JP3755206B2 - Editing device and editing method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a video material editing apparatus and editing method . Specifically, when editing using multiple video materials such as digital video signals compressed by inter-frame prediction such as MPEG, the video material decompression and recompression processing section in the editing section should be made as short as possible. The present invention relates to an editing apparatus and an editing method capable of shortening the time required for image processing and preventing image quality deterioration associated with editing.
[0002]
[Prior art]
In an editing apparatus that edits video material, editing is often performed in which a plurality of, for example, two video materials are connected at an appropriate place, or a part of one video material is cut out.
[0003]
[Problems to be solved by the invention]
In recent video materials, data is often compressed. For example, the opportunity to use a compressed video signal using inter-frame prediction such as MPEG is increasing. For this reason, compressed video signals are often provided to editing apparatuses as video material.
[0004]
In this way, when editing video materials of video signals compressed by inter-frame prediction etc., the video material after cut-in is the original video material for the cut-in side of the video material to be edited. The image is once decompressed into an uncompressed state, editing is performed in that state, and then compressed again to form an edited video material.
[0005]
For example, as shown in FIGS. 5A and 5B, two image materials A and B are prepared. For the image material A, up to a point p (cutout point) is used, and in the other image material B, a point q (cut) The editing is assumed to be the video material C shown in FIG.
[0006]
In this case, as shown in FIG. 6D, in one video material A, data expansion processing is performed on a predetermined area including at least the editing point p. In the other video material B, as shown in FIG. 5E, all the data after the area immediately before including the editing point q is expanded. Then, before and after the editing points p and q, the joint processing (recompression using the materials A and B) is performed using the decompressed video material, and after that, the decompressed video material B is recompressed. Thus, a video material C as shown in FIG.
[0007]
This is due to the following reason.
As is well known, there are three types of video signals such as I picture, P picture, and B picture in a video signal compressed by inter-frame prediction such as MPEG.
An I picture (Intra Picture: an intra-frame encoded image) is a screen encoded only from information on the frame, and is generated without using inter-frame prediction.
A P picture (Predictive Picture: an inter-frame forward direction predictive encoded image) is generated from information on the frame and information on a frame immediately preceding the frame.
A B picture (Bidirectionally predictive picture) is generated based on information on the frame and information on frames preceding and following the frame.
[0008]
The P picture and B picture use the information of the previous frame in time or the information of the previous and subsequent frames in the compression generation of these signals, but the picture arrangement, that is, the image type and the image type image When the sequence (GOP) is ignored and the image is simply cut and pasted, the P picture or B picture that is decompressed with reference to other frame images may not be successfully decompressed at this joint, and the image quality deteriorates.
[0009]
Therefore, in the conventional method, all the materials necessary for editing after editing are temporarily decompressed to the original uncompressed state, and the editing operation is performed again. We are doing compression work.
[0010]
In this way, all the materials necessary for editing need to be expanded once after the video material B is uncompressed, and then compressed again. Need a lot of time.
[0011]
Also, in the picture reconstruction after the editing point, the video material B is expanded and recompressed. Therefore, a quantization error occurs in the recompressed image data, and the image quality of the entire video is deteriorated. It will end up.
[0012]
The present invention has been made in consideration of the above points, and proposes an editing apparatus and an editing method capable of editing a video signal compressed by inter-frame prediction in a short time while minimizing deterioration in image quality. It is something to try.
[0013]
[Means for Solving the Problems]
In order to solve the above-described problems, in the editing apparatus of the present invention, the first image type identifying means for identifying the image type of the compressed video data of the first video material, and the compressed video data of the second video material. Second image type identifying means for identifying the image type, and the first I identified by the first image type identifying means by performing a forward search in time from the cutout point set in the first video material. The first video material with the picture as a decoding start point and the first I picture or P picture identified by the first image type identification means as a result of backward search from the cutout point in time. A first decoding means for decoding the compressed video data, and a forward search in time from the cut-in point set in the second video material, and the second image type identification means The first I picture is set as a decoding start point, the first I picture or P picture identified by the second image type identifying means by backward searching in time from the cut-in point is set as the decoding end point, and the second A second decoding means for decoding the compressed video data of the video material, and the first video data decoded by the first decoding means and the second video using the cut-out point and the cut-in point as connection points. Connecting means for connecting the second video data decoded by the decoding means, and re-encoding means for re-encoding the video data connected by the connecting means to generate re-compressed video data. is there. The editing method of the present invention also includes a first image type identifying step for identifying the image type of the compressed video data of the first video material, and a second for identifying the image type of the compressed video data of the second video material. The first I picture identified in the first image type identification step by temporally searching forward from the cut-out point set in the first video material, and the decoding start point, The compressed video data of the first video material is decoded with the first I picture or P picture identified in the first image type identification step as a decoding end point by backward searching from the cutout point in time. First decoding step and starting to decode the first I picture identified in the second image type identification step by temporally searching forward from the cut-in point set in the second video material And using the first I picture or P picture identified in the second image type identification step as a result of backward searching from the cut-in point in time, the compressed video data of the second video material is used as the decoding end point. A second decoding step for decoding, the first video data decoded by the first decoding step and the second decoding step decoded by the second decoding step, using the cut-out point and the cut-in point as connection points. And a re-encoding step of re-encoding the video data connected in the connection step to generate re-compressed video data.
