JP2006129133A - Content reproducing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently reproduce content consisting of the same program separately video-recorded, using meta data created by a third person. <P>SOLUTION: A synchronization tag is previously added to the meta data. The tag indicates the difference in feature between a frame to be indicated by the tag and screens before and after the frame, a synchronization detecting means 7 detects the synchronization tag from the meta data read from a meta data accumulating means 2, and it is detected that to which frame the corresponding tag is equivalent on the basis of the difference in feature in the content read from a content accumulating means 6. The difference is obtained between the detected frame and a time described in the meta data, a meta data correcting means 8 corrects the meta data, a reproduction control means 3 controls a content reproduction means 5 on the basis of the corrected meta data to decide which part should be reproduced, and the content is reproduced. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a content reproduction apparatus that efficiently reproduces content obtained by recording a general broadcast program using metadata.

2. Description of the Related Art Conventionally, a content reproduction apparatus using metadata is known as described in Patent Document 1. According to this, the video playback device receives and stores the video material recording means (content storage device) for recording the video material and the associated information data (metadata) defining the information content of the video material. A storage device and playback control means for controlling playback of the video material recorded in the video material recording means based on the metadata are provided. The metadata includes a unique identification index, UMID, and start / end time codes. Based on this information, video that does not move much is fast-forwarded to reduce the time required to view the entire content. It is.
JP 2002-281457 (first page, fifth page)

  However, in the configuration as described above, the metadata needs to be specialized for video content. For example, a third party (another device) for a program (content) recorded in each home. Considering that content reproduction is performed using metadata for the same content created by the above, the following inconvenience occurs. That is, 1) The time at which content recording is started differs depending on the device that performed the recording, and the time from the recording start time is different even for the same video. 2) Each recording device has a different clock, and even if recording starts simultaneously, the recording start time added to the recorded content differs. In addition, depending on the bit rate at the time of compression and, more specifically, the state of the radio wave transmitted to each home, the received image itself is not always the same due to noise or ghost.

  Therefore, if playback is performed based on the time described in the metadata, it cannot be correctly played back from the desired scene.

  SUMMARY OF THE INVENTION In view of the above problems, the present invention provides metadata and a content playback apparatus that can perform content playback as desired even when metadata for the same content created by a third party is used.

  In order to solve this problem, a content playback apparatus according to the present invention includes metadata storage means for storing video content information and storing metadata including synchronization information based on scene changes in the video content used for extraction. A content storage unit that stores video content, a synchronization unit that synchronizes the metadata and the video content stored in the content storage unit, and a correction unit that corrects the metadata based on the output of the synchronization unit. It is a thing.

  Furthermore, the content reproduction apparatus of the present invention is configured such that the synchronization information includes information regarding a feature difference indicating how the synchronization information is changed with respect to the screen immediately before the location where the synchronization information is added. is there.

  As described above, according to the present invention, it is possible to reproduce video content uniquely recorded using metadata created by a third party, and conversely, using metadata created from uniquely recorded video content, There is an effect that the video content recorded by the three parties can be reproduced.

  According to a first aspect of the present invention, there is provided a metadata storage means for storing video content information, and storing metadata including synchronization information based on a scene change in the video content used for the extraction, and the video content Content storage means for storing data, synchronization means for synchronizing the metadata and video content stored in the content storage means, and correction means for correcting the metadata based on the output of the synchronization means As a result, if it is possible to determine which frame is the screen indicated by the metadata at one location, the number of frames from one screen to another is the image quality, bit of recorded content. Because it is the same regardless of the rate, it has its own using metadata created by a third party An effect that can be obtained metadata of the video content and synchronization is established.

  The invention according to claim 2 is the content playback apparatus according to claim 1, wherein the synchronization information changes with respect to the screen immediately before the location where the synchronization information is added. It is characterized in that it includes information on the characteristic difference to be shown, and this makes it possible to more efficiently use the synchronization information by checking whether the comparison result shown in advance by the metadata matches the actual comparison result. It has the effect that the indicated location can be narrowed down.

  Further, the invention according to claim 3 is the content playback apparatus according to claim 1, wherein the video content used for the extraction and the video content stored in the content storage means are a stream compressed by the MPEG system. As a result, the feature of each frame in the video content stored in the content storage means appears more prominently with a smaller amount of data.

