JP2009284035A - Decoding device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a decoding device enabling a receiver side to suppress the expression of a "ghost phenomenon" caused by a scene change. <P>SOLUTION: An accumulation quantity extracting section 5 extracts the accumulated code quantity of each image from a compressed video stream and delivers the extracted code quantity to an image information buffer control section 8. The image information buffer control section 8 writes the accumulated code quantity to an image information buffer 7. A decode processing section 6 performs decoding processing and detects an error in decoding, and delivers error information in decoding to a determination instruction section 9. When the decoding processing section 6 detects an error in decoding, the determination instruction section 9 determines whether the error in decoding has been caused by a scene change on the basis of the information of the accumulated code quantity in the image information buffer 7, and if determining that the error has been caused by the scene change, the determination instruction section 9 instructs the decoding processing section 6 to execute freezing or interconcealment. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a decoding device suitable for use in a portable terminal or the like that supports broadcasting using one segment (hereinafter referred to as “one-segment”) among terrestrial digital television broadcasting.

  With the promotion of the transition from terrestrial analog TV broadcasting to terrestrial digital TV broadcasting, 1Seg is attracting attention, and 1Seg-compatible mobile terminals that can receive this are popular. However, with 1Seg, it is difficult to improve image quality due to limitations such as bit rate. The bit rate is suppressed to about 312 kbps (kilo bits per second) or less, and the frame rate is 15 fps (frames per second). Transmission of information exceeding this amount is impossible in standard, and video creation in the direction of increasing transmission amount is impossible in standard.

  By the way, “scene change” is inevitable in the video. A scene change is a content change that occurs in a video stream, and often occurs in a series of video streams. When encoding without taking any measures against a scene change, the contents before and after the scene change are completely different, so an error occurred during decoding and the current picture was decoded within the previous picture. An image that looks like a “moving” picture appears. Some people feel that the edge of the object in the picture currently being decoded is visible in the previous picture. Here, this is defined as a “ghost phenomenon”.

  FIG. 5 is a diagram for explaining the “ghost phenomenon”. In FIG. 5, 1 is an image before a scene change (map image), 2 is an image at the time of a scene change, and 3 is an image after a scene change (an image of a human face). In this example, the image 2 at the time of the scene change is an image in which the image 3 after the scene change can be seen in the image 1 before the scene change. This is the “ghost phenomenon”.

  As a countermeasure for the “ghost phenomenon”, it is conceivable to perform some processing on the transmission side. For example, a process of detecting a scene change on the transmission side and converting a picture immediately after the scene change into a picture that does not refer to any picture, for example, an IDR (Instantaneous Decoder Refresh) picture in the case of one segment. Can be considered.

However, for example, in the case of one segment, a scene change does not necessarily appear within the standard range of inserting an IDR picture once every two seconds. For this reason, when the process of detecting a scene change on the transmission side and converting the picture immediately after the scene change into an IDR picture, the number of IDR pictures is surely increased and the code amount is within the standard range (maximum code amount) of one seg. The possibility of exceeding is very high. This is a fatal demerit in One Seg where the total amount of information is determined. Therefore, even in the current one-segment broadcasting, no scene change detection / measure is taken in order not to increase the total amount of information. Therefore, the “ghost phenomenon” may occur.
JP-A-6-54315 JP 2002-10254 A

  In view of this point, an object of the present invention is to provide a decoding apparatus capable of suppressing the expression of “ghost phenomenon” caused by a scene change on the receiving side.

  The decoding device disclosed herein decodes a compressed video stream, and when the decoding processing unit detects a decoding error, whether or not the decoding error is caused by a scene change. A determination instruction unit for determining and instructing the decoding processing unit on a concealment method corresponding to the determination result.

  In the disclosed decoding device, when the decoding processing unit detects a decoding error, the determination instruction unit determines whether the decoding error is caused by a scene change, and the determination is made. A concealment method corresponding to the result is instructed to the decoding processing unit. Therefore, the reception side can suppress the expression of the “ghost phenomenon” due to the scene change.

  FIG. 1 is a circuit diagram showing an embodiment of the present invention. One embodiment of the present invention is used for a one-segment mobile terminal, and includes a cumulative code amount extraction unit 5, a decoding processing unit 6, a picture information buffer 7 serving as a storage unit, and a picture serving as a storage management unit. An information buffer management unit 8 and a determination instruction unit 9 are provided.

  The accumulated code amount extraction unit 5 inputs a compressed video stream that is a compressed video data sequence obtained by demodulating a received signal of one seg by OFDM (Orthogonal Frequency Division Multiplexing), and the input compressed video The picture type and cumulative code amount of each picture are extracted from the information that the stream has. The extracted picture type and accumulated code amount are transferred to the picture information buffer management unit 8.