[0014]
In the present invention, in the video material A on the cutout side, decompression processing (baseband processing) is performed on the compressed data of the smallest image type that can be decompressed using the I picture, P picture, and B picture. Is called. For video material B on the cut-in side, similar decompression processing is performed on the compressed data in at least one image sequence (GOP) including the cut-in point. This is because the 1 GOP image data including the cut-in side needs to be reused as the image data after editing on the cut-in side.
[0015]
Then, it is recompressed using the decompressed data A ′ and B ′. In the example of FIG. 2, the recompressed data is used from two frames before the cutout point Aout. When the cutin point Bin is reached, recompression processing is performed according to the image type using the decompressed data A ′ and B ′. The arrangement of image types in the editing section follows the picture arrangement of the video material A, for example. The compressed data itself of the video material B is used from the frame where the I picture of the video material B is located.
[0016]
This minimizes the number of frames to be decompressed and compressed, thereby shortening the editing processing time. In addition, since the image data is decompressed and recompressed only in the minimum necessary area, the compressed video materials A and B are used as they are for the material away from the editing point. Therefore, it is possible to minimize the deterioration of image quality due to the decompression and compression processing.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
Next, an embodiment of an editing apparatus according to the present invention will be described in detail with reference to the drawings.
[0018]
The editing apparatus 10 shown in FIG. 1 targets two video materials (editing materials) A and B. In the figure, a video material A is recorded in the data reproducing apparatus 11 in a state compressed in advance by inter-frame prediction using a technique such as MPEG. Similarly, the video material B is recorded in the data reproducing apparatus 14 in a state compressed in advance by inter-frame prediction using a technique such as MPEG.
[0019]
The reproduced video materials A and B are sent to the data decompressors 13 and 16 and the image type discriminators 12 and 15, respectively. The image type discriminators 12 and 15 select the editing points of the reproduced video materials A and B. The picture configuration (I / P / B picture) in the included image sequence (1 GOP) is automatically determined. The reason why the image type that is the type of picture is determined is that another picture having a different image type may be used in order to correctly restore the compressed video data to the original video signal.
[0020]
The determined image type information is sent to the CPU 20 and used as recompression information at the time of editing. The data decompressors 13 and 16 decompress the compressed data so that the pictures in the minimum image sequence required for editing are in the original uncompressed video signal state. The compressed video materials A and B are sent to the switcher 17.
[0021]
In the switcher 17, a picture (expanded data) to be sent to the data recording device 19 via the data compressor 18 and a picture (compressed data) to be sent directly to the data recording device 19 based on a control signal (not shown) from the CPU 20. Is selected.
[0022]
Therefore, in the CPU 20, the configuration of the image type relating to the compressed data in the compression sequence including the editing point (cutout point) Aout of the video material A and the compression sequence including the editing point (cut-in point) Bin of the video material B are included. Each analysis of the composition of the image type with respect to the compressed data is performed. Then, editing control information such as an image data switching point at the switcher 17, a compression start point and a compression end point at the data compressor 18, starting from the decompression start point and decompression end point at the decompressors 13 and 16, is generated. Then, the data is sent to the decompressors 13 and 16, the switcher 17, the data compressor 18, and the like, and the corresponding editing processing operation is executed. The contents will be described below.
[0023]
First, the data reproducing apparatus 11 reproduces a predetermined number of frames of compressed data including the cut-out point Aout, and the reproduced compressed data is transferred to the decompressor 13 and decompressed data A ′ corresponding to a predetermined number of frames necessary for editing processing. Is sent to the switcher 17. Similarly, at a timing so that the connection between the video material A and the video material B is successful, compressed data for a predetermined number of frames including the cut-in point Bin is reproduced from the other data reproduction device 14 and this is expanded. 16 and the decompressed image data necessary for the editing process is transferred to the switcher 17.