  According to a fourth aspect of the present invention, there is provided the content reproduction apparatus according to the third aspect, wherein the synchronization information is applied to at least one of the location where the synchronization information is added and the immediately preceding and immediately following I picture. It is characterized by including information on the feature difference indicating how it has changed, and by this, it is more possible to see whether the comparison result shown in advance by the metadata matches the actual comparison result It has the effect that the location indicated by the synchronization information can be narrowed down efficiently.

  Further, the invention according to claim 5 is the content playback apparatus according to claim 3, wherein the synchronization information is information attached to a location where the feature difference from the immediately preceding I picture is particularly large. As a result, when detecting synchronization, the features of adjacent I pictures are compared, and the point where the difference is the largest is the first candidate, so that synchronization can be achieved more efficiently. Has an effect.

  The content reproduction apparatus according to claim 6, wherein the feature difference includes at least one of the size, luminance, hue, and color density of the I picture. As a result, when an incorrect I picture is selected, the probability that the comparison result indicated in advance by the metadata matches the actual comparison result is low, and is indicated by the synchronization information. This has the effect that the I picture can be narrowed down more efficiently.

  Further, according to a seventh aspect of the present invention, in the content reproduction device according to the third aspect, the synchronization information is located at a place where there is no scene change for at least several seconds before and after the scene change indicated by the synchronization information. As a result, the difference from the immediately preceding I picture increases only at the location indicated by the synchronization information, and which corresponds to the I picture indicated by the synchronization information. This has the effect of making it easier to obtain.

  The invention according to claim 8 is the content playback apparatus according to claim 3, characterized in that a plurality of the synchronization information exist in the metadata. It has the effect of preventing.

  The content reproduction apparatus according to claim 9 is the content reproduction apparatus according to claim 8, wherein the synchronization unit calculates the number of frames between locations to which synchronization information is added, and the video content is in accordance with the calculated value. It is characterized in that synchronization detection is performed by utilizing the presence of the B picture or P picture having the maximum size among the I pictures or the I pictures existing between the pictures at As a result, the detection process can be performed only by examining the data amount of the I picture, B picture, and P picture, and synchronization can be achieved without decoding the MPEG compressed content. .

  Further, the invention according to claim 10 is the content playback apparatus according to claim 1, wherein the storage means for storing the playback conditions for playing back the video content stored in the content storage means, the output of the correction means, And a control unit that controls playback of the video content based on the playback condition, and a playback unit that plays back the video content based on the output of the content storage unit and the output of the control unit. With this, the content can be freely reproduced under desired conditions using the corrected metadata.

  The invention according to claim 11 is the content playback apparatus according to claim 1, further comprising means for storing the metadata corrected by the correcting means or replacing the metadata before correction. As a result, once the metadata and the accumulated video content are synchronized, there is no need to re-synchronize the next time the video content is reproduced using the metadata.

  The invention according to claim 12 is the content playback apparatus according to claim 11, wherein whether the metadata is already synchronized with the video content stored in the first storage means is identified. Therefore, when reproducing video content using metadata, it is possible to instantaneously determine whether or not synchronization is required in advance.

  The invention described in claim 13 has information indicating in what manner the video content stored in the content storage means is compressed in the content playback apparatus described in claim 1. Thus, the video content created by a third party can be synchronized with the metadata.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

(Embodiment)
FIG. 1 is a block diagram showing a configuration of a content reproduction apparatus according to an embodiment of the present invention. Referring to this figure, video content is recorded in the content storage means 6, and here, video content that is generally MPEG2 compressed and recorded by a recording device is stored. Video contents are filed, and by reading the file attributes, the recording date and time and the outline of the contents can be known.

  The metadata storage unit 2 stores metadata corresponding to the video content stored in the content storage unit 6. Here, metadata created by a third party (separate device) individually recording the above-described television broadcast and created based on the recorded video is stored. The metadata includes the date and time of recording, the contents thereof, where and what scene is recorded in the corresponding video content, and the same program recorded by other video recording devices, for example. Information for synchronization to be used when adapting to is shown in a predetermined format.

