the end of S2c, this structure or "image" is recompressed as all the internally encoded blocks, and the recompressed image is inserted at the end of S2c. beginning of the insertion point t2. The rest of the SI current may be the same as in the original stream. In another example, both SI and S2 are progressively renewed. In this case, when switching from the current SI to the video segment S2c, the decoding of S2c starts at some time dt 'before the insertion tl, that is, in (tl-dt'). A first structure retrieved (in order of coding) from S2c is recompressed as all blocks I. This first structure is inserted as the first structure of the video segment S2c at the insertion point tl. No special processing is required for the rest of the S2c video segment. When switching again from S2c to SI, the same technique followed in the previous example is used, that is, the first structure (in order of coding) of the SI current to be inserted at the insertion point t2, is recovered and recompresses all the blocks I for insertion in t2. Figure 2 illustrates a special case referred to as "switching". In this case, a current is switched (for example, a power coming from a first camera) to another current (for example, a power coming from a second camera). As illustrated in FIG. 2, current 16 (current S4) is created by switching SI to current S2 at the insertion point (or "switching point) tl, in the event that SI is progressively renewed but S2 does not , no special processing is required, however, in the case where S2 is progressively renewed (without taking into account if SI is progressively renewed), the IS decoding will start at a time dt before the switching point tl in order to that a first structure of the same (in order of coding) can be recompressed as an I structure to be used as the first structure when the currents are switched at the switching point tl. The moments dt and dt 'are established by knowing the number of structures that take to complete the progressive renewal of an image. For example, if the progressive renovation is completed within fifteen structures, dt and dt 'will be of a period of time at least equivalent to fifteen moments of structure and, more typically, will be greater than these (for example, thirty structures) to be sure to provide enough time to decode and recompress the first necessary structure of the inserted video portion. It should now be appreciated that the present invention avoids the artifacts of progressive renewal by beginning to decode a main stream, which is progressively renewed, several structures before an inserted portion of video ends (eg, a commercial advertisement). At the time when a commercial announcement is completed, the decoded image (eg, a P image) of the main stream is captured and recompiled as a structure I. The main stream is It then inserts by switching, starting with that structure I. The rest of the structures in the mainstream can be progressively renewed structures. It is noted that some artifacts may still exist due to the fact that structure I is not the true reference structure used to encode the main stream. However, generally such artifacts are significantly less objectionable than artifacts of progressive renovation that would otherwise appear. In addition, it is also possible to store the commercial advertisement or another inserted video segment in such a way that only the first structure of the segment is a structure I and the rest of the structures are progressively renewed. In this way, a progressively renewed commercial advertisement can be used without progressively renewed, harmful artifacts, at the time when the commercial advertisement is inserted by switching in the video stream. Although the invention has been described in connection with various specific embodiments, it should be appreciated that numerous adaptations and modifications may be made thereto without depart from the scope of the invention as set forth in the claims.

Claims (8)

1. NOVELTY OF THE INVENTION Having described the present invention, it is considered as a novelty and therefore the property described in the following claims is claimed as property. A method for replacing compressed video from a first source with compressed video from a second source in a video data stream, comprising the steps of: determining whether the video from the second source is progressively renewed video, and it is like this: (a) decompress the video from the second source sufficiently before the replacement of the first video source to allow recovery of a first replacement video structure from the second source before the replacement begins; (b) recompressing the first replaced video structure recovered as a mtra-coded structure (I); and (c) starting the replacement of the first video source with the second video source when using structure I.
2. 2. A method according to claim 1, characterized in that the replacement of the first video source with the second video source. begins with the insertion of structure I at a predefined insertion point in the video data stream.
3. A method according to claim 2, characterized in that the insertion point indicates the beginning of a commercial cut in the video stream and the second source provides a commercial advertisement for insertion in the commercial cut.
4. A method according to claim 3, characterized in that at the end of the commercial cut, the video from the first source starts again at a second point of insertion into the data stream, comprising the additional steps of: determining whether the video comes from from the first source is progressively renewed video, and if so: (a) decompress the video from the first source sufficiently before the second insertion point to allow the retrieval of a first video structure from the first source, to inserted in the second insertion point; (b) recompressing the first recovered video structure from the first source to produce an intra-codified structure (I); and (c) begin inserting the video from the first source in the second insertion point by using the structure I produced from the first source.
5. A method for replacing video from a first source with compressed video, progressively renewed, substitute, in a video data stream, characterized in that it comprises the steps of: storing the substitute video with a first structure thereof, coded as a embedded structure (I); and begin inserting the substitute video into the video data stream by using structure I as the first inserted structure.
6. 6. A method according to claim 5, characterized in that the insertion of the substitute video begins with the insertion of the structure I at a predefined insertion point in the video data stream. A method according to claim 6, wherein the compressed video from the first source is progressively renewed and restarted at a second insertion point in the data stream following the first insertion point, characterized in that the method comprises the additional steps of: (a) decompressing the video from the first source sufficiently before the second insertion point to allow recovery and recompression of a first video structure from the first source to be inserted at the second insertion point; (b) recompress the first recovered video structure, from the first source, to produce an intra-codified structure (I); and (c) starting the insertion of the video from the first source at the second insertion point by using structure I produced from the first source. A method according to claim 7, characterized in that the insertion points, first and second, define a commercial cut in the video stream and the substitute video is a commercial advertisement for its insertion in the commercial cut.