[0024]
The decompressed data A ′ and B ′ selected by the switcher 17 are transferred to the data compressor 18 and temporarily stored. The data compressor 18 performs recompression processing according to the image type using the decompressed data A ′ and A ′ and B ′ from the recompression start point determined by the CPU 20 to the recompression end point. The compressed image data is finally recorded by the data recording device 19.
[0025]
In this example, the picture arrangement for recompression follows the picture arrangement for video material A. Accordingly, the picture arrangement of the video material A is made up to the frame using the recompressed compressed data, and thereafter the picture arrangement of the video material B is the same. As a result, it may happen that 1 GOP is not composed of a complete number of frames in the editing section.
[0026]
Finally, the compressed data outside the editing area is not subjected to the editing process and is directly passed through the switcher 17 and transferred to the data recording device 19 for recording.
[0027]
Next, specific editing processing will be described with reference to FIG. FIG. 3 shows details in the editing section, and FIG. 4 is an example of a flowchart for realizing editing processing. As shown in FIG. 2a, video materials A and B are composed of 15 frames per GOP, and the GOP picture array, that is, the image type array is “BBIBBPBBPBBPBBP”. It is assumed that the compressed data has been subjected to.
[0028]
FIG. 2b shows a compressed video material A by inter-frame prediction in the data reproducing apparatus. The 15 frames having the regularity of “BBIBBPBBPBBPBBP” are defined as one unit (1 GOP), and each GOP is represented by GA0 to GA4. To do. Further, numbers A000, A001,..., A014, A100 to A414 are assigned to each frame.
[0029]
Similarly, FIG. 5C shows a compressed video material B reproduced from the data reproducing apparatus 14 by inter-frame prediction, and 15 frames having the same regularity are grouped into one, and each GOP is GB0 to GB4. . Further, numbers B000, B001,..., B014, B100 to B414 are assigned to each frame.
[0030]
Here, the arrangement of compressed and encoded data is a picture arrangement such as “IBBPBBPBBPBBPBB” unlike MPEG in the case of MPEG. However, hereinafter, for convenience of explanation, it is assumed that pictures are arranged and recorded as shown in FIG.
[0031]
For the video material A, an appropriate frame of the group GA3 is selected as the cut-out point Aout, and in the other video material B, the group GB2 is selected as the cut-in point Bin. Accordingly, the frame configurations (picture configurations) of these video materials A and B are d and f in the figure, and the edited picture configuration is as in e in the figure.
[0032]
FIG. 3 shows details in the editing section. As will be described later, the number of frames used for editing will be described later. Data (baseband components) obtained by decompressing the video materials A and B are A ′ and B ′. .
[0033]
Now, an example of the editing process according to the present invention will be described with reference to FIG. 4. In accordance with an instruction from the CPU 20, compressed data is sequentially displayed from the data reproducing apparatus 11 from a frame a little before the cutout point Aout of the video material A. It is transferred to the type identification unit 12. The image type identification unit 12 identifies the image type (I, B, P picture) of the compressed data near the frame including the cut-out point Aout of the video material A (step 31), and sends the information to the CPU 20. Thereby, the CPU 20 can determine that the configuration of the image type in the vicinity of the cutout point Aout is the configuration (sequence) as shown in FIG.
[0034]
Then, the first I picture A302 is determined as the expansion start point by searching forward from the cutout point Aout (forward search). Then, this time, a picture (I picture or P picture) used to decompress a B picture (possibly a P picture in some cases) existing in the meantime is searched backwards (backward search). Is the extension end point (step 32).
[0035]
In the example of FIG. 3, if the I picture A302 and the first P picture A305 following this are used, the two B pictures existing between them can be correctly decompressed. In this example, the P picture of A305 is the decompression end point.
[0036]
When pictures to be used for editing are determined, decompression processing is performed on them (step 33). As is well known, in this extension process, the I picture A302 is extended into an extended I picture A′302 (see FIG. 3). An expanded P picture A ′ 305 is generated from the expanded I picture A ′ 302 and P picture A 305. Subsequently, from the expanded I picture A′302 and P picture A′305, B pictures A303 and A304 are expanded using these image data (A303, A304) to obtain A′303 (A′304 is (Not shown).