  The decoding processing unit 6 receives the input compressed video stream, and can perform decoding processing, vector movement tendency calculation, color difference distribution calculation, decoding error detection, concealment processing described later for decoding error, and the like. It was made possible. The vector movement tendency calculation and the color difference distribution calculation are performed only when set by the user, and the results are sent to the picture information buffer management unit 8.

  Further, the decoding processing unit 6 passes decoding error information indicating the presence or absence of decoding error to the picture information buffer management unit 8 for each picture. As a result of decoding, the decoding processing unit 6 conforms to the syntax of H.264 / MPEG-4 AVC. When there is no data, it is determined that an error has occurred during decoding.

  The picture information buffer 7 can write the picture type, the accumulated code amount, the vector movement tendency, and the color difference distribution for each picture. However, since the information that can be stored in the picture information buffer 7 is limited, the picture information buffer 7 is preferably a structure (such as a ring buffer) that can overwrite information at the oldest time.

  The picture information buffer management unit 8 manages the picture information buffer 7 and is used by the user to write the picture type and cumulative code amount passed from the cumulative code amount extraction unit 5 to the picture information buffer 7. Is set, the vector movement tendency or color difference distribution or vector movement tendency and color difference distribution passed from the decoding processing unit 6 is written into the picture information buffer 7 in accordance with the contents.

  Note that the accumulated code amount of IDR pictures is considerably larger than the accumulated code amount of P pictures and B pictures due to its nature. Therefore, the cumulative code amount of the IDR picture is set to a small value within a range that does not affect the determination of whether or not the decoding error is caused by a scene change compared to the cumulative code amount of other pictures. It may be converted (for example, converted to a value divided by 4) and stored in the picture information buffer 7.

  In addition, the vector movement tendency is stored in the picture information buffer 7 by expressing the vector movement tendency as, for example, “small”, “medium”, and “large” and the vector movement tendency information is valid Is also stored together (indicated by a circle in the figure).

  The color difference distribution is stored in the picture information buffer 7 by expressing the color difference distribution as, for example, “small”, “medium”, and “large” and indicating that the color difference distribution information is valid. (Indicated by a circle in the figure) is written together. Note that the color difference distribution of the IDR picture is considerably larger than the color difference distribution of the P picture and B picture due to its nature. Therefore, the color difference distribution of the IDR picture is “large” so that it is convenient when determining whether or not the decoding error is caused by a scene change compared to other pictures. Alternatively, it may be stored as “medium”.

  When the decoding error information received from the decoding processing unit 6 indicates “decoding error”, the determination instruction unit 9 reads the picture information in the picture information buffer 7 and the decoding error is caused by a scene change. It is determined whether or not it is an object, and an instruction as to what concealment method is to be performed is given to the decode processing unit 6.

  In an embodiment of the present invention, the determination instructing unit 9 performs decoding when the cumulative code amount of a picture in which an error has occurred during decoding changes in a direction that increases compared to the cumulative code amount of the previous picture. It is determined that the time error is caused by a scene change. In this case, as a concealment method corresponding to this, the determination instruction unit 9 freezes the previous picture in place of the picture having the decoding error or displays a macroblock having the decoding error in the past at the same position. Indicates an inter concealment to be replaced with a macroblock. Whether to instruct freezing or to instruct interconcealment depends on the setting of the user.

  On the other hand, when the cumulative code amount of a picture in which a decoding error has occurred has changed in a direction that decreases compared to the cumulative code amount of the previous picture, the determination instruction unit 9 It is determined that it is not caused by a scene change. In this case, as a concealment method corresponding to this, the determination instruction unit 9 has an intra concealment in which a macroblock having a decoding error is filled with an average of pixels of surrounding macroblocks or a decoding error. A single color (for example, gray to fill in gray) that fills the macro block with a single color is instructed. Note that whether to indicate intra concealment or single colorization depends on the user's setting.

  2 to 4 are circuit diagrams for explaining the operation of the embodiment of the present invention, and only the accumulated code amount is used as a determination material for determining whether or not the decoding error is caused by a scene change. Take the case as an example. FIG. 2 shows a case where no decoding error has occurred, FIG. 3 shows a decoding error determined to be caused by a scene change, and FIG. 4 shows a decoding error which is not judged to be caused by a scene change. It is.

  In one embodiment of the present invention, the accumulated code amount extraction unit 5 extracts the picture type and accumulated code amount of each picture from the information held in the input compressed video stream, and extracts the extracted picture type and accumulated code. The amount is passed to the picture information buffer management unit 8. The picture information buffer management unit 8 writes the picture type and cumulative code amount passed from the cumulative code amount extraction unit 5 to the picture information buffer 7 for each picture.

  In addition, the decoding processing unit 6 performs decoding processing of the input compressed video stream, and when no decoding error is detected, as shown in FIG. Pass the information “no error”. In this case, the determination processing unit 9 sends “no concealment” as the determination instruction. In response to this, the decoding processing unit 6 outputs the decoded video data as it is without performing the concealment processing.