  FIG. 2 shows a specific example of metadata applicable to the present invention. Here, the metadata is described in XML notation (eXtensible Markup Language), and is composed of a header portion indicating the general content of the metadata and a body portion indicating the specific content of the video content.

  The header part is indicated by a header tag (<header>), and the title of the video content program (<title>), program genre (<category>), recording date and time (date), recording time (duration), A keyword indicating the content (kywd) and encoding information (<enc>) indicating how the video content used when the metadata is created are compressed are set. The encoding information includes a method parameter indicating the encoding method and a mode parameter indicating the mode. There are three method parameters, mpeg2, mpeg4, and none, and mpeg2 and mpeg4 literally indicate that the video content was compressed by the MPEG2 or MPEG4 system. “None” indicates that the video content is not MPEG-compressed. The mode parameter is effective when the value of the method parameter is other than “none”. Here, “constant” means a mode in which an I picture is inserted at a constant period regardless of a scene change, and an I picture for each scene change. Are forcibly assigned, and two types of optimum, which means a mode in which an I picture is inserted at a constant period until the next scene change, are prepared. In the present embodiment, it is shown to be constant. The offset tag (<offset>) is a metadata correction tag, which will be described later.

  Here, the genre of the program is “sports / baseball”, which indicates that the program is a sports program and that the sport is baseball. Here, “/” is used to further categorize the genre. Further, the encoding information is “method = mpeg2; mode = constant”, which is encoded in MPEG2 and is in a constant mode in which an I picture is not forcibly assigned at the time of a scene change. In addition, although all are described using Roman characters here, it is a matter of course that katakana and kanji may be used.

  The start of the body part is indicated by the body tag (<body>). The scene tag is recorded as a segment tag (<seg>), where each scene is regarded as a segment. ing. The segment tag includes a position tag (<pos>) indicating which recording period the block is in the entire video content, a keyword tag (<kywd>) indicating the content, and the like. In the position tag, the from parameter indicating from which time the segment starts on the video content, the duration parameter indicating the length of the segment, and the unit for specifying the unit of the numerical value given to them Has parameters.

  Here, in segment 1, as keywords, N.I. Y. , Bitter # 1, etc., the scene of New York, the first batter, the from parameter is 12000, the duration parameter is 5500, and the unit is ms, so 12000ms from the start of the video content, that is, 5500ms after 12 seconds, That is, it is shown that 5.5 seconds corresponds to the scene. Segment 4 is displayed as “sync = on” in the segment tag, which indicates that this segment should be used to synchronize with the video content. To facilitate synchronization, here, As shown in the header section, this metadata is created based on video content compressed in MPEG2, so that information indicating the feature difference between the start frame of the segment and the immediately preceding and immediately following I picture is displayed. Yes. Here, the feature difference from the immediately preceding and immediately following I picture is shown, but usually there is not much difference in the video immediately after the scene change, so only the feature difference from the immediately preceding I picture is shown. It may be.

  The operation will be described below with reference to FIG. The control means 10 selects video contents and corresponding metadata from the content storage means 6 and the metadata storage means 2 (step 31). The content storage unit 6 and the metadata storage unit 2 start reading the video content and metadata (step 32), and the selected metadata is displayed (step 33). Since the display device is not particularly new, it is not shown.

  Based on the displayed metadata, the user sets playback conditions via the control means 10 (step 34). The setting method can be freely set according to the preference of the setter. For example, a specific scene can be repeatedly reproduced several times, or a specific scene can be skipped.

  On the other hand, the synchronization tag is searched from the metadata read in step 32 (step 35). Here, as shown in FIG. 2, segment 4 and segment 9, the synchronization tag is displayed as “sync = on” in the segment tag, and this segment is used for synchronization to synchronize with the video content. It means you can take it. The parameters displayed after “sync = on” are not used here.

  When the synchronization tag is found, the video content is analyzed in order to find out where the corresponding segment actually exists (step 36). When the analysis is completed and synchronization is established, “synchronization complete” is selected in step 37, and the process proceeds to step 38. How the content analysis is performed will be described later.