[0037]
Of the decompressed data A ′, the data up to the cut-out point Aout is supplied to the data compression unit 18 and temporarily stored (step 34). Here, the reason why the decompressed data after the cut-out point Aout is not used as the image data for editing is that it is necessary to switch to the information of the video material B after the cut-out point Aout.
[0038]
Similar processing is performed on the video material B (steps 35 to 38). Compressed data of at least 1 GOP counted from the data playback device 14 from the frame slightly before the cut-in point Bin of the video material B is sequentially transferred to the image type identification unit 15.
[0039]
In the image type identification unit 15, the image type (I, B, P picture) of the compressed data near the frame including the cut-in point Bin of the video material B is identified (step 35), and the information is sent to the CPU 20. Thereby, the CPU 20 can determine that the configuration of the image type in the vicinity of the cut-in point Bin is a configuration (sequence) as shown in FIG.
[0040]
Then, a forward search is performed from the cut-in point Bin and the first I picture B202 is determined as an expansion start point, and a picture (I picture or P picture) used to search backward and expand a plurality of pictures existing therebetween. ) And the picture is set as the end point of expansion (step 36).
[0041]
That is, in order to succeed the video material B after the cut-in point Bin, it is necessary to obtain a minimum picture section having a good cut in the picture arrangement of the video material B. Referring to FIG. 3c, in order to extend a picture near the cut-in point Bin, it is necessary to extend to the first I picture B302 after the cut-in point Bin.
[0042]
In the example of FIG. 3c, it is necessary to use picture B305, which is three frames after I picture B302 (1 GOP + 3 frames). This is because the cut-in point Bin to the I picture B302 follow the picture arrangement of the video material A, as is apparent from FIG. For this reason, in the example of FIG. 3c, decompression processing is performed from I picture B 202 to P picture B 305 (step 37).
[0043]
Of the decompressed data B ′, the necessary number of frames from the cut-in point Bin is transferred to the data compression unit 18 up to B picture B303 in this example and stored (step 38).
[0044]
The data compression unit 18 re-compresses the image data based on the decompressed data A ′ and B ′ (step 39). An example of recompression processing when the picture arrangement of the editing section is set to be the picture arrangement of the video material A will be described next.
(1) Recompression is performed using the expanded I picture A′302 (A302 ′).
(2) P picture B209 'is calculated from the expanded B picture B'209 and I picture A302'.
(3) A new B picture A303 ′ is calculated using the I picture A302 ′, the P picture B209 ′, and the B picture A′303.
(4) A new B picture B 208 ′ is calculated using the I picture A 302 ′, P picture B 209 ′, and B picture B ′ 208. With the above processing, a 4-frame picture (I, B, B, P picture) including edit points is obtained.
(5) Similar processing is performed for subsequent pictures. However, the P picture B 212 ′ after three frames is generated based on the immediately preceding P picture B 209 ′. That is, P picture is used instead of I picture. Similar processing is performed on the subsequent P pictures, and recompressed pictures are sequentially generated.
(6) When this recompression process is sequentially performed, a B picture B301 ′ positioned immediately before the I picture B302 is generated using a P picture B303 ′ after two frames, and the like. However, since the picture material B itself is used after the B picture B301 ′, the B pictures B′302 and B′303 are not used as recompressed data. The recompression process described above does not change regardless of the frame where the edit points Aout and Bin are in one GOP. However, the number of each extended frame varies slightly depending on the frame position.
[0045]
The recompressed image data (11 frames from A302 ′ to B301 ′) is recorded as compressed data after editing by the data recording unit 19, and the original video materials A and B are recorded as they are before and after that. By this editing process, a new video material C in which the video materials A and B are joined can be obtained.
[0046]
As described above, the video material A is subjected to the decompression process for only four frames (A302 to A305) in the example of FIG. Regarding the video material B, since several frames are used as frames to be recompressed, only 1 GOP + 3 frames (B202 to B305) are decompressed, and the video materials A and B can be joined together.
[0047]
In particular, the video material B on the cut-in side uses the original compressed data itself as the edited image data from the first I picture obtained by backward searching from the cut-in point Bin. Can be kept in the minimum necessary frame section. In some cases, the video material B may be decompressed only within one GOP.
[0048]
As a result, the editing time can be greatly shortened, and the image quality deterioration caused by the decompression / recompression process can be greatly improved.
[0049]
When the image sequence is set so that the picture arrangement of the editing section is not the picture material A but the picture material B, the recompression processing is performed so that the picture arrangement of the picture material B is made after the cut-in point. Is done.