  On the other hand, when the decoding processing unit 6 detects a decoding error, as shown in FIGS. 3 and 4, the decoding processing unit 6 determines information “decoding error exists” as decoding error information. It passes to the instruction unit 9. In this case, the determination instructing unit 9 reads out the accumulated code amount of the picture in which an error has occurred during decoding and the previous picture from the picture information buffer 7, and from the change in the accumulated code amount, the error in decoding is caused by a scene change. It is determined whether or not it is to be performed, and a concealment method corresponding to the determined content is instructed to the decode processing unit 6.

  Here, FIG. 3 exemplifies a case where a P picture having a cumulative code amount of 32 causes an error during decoding. In this example, the cumulative code amount of the P picture before the P picture in which the decoding error has occurred is 16, and the cumulative code amount of the P picture in which the decoding error has occurred has increased to 32. Therefore, the determination instruction unit 9 determines that the decoding error is caused by a scene change, and instructs freeze or interconcealment according to the contents set by the user.

  FIG. 4 shows an example in which a P picture having a cumulative code amount of 8 has an error during decoding. In the case of this example, the cumulative code amount of the P picture before the P picture in which the decoding error has occurred is 16, and the cumulative code amount of the P picture in which the decoding error has occurred has changed to 8 and decreases. Therefore, the determination instructing unit 9 determines that the decoding error is not caused by a scene change, and instructs intra concealment or single color (for example, graying) according to the content set by the user.

  As described above, the determination instruction unit 9 determines that the decoding error is caused by the scene change when the cumulative code amount of the picture in which the decoding error has occurred is larger than the cumulative code amount of the previous picture. However, when the cumulative code amount of a picture in which an error has occurred during decoding changes with respect to the cumulative code amount of the previous picture, the user determines whether the error during decoding is due to a scene change. May be configured to be settable.

  As described above, in the embodiment of the present invention, when the decoding processing unit 6 detects an error during decoding, the determination instruction unit 9 reads the accumulated code amount from the picture information buffer 7 and detects the change in the accumulated code amount. It is determined whether or not the decoding error is caused by a scene change. When it is determined that the error is caused by a scene change, the decoding processing unit 6 is informed by a concealment method suitable for the scene change. It is supposed to indicate a freeze or interconcealment. Therefore, the reception side can suppress the expression of the “ghost phenomenon” due to the scene change. As a result, the scene change process on the transmission side is unnecessary, and the use of the existing device can be continued on the transmission side.

  In one embodiment of the present invention, when determining whether or not the decoding error detected by the decoding processing unit 6 is caused by a scene change, the vector movement tendency or the color difference distribution or the vector movement tendency and the color difference are determined. When using a distribution, the user determines whether the decoding error is caused by a scene change when the vector movement tendency or color difference distribution or vector movement tendency and color difference distribution changes. It can depend on the content to be.

  Further, one embodiment of the present invention is used for a one-segment mobile terminal, but the present invention is not limited to this, and may be used for various devices that decode a compressed video stream. it can.

It is a circuit diagram showing one embodiment of the present invention. It is a circuit diagram for demonstrating operation | movement of one Embodiment of this invention. It is a circuit diagram for demonstrating operation | movement of one Embodiment of this invention. It is a circuit diagram for demonstrating operation | movement of one Embodiment of this invention. It is a figure for demonstrating a "ghost phenomenon."

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Image before scene change 2 ... Image at the time of scene change 3 ... Image after scene change 5 ... Cumulative code amount extraction unit 6 ... Decoding processing unit 7 ... Picture information buffer 8 ... Picture information buffer management unit 9 ... Determination instruction unit

Claims (4)

A decoding processing unit for decoding the compressed video stream;
When the decoding processing unit detects a decoding error, it is determined whether the decoding error is caused by a scene change, and a concealment method corresponding to the determined result is transmitted to the decoding processing unit. A determination instruction unit for instructing;
A decoding device comprising: A cumulative code amount extracting unit that extracts a cumulative code amount of each picture from the compressed video stream;
The decoding apparatus according to claim 1, wherein the determination instruction unit determines whether the decoding error is caused by a scene change based on the accumulated code amount. A cumulative code amount extracting unit that extracts a cumulative code amount of each picture from the compressed video stream;
The decoding processing unit calculates a vector movement tendency or a color difference distribution,
The determination instruction unit determines whether or not the decoding error is caused by a scene change based on the accumulated code amount and the vector movement tendency or the color difference distribution. Item 2. The decoding device according to Item 1. A cumulative code amount extracting unit that extracts a cumulative code amount of each picture from the compressed video stream;
The decoding processing unit calculates a vector movement tendency and a color difference distribution,
The determination instruction unit determines whether the decoding error is caused by a scene change based on the accumulated code amount, the vector movement tendency, and the color difference distribution. The decoding device according to claim 1.

Images