  If the synchronization is OK in step 37, the process proceeds to step 38 to correct the metadata. That is, the corresponding actual value obtained from the start time of each scene shown in the metadata obtained by a series of operations in steps 36, 37, and 41, the attribute of the video content stored in the content storage means 6, and the like. Correct the difference from the scene start time. Here, in the header portion of the metadata shown in FIG. 2, the offset tag is used to correct the difference between the start time of each scene indicated by the metadata and the actual start time.

  In other words, the offset tag includes a numerical value indicating the difference between the start time of the segment indicated in the metadata, the time difference in the actual video content to be reproduced, and the unit of the numerical value by the time parameter and the unit parameter. In addition, information indicating which video playback device has corrected the offset data is clearly indicated by a base parameter. Here, the serial number of the playback device is written in the base parameter. Conversely, if the value of the base parameter matches the serial number of the playback device, the metadata is treated as corrected.

  The offset tag value is used as follows. That is, assuming that the reproduction of the segment 1 is designated in the reproduction condition data storage unit 4, the reproduction control unit 3 reads the offset tag in the metadata header portion from the metadata correction unit 8. Since time = 1-1233 and unit = ms, −1233 ms is added to the from parameter of each segment.

  Next, the value of the position tag in the segment tag <seg = 1> is read. Here, since the from parameter is 12000 and the unit parameter is ms, the content reproduction means 5 is configured to reproduce the video image from 10767 ms after adding 1233 ms to the previous 20001 offset tag from the start of the video content. Give a command to it.

  Even if the recording start time is different for each video recorded at each home, the same video is always sent to each home when viewed in frame units, and once synchronized, all video Full synchronization is obtained in the part. Therefore, as described above, if synchronization correction is performed at one location, it is possible to ensure synchronization at all locations.

  When the correction step (step 38) is completed, normal reproduction is started (step 39). Here, the reproduction control means 3 selects the segment tag of the corresponding scene from the metadata accumulation means 2 based on the information stored in the reproduction condition data accumulation means 4 such as which scene is reproduced, and the metadata. Based on the tag information corrected by the correcting unit 8, the content reproducing unit 5 is instructed which part of the video content is to be reproduced.

  In the same manner, video contents are reproduced in the order designated by the reproduction condition data storage means 4.

  If it is not possible to analyze where the segment with the synchronization tag exists in step 36, “No” is selected in step 37, the process proceeds to step 41, indicating that synchronization has not been achieved, and video The content is normally reproduced from the beginning (step 42).

  In the above embodiment, ms (millisecond) is used as the unit parameter indicating the unit. However, the present invention is not limited to this, and the unit may be seconds, minutes, or a frame or field. In particular, since video images are often managed in units of frames, using a frame as a unit is effective because it can suppress variations in source oscillators that generate clock signals in individual content playback apparatuses.

  With this configuration, the user can select the video content and the corresponding metadata in the content playback apparatus 1, and simply enter the scene that he / she wants to view, and the metadata created by the third party can be easily obtained. The video content can be reproduced using the.

  Next, a content analysis method will be described. In order to make the explanation easy to understand, before describing a specific method, image compression generally performed in a video recorder using MPEG compression will be described with reference to FIG. FIG. 3 (1) shows a video image before compression, that is, a television image program transmitted from a broadcasting station. Each break represents a frame. (2) is the result of MPEG compression of the signal. I represents an I picture, P represents a P picture, and B represents a B picture.

  Usually, when video images are compressed / recorded, an I picture uncorrelated with other frames, a P picture indicating a difference from the I picture, an I picture before or after the I picture, or a B picture indicating a difference from the P picture Thus, the compression sequence is simplified and the convenience during reproduction is achieved. In many cases, as shown in the part before time A in FIG. 3B, a method of periodically repeating “IBBPBBPBBPBBPBB” is employed. However, for example, in FIG. 3 (1), when there is a scene change at time A, if the difference from the previous video data is taken, the amount of data becomes huge. It is composed of pictures, and the cycle of IBBPBB... Described above is repeated starting from this. This is equivalent to encoding information in the metadata header part, mode parameter = optimum. However, depending on the model, the calculation amount required for encoding is given priority over the increase in the amount of data. Therefore, as shown in FIG. 4 (2), the cycle of IBBPBB... Some have. This is equivalent to encoding information / mode parameter = constant in the metadata header section.