[0050]
In the editing apparatus 10 shown in FIG. 1, the data reproduction apparatus 11 in which the video material A is recorded, the data reproduction apparatus 14 in which the video material B is recorded, and the data recording apparatus 19 in which the edited data is recorded. Although it is composed of three data recording / reproducing apparatuses, if a recording / reproducing apparatus capable of random access, such as a hard disk drive, is used, they can be integrated into one. In this case, it is only necessary to manage the logical data connection, and the editing operation can be performed in a short time only by processing the editing points.
[0051]
【The invention's effect】
As described above, according to the present invention, only image data near the editing point is expanded while determining its image type, and editing is performed only within the sequence of the minimum image type. The compressed data is used as it is.
[0052]
As a result, the recompression processing section is shortened, so that the editing processing time is shortened. Further, since the data decompression processing and recompression processing are performed only for the minimum necessary area, the compressed video material is used as it is for the material away from the editing point. As a result, it is possible to minimize degradation of image quality due to the decompression and compression processing.
[0053]
Therefore, according to the present invention, it is possible to realize an editing process for a video signal compressed by inter-frame prediction in a short time and with minimal deterioration in image quality.
[Brief description of the drawings]
FIG. 1 is a system diagram of a main part showing an embodiment of an editing apparatus according to the present invention.
FIG. 2 is an explanatory diagram of an editing operation.
FIG. 3 is an explanatory diagram of an editing operation.
FIG. 4 is a flowchart illustrating an example of editing processing.
FIG. 5 is an explanatory diagram of a conventional editing process.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Editing apparatus, 11, 14 ... Data reproduction apparatus, 12, 15 ... Image type identification part, 13, 16 ... Expander, 17 ... Switcher, 18 ... Data compressor 19 Data recording device

Claims (5)

First image type identifying means for identifying the image type of the compressed video data of the first video material;
Second image type identifying means for identifying the image type of the compressed video data of the second video material;
The first I picture identified by the first image type identifying means by temporally searching forward from the cutout point set in the first video material is used as a decoding start point, and temporally from the cutout point. First decoding means for decoding the compressed video data of the first video material, using the first I picture or P picture identified backward by the first image type identification means as a decoding end point;
The first I picture identified by the second image type identifying means by temporally searching forward from the cut-in point set for the second video material is used as a decoding start point, and temporally from the cut-in point. Second decoding means for decoding the compressed video data of the second video material with the first I picture or P picture identified by the second image type identification means as the end point of decoding,
A connection for connecting the first video data decoded by the first decoding means and the second video data decoded by the second decoding means, using the cut-out point and the cut-in point as connection points. Means,
An editing apparatus comprising: re-encoding means for re-encoding the video data connected by the connecting means to generate re-compressed video data . Editing means for generating edited edited compressed video data by switching and outputting the compressed video data of the first video material, the compressed video data of the second video material, and the recompressed video data. The editing apparatus according to claim 1, further comprising: an editing apparatus. It further has recording means for recording the edited compressed video data generated by the re-encoding means on a recording medium.
The editing apparatus according to claim 1 . The re-encoding means re-encodes the video data connected by the connecting means in correspondence with the picture arrangement of the compressed video data of the first video material,
The second decoding means performs a backward search in time from the cut-in point, and re-encodes a video range before the first I picture identified by the second image type identifying means. To move the decoding end point so that it can be re-encoded corresponding to the picture arrangement of the compressed video data of the first video material.
The editing apparatus according to claim 1 . A first image type identifying step for identifying the image type of the compressed video data of the first video material;
A second image type identifying step for identifying the image type of the compressed video data of the second video material;
The first I picture identified in the first image type identification step by temporally searching forward from the cutout point set in the first video material is used as a decoding start point, and temporally from the cutout point. A first decoding step of decoding the compressed video data of the first video material using the first I picture or P picture identified backward in the first image type identification step as a decoding end point;
The second image is searched forward in time from the cut-in point set in the second video material. The first I picture identified in the image type identification process is used as a decoding start point, and the first I picture or P picture identified in the second image type identification process is searched backward in time from the cut-in point. As a decoding end point, a second decoding step of decoding the compressed video data of the second video material,
A connection for connecting the first video data decoded by the first decoding step and the second video data decoded by the second decoding step, using the cut-out point and the cut-in point as connection points. Process,
A re-encoding step for re-encoding the video data connected in the connection step to generate re-compressed video data.
An editing method characterized by that.

Images