  FIG. 3 (3) shows an MPEG compression result when the video content shown in FIG. 3 (1) is recorded by another video recorder. Since the recording start time is different, it is up to the video recorder whether the transmitted video is assigned to I, P, or B for a period of time from the start of recording. In this figure, since the encoder keeps the above-described cycle of IBBPBB... Regardless of whether or not there is a scene change, even after a scene change at time A occurs, FIGS. ) That I, P, and B cannot be synchronized.

  FIG. 4 (3) shows a case where the encoding method in the video recorder is different. In this figure, the encoder allocates an I picture at every scene change, and this point is used as a starting point for the IBBPBB. The cycle is repeated.

  The type of the video content stored in the content storage means 6 can be determined as a matter of course as long as this content is recorded by itself, as shown in FIG. 3 (3) or FIG. 4 (3). However, there is a case where it is not clear if it is recorded by a third party. Therefore, here, in the header portion of each video content stored in the content storage means 6, information indicating which type is shown in FIG. 3 (3) or FIG. , A constant mode is added.

  On the other hand, a synchronization tag in metadata is added mainly when a scene change occurs. In particular, it is effective to select and attach a scene change that has a particularly large amount of picture data.

  Returning to the content analysis method, FIG. 6 is a sequence diagram showing a specific example of content analysis in step 36. When the analysis is started, the synchronization detection means 7 extracts segments with synchronization tags and calculates the interval between the segments (step 51). Since each segment has a position tag indicating its start time, the value can be easily obtained by reading the from parameter displayed on the segment and obtaining the difference. Here, the interval between these segments is obtained in units of frames, and the interval between the segment to which the Nth synchronization tag is added and the segment to which the (N + 1) th synchronization tag is added is the FN frame.

  When the interval between the segments to which the synchronization tag is added is obtained, the next candidate selection for the first segment to which the synchronization tag is added is entered.

  If the video content to be synchronized now uses the method shown in FIG. 4 (3) at the time of MPEG compression, that is, the optimum mode, an I picture is assigned for each scene change, and the synchronization tag is set to the scene change. In step 52, an I picture may be selected as a candidate frame. Actually, an approximate I picture may be calculated from the recording start time indicated by the attribute of the video content or the time of the synchronization tag indicated by the segment tag, and an adjacent I picture may be selected.

  On the contrary, if the video content to be synchronized now uses the method as shown in FIG. 3 (3) at the time of MPEG compression, that is, the constant mode, the I picture, Since P pictures and B pictures are allocated, in step 52, approximate positions are calculated as candidate frames from the recording start time indicated in the video content attribute and the time of the synchronization tag indicated in the segment tag. A P picture, a B picture, or an I picture that is in the vicinity and has a maximum size may be selected. “Maximum” as used herein means that the size is the largest among the I pictures existing between the pictures. The same applies to the following unless otherwise specified.

  In step 53, if the video content to be synchronized is using the optimum mode, an I picture is assigned for each scene change. Therefore, the I picture selected as a candidate is indicated by the synchronization tag. If it is a scene, all the pictures selected in step 53 should be I pictures. In this case, it is determined that synchronization has been achieved, and the process proceeds to step 55, where the difference between the time in the actual video content of the scene selected as a candidate and the data indicated by the metadata is extracted, and the analysis is terminated.

  If the video content to be synchronized now uses the constant mode at the time of MPEG compression, as described above, the synchronization tag is always added at the time of the scene change, and at the time of the scene change, the picture size Is always the maximum or happens to be assigned an I picture, so if the picture selected as a candidate is a corresponding scene, the picture selected in step 53 must have a maximum size. Either a P picture or a B picture, or an I picture. In this case, it is determined that synchronization has been achieved, and the process proceeds to step 55, where the difference between the time in the actual video content of the scene selected as a candidate and the data indicated by the metadata is extracted, and the analysis is terminated.

  Conversely, if one or more of the pictures selected in step 53 is not a picture that satisfies the condition, it is determined that synchronization is not achieved, and the next candidate must be extracted. In this case, the process proceeds to step 57, and it is determined whether there is still an I picture that can be a candidate, or a P picture or B picture having a maximum size. Actually, either an I picture before or after the previously selected picture, or a P picture or B picture having a maximum size is selected.

  If the video content to be synchronized now uses the optimum mode during MPEG compression, and the 20th I picture is the first candidate, the next candidate is the 21st, It is preferable to select an I picture that gradually moves away from the first candidate as a candidate, such as 19th, 22nd, 18th,.

  If the video content to be synchronized now uses the constant mode at the time of MPEG compression, the next candidate is the P picture, B picture, or I picture, which is the maximum before that, and behind the first candidate. In this case, the candidate may be selected so as to gradually move away from the first candidate, such as a P picture, a B picture, or an I picture, which is a maximum, and before or after that.

  As for the presence / absence of the next candidate, “no next candidate” is set when the distance from the I picture selected as the first candidate is more than a certain distance. The user can select how much “a certain level or more” is set.

  If “no next candidate” is found in step 57, the process proceeds to step 59, where the “detection impossible flag” indicating that synchronization cannot be established is set and the sequence is terminated.

  In the above embodiment, the metadata is described as being created based on the MPEG compressed video content. However, if the metadata is not a video content compressed by MPEG2 or MPEG4, it is a digital video movie. If it is created based on video content recorded in the MiniDV format, etc., the metadata is displayed so that information indicating the feature difference between the start frame of the segment and the immediately preceding and immediately following frames is displayed. create. Here, the feature difference from the immediately subsequent frame is merely an example, and for example, the feature difference from the frame after 15 frames, which is the period in which the I picture in FIGS. 3 and 4 is inserted, may be used. In this way, in the content reproduction apparatus 1, if the video content is MPEG compression and the optimum mode is used, the degree of coincidence with the feature difference indicated in the metadata can be further increased. it can.

  FIG. 7 shows another embodiment of the content reproduction apparatus according to the present invention. In this figure, steps having the same functions as those in FIGS. 5 and 6 are denoted by the same reference numerals, and detailed description thereof is omitted. In this embodiment, the same metadata is used, but the content analysis step is different from FIG.

  In step 35, a search for a synchronization tag is performed from the metadata read in step 32. When a synchronization tag is found, the video content is analyzed in order to find out where the corresponding segment actually exists. Performed (step 46). When the analysis is completed and synchronization is established, “synchronization complete” is selected in step 37, and the process proceeds to step 38. Thereafter, as in the case shown in FIG. 5, correction of metadata and playback of video content are started. Details of the content analysis will be described later.

  If it is not possible to analyze where the segment with the synchronization tag exists in step 46, "No" is selected in step 37, and the next synchronization tag is searched (step 43). If the synchronization tag is found, the process returns to step 46 and the content analysis is performed again. On the contrary, if the synchronization tag is not found, the process proceeds to step 41 to display that the synchronization is not established, and normal reproduction of the video content is performed from the beginning (step 42).

  Next, the content analysis method will be described with reference to FIG. As described above, since the segment to which the synchronization tag is added is added at the time of the scene change, if the video content to be synchronized now uses the optimum mode at the time of MPEG compression Since an I picture is assigned for each scene change, an I picture is selected as a candidate frame in step 61. Conversely, if the video content to be synchronized now uses the constant mode during MPEG compression, I pictures, P pictures, and B pictures will be allocated at a constant cycle regardless of the presence or absence of a scene change. In step 61, as a candidate frame, an approximate position is calculated from the recording start time indicated in the video content attribute and the time of the synchronization tag indicated in the segment tag. Select P picture, B picture, or I picture.

  Actually, a rough picture is calculated from the recording start time indicated by the attribute of the video content and the time of the synchronization tag indicated by the segment tag, and a picture that satisfies the above condition is selected as a candidate.

  In steps 62 and 63, if the video content to be synchronized now uses the optimum mode at the time of MPEG compression, the following processing is performed. That is, first, the ninth, tenth, and eleventh I pictures are extracted before the candidate I picture. Assuming that the candidate I picture is the 50th picture from the top of the video content, IP41, IP40, and IP39 are extracted. Since one picture can be decoded by itself as an I picture, these are decoded in step 63, and the feature difference is extracted. Here, the size of the I picture, the overall luminance, the change in hue, and the volume difference of the audio signal added to each I picture are compared. How these changes should be described is described after the synchronization tag, and if the segment 4 is examined now, the tag of the segment 4 has frame = ip brite = updc size = downup col. = Updint tint = updaudio audio = upnc, which are the type of picture (I, P, or B) to which the tag is added, the change in luminance, the change in picture size, and the darkness of the color, respectively. A change in depth, a change in hue, and a change in volume are shown, and a comparison with the immediately preceding and immediately following I pictures is displayed. If it is “down”, it means that it has increased from the immediately preceding I picture and the immediately following I picture has decreased from the current one. dc means “don't care” in don't care, and nc means “no change” in no change.

  In the video content used in creating the metadata, it is described that the luminance is increased, no question, the size is decreased, increased, the hue is increased, decreased, the volume is increased, and there is no change. It is checked whether or not IP40 and IP39 have changed in that way. If all items match the description, it is considered that synchronization has been achieved, the process proceeds to step 55, a time difference is extracted, and the content analysis process is terminated. If even one item does not match, go to step 67.

  On the other hand, if the video content to be synchronized now uses the constant mode at the time of MPEG compression, the following processing is performed in steps 62 and 63. That is, the video content to be analyzed is assigned an I picture, a P picture, and a B picture at a constant cycle regardless of the presence or absence of a scene change, so the ninth and tenth pictures before the candidate pictures. A maximum P picture or B picture and an I picture existing between them are extracted. Assuming that a candidate picture is a B picture and is a picture sandwiched between the 50th and 51st I pictures from the top of the video content, IP42, IP41, and B pictures are extracted. If the candidate picture is an I picture, I pictures before and after that are extracted.

  When a B picture is reproduced from the picture, it is necessary to perform decoding processing using the previous I picture and P picture. However, since the decoding from the immediately preceding I picture is sufficient, the time required for processing is human. From this sense, it can be ignored. Since one picture can be decoded by itself as an I picture, these are decoded in step 63, and the feature difference is extracted. Hereinafter, as in the optimum mode, the differences in size, brightness, hue, and volume are compared, and if all items match the contents described in the metadata, it is considered that synchronization has been achieved and the process proceeds to step 55. Proceed to extract the time difference and end the content analysis process. If even one item does not match, go to step 67.

  In step 67, it is confirmed that the range for performing synchronization detection is within a certain range. Here, the comparison process in step 63 is performed 20 times or more in the optimum mode and 40 times or more in the constant mode. Since the synchronization cannot be detected, the process proceeds to step 59. If the specified number of times is as follows, the next candidate picture is extracted in step 68. In the case of the optimum mode, the next existing I picture is extracted. In the case of the constant mode, if the previously extracted picture is an I picture, the B or P picture having the next maximum value is not selected. Extracts the I picture immediately after the picture.

  The content analysis as described above is performed, and synchronization detection is performed.

  The video content used to create the metadata and the video content to be played are not necessarily the same in the MPEG compression algorithm. As described above, depending on the radio wave condition, noise and ghost Depending on the characteristics of the screen, it is considered that there is almost no difference between the two. The metadata and the video content can be easily synchronized.

  In the present embodiment, in step 62, the analysis is started with the tenth I picture before the candidate picture as a start. However, the user can set which range is the analysis range. It goes without saying that it is good. In addition, the match is sequentially checked from the previous ten, but the candidate picture is centered away from the picture (such as +1, -1, +2, -2,...). May be.

  In this embodiment, the feature comparison is performed only for one picture to which a synchronization tag is added. However, after the time difference is corrected in step 65, a picture to which the next synchronization tag is added is obtained. The processing of steps 62 and 63 may be performed on this picture, and it may be confirmed in step 64 that a match is obtained. If no match is obtained, the next candidate picture is extracted within the range in which the determination in step 67 is Yes, and a loop process including steps 62, 63, 64, 67, and 68 is performed, and all items match. You may make it search whether there exists an I picture to do.

  In this embodiment, the feature difference between the previous and next I pictures of the candidate picture is performed. However, the difference from only the previous I picture is obtained and compared with the content described in the metadata. It may be good.

  In addition, when obtaining values such as the brightness and hue of a picture, for example, if it is luminance, the average brightness (luminance) of the entire screen is obtained, and in the case of hue, the average of only the center of the screen is obtained. If the hue is determined to be obtained, the feature of the picture becomes clearer, and it is possible to reduce misjudgment.

  Note that the number of synchronization tags and the interval thereof depend on the length of the video content, but there should be at least 4-5 locations at the beginning of the content at a rate of about 1 location every 3-5 minutes. Then, it is considered easy to synchronize and is appropriate.

  The content reproduction apparatus according to the present invention can reproduce video content uniquely recorded using metadata created by different devices, and conversely, by using metadata created from video content originally recorded, It is useful as a content playback device, video playback device, etc. that can efficiently play back video content recorded by using the metadata. .

1 is a block diagram showing a configuration of a content reproduction apparatus according to an embodiment of the present invention The figure which showed the specific example of the metadata in the same content reproduction apparatus Conceptual diagram showing the relationship between video images and MPEG-compressed streams in the content playback apparatus Conceptual diagram showing the relationship between video images and MPEG-compressed streams in the content playback apparatus Flow chart showing the flow of synchronization detection and content playback in the content playback apparatus A flowchart showing a flow of a synchronization detection method in the content reproduction apparatus A flowchart showing a flow of second synchronization detection and content reproduction in the content reproduction apparatus The flowchart which showed the flow of the 2nd synchronization detection method in the same content reproduction apparatus

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Content playback apparatus 2 Metadata storage means 3 Playback control means 4 Playback condition data storage means 5 Content playback means 6 Content storage means 7 Synchronization detection means

Claims (13)

Metadata storage means for extracting video content information and storing metadata including information for synchronization based on scene changes in the video content used for extraction;
Content storage means for storing video content;
Synchronization means for synchronizing the metadata and the video content stored in the content storage means;
A content reproducing apparatus comprising: a correcting unit that corrects the metadata based on the output of the synchronizing unit. 2. The content reproduction apparatus according to claim 1, wherein the synchronization information includes information regarding a feature difference indicating how the screen has changed with respect to a screen immediately before the portion to which the synchronization information is added. 2. The content playback apparatus according to claim 1, wherein the video content used for the extraction and the video content stored in the content storage means are a stream compressed by an MPEG system. The information for synchronization includes information on a feature difference indicating how the synchronization information is changed with respect to at least one of a location where the synchronization information is added and an immediately preceding or immediately following I picture. 3. The content reproduction apparatus according to 3. 4. The content reproduction apparatus according to claim 3, wherein the synchronization information is information attached to a portion having a particularly large feature difference from the immediately preceding I picture. 6. The content reproduction apparatus according to claim 4, wherein the feature difference includes at least one of a size, luminance, hue, and color density of an I picture. 4. The content reproducing apparatus according to claim 3, wherein the synchronization information is information attached to a portion where there is no scene change for at least several seconds before and after the scene change instructed by the synchronization information. 4. The content reproduction apparatus according to claim 3, wherein a plurality of the synchronization information exist in metadata. The synchronization means calculates the number of frames between locations to which synchronization information is added, and the size of the I picture or the I picture existing in a form sandwiching the picture at the location according to the calculated value in the video content. 9. The content reproduction apparatus according to claim 8, wherein synchronization detection is performed by utilizing the presence of the largest B picture or P picture. The content reproduction apparatus is
Accumulation means for accumulating reproduction conditions for reproducing the video content stored in the content accumulation means;
Control means for controlling reproduction of the video content based on the output of the correction means and the reproduction condition;
2. The content reproduction apparatus according to claim 1, further comprising: reproduction means for reproducing the video content based on the output of the content storage means and the output of the control means. 2. The content reproducing apparatus according to claim 1, further comprising means for storing the metadata corrected by the correcting means or replacing the metadata before correction. 12. The content playback apparatus according to claim 11, further comprising means for identifying whether the metadata is already synchronized with the video content stored in the content storage means. 2. The content reproducing apparatus according to claim 1, further comprising information indicating in what manner the video content stored in the content storage means is compressed.

Images