JP2003284069A - Method and apparatus for decoding multiple data - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus which multiplexes, transmits, and decodes image and sound data, lowers jitters of a clock, and performs decoding reliably, even when the content of the data to be transmitted changes. <P>SOLUTION: When a first packet train composed by multiplexing a packet comprising a kind or more of digital data and a packet comprising a reference time signal of a decoder is inputted, and a second packet train composed by multiplexing a packet comprising a kind or more of digital data and a packet comprising a reference time signal of the decoder is inputted, and the first packet train and the second packet train are multiplexed and a third packet train is outputted, the reference time signal is removed from the packet comprising the reference time signal in the first packet train in this data multiplexing method, when the packet comprising the reference time signal in the first packet train, and the packet comprising the reference time signal in the second packet train are inputted within a specified time. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data multiplexing method for multiplexing and transmitting digital images, voices, digital data, etc., a data transmission method, and a decoding method and apparatus for decoding the multiplexed data. Is.

[0002]

2. Description of the Related Art In recent years, with the progress of compression technology, broadcasting, CA
Some digital image and voice services have been put to practical use on TVs. When transmitting digital images, voice, and other data, how to multiplex those data as well as compression techniques is an important issue.

An MPEG data multiplexing method, which is an international standard of the above-described data multiplexing method, and a multiplex data decoding apparatus for decoding multiplexed data will be described below with reference to the drawings.

11, 12, and 13 are explanatory views of a conventional MPEG2 data multiplexing method. FIG. 14 is a block diagram of a multiplex data decoding apparatus for decoding data multiplexed by the MPEG2 data multiplexing method. In FIG. 14, reference numeral 141 denotes a separating unit.
It is composed of a buffer 411 and a CPU 1412. Further, 142 is a clock control unit, 143 is an image decoder, 1
Reference numeral 44 is an audio decoder. FIG. 15 shows the CPU 14 of the separation unit.
12 is a flowchart showing the operation of No. 12.

The operation of the conventional MPEG2 data multiplexing method and multiplex data decoding apparatus configured as described above will be described below.

Image data and audio data are composed of a fixed number of samples (MP
In the EG, even in the case of voice, this is also called a frame) and compression-coded for each frame, or one frame or multiple frames
Create a packet called a PES packet. FIG. 13 shows the outline of the PES packet format. A header is attached to the PES packet, but in the header, the stream ID indicating whether the subsequent data area is image data, audio data, or other data, or special reproduction It includes information such as a trick mode flag indicating that the data is for. The PES packet is divided into a plurality of 188-byte-long TS packets, which will be described later, and transmitted.

FIG. 11 shows an outline of the format of a TS packet. Each TS packet has a packet-specific number called PID. The PID is not different for each TS packet, but the same PES packet has the same PID. As shown in FIG. 11 (b), the TS packet sends information to be used for separating data or multiplexed data called an adaptation field, following the header. Since the length of the adaptation field depends on the information to be sent, if the length is short, it is possible to send data following the adaptation field. The PID of the TS packet is sent as a part of the header as shown in FIG. 11 (c). FIG. 11 (d) is a diagram showing a part of the contents of the adaptation field, and FIG. 11 (e) is an explanatory diagram showing a part of the contents of the optional fields of the adaptation field. The adaptation field is shown in Fig. 11 (e) as an optional field.
It is possible to send the reference time clock PCR necessary for decoding as in. In MPEG2, the clock frequency of the multiplexer and demultiplexer is set to 27MHz, and the clock of the multiplexer is set to prevent overflow underflow of the buffer.
I am sending the PCR, the number counted at 27MHz. The demultiplexer regenerates the clock synchronized with the multiplexer from the received PCR. This PCR does not have to be sent at all times, but is sent at intervals necessary for synchronizing the separation device, for example, once every 0.7 seconds. Therefore, the adaptation field has a structure such that it may or may not be sent as an optional field. Also, when sending different programs, it is often the case that different image encoders encode.
The reference clock is generally different. When sending different PCRs with the same program number or the like, the discontinuity flag in the adaptation field is set to 1 to notify the separator that the PCR discontinuity exists.

In addition to the PES packet shown in FIG. 11, it is possible to send program selection information called PSI to the data area of the TS packet. FIG. 12 shows the structure of the program table PMT which is one of PSI. A program number indicating the number of the program to be transmitted, and PIDs of PCR, image data PES, audio data PES, and data PES packet, which are necessary for reproducing the program, are multiplexed and sent to the PMT.

FIG. 14 shows an example of a conventional multiplex data decoding apparatus for receiving the data multiplexed as described above and reproducing the data. First, the received multiplexed data is buffer 141
Remember with 1. The CPU 1412 operates based on the flowchart shown in FIG. 15, and searches the first transmitted data for the PMT having the program number set from the outside. When the PMT is detected, the PID of the PCR, the image data PES, the audio data PES, and the data PES packet is extracted from the PMT. Next, looking at the PID of the TS packet in the received data, the TS packet necessary for reproducing the program is extracted. The packet including PCR among the TS packets is the clock control unit 142.
Sends to the picture, reproduces the reference clock required for the audio decoder. The image and audio packets of the TS packets are sent to the respective decoders 143 and 144. The decoding unit can decode the image and audio data using the reproduced reference clock and decompress and reproduce the program (for example, ISO / IEC JTC1 / S
C29N801, &quot; ISO / IEC CD 13818-1: Information tec
hnology-Generic codingof moving pictures and ass
ociated audio: Systems &quot;, 1994.11).

[0010]

However, various problems occur in the decoding device only with the above-mentioned configuration. First, when TS packets of a plurality of programs are multiplexed, TS packets having a reference clock PCR may collide during packet multiplexing, resulting in fluctuation jitter in the PCR packet intervals. In the decoding device, a PLL is used to regenerate the reference clock. However, if the PCR interval exceeds a certain value, the jitter cannot be absorbed and the clock cannot be regenerated.

Further, PCR becomes discontinuous at the beginning of the program or when the program is switched, but the discontinuity flag in FIG. 11 is at a position indicating discontinuity before the discontinuous packet arrives. There is no guarantee. Therefore, in the worst case, the next discontinuous PCR arrives without the discontinuity flag, and the decoding device tries to synchronize with a completely different PCR, resulting in unstable operation.

Further, when the frame frequency of the image or the sampling frequency of the audio changes, the output clock from the decoding device also needs to be changed, but in general, a certain time is required to establish synchronization, and the sampling frequency is If different data is received, the synchronization system cannot be established and the output may be disturbed.

On the other hand, in FIG. 13, there is a trick mode flag in the PES header, and commands such as fast forward and still can be sent together with the PES packet, but there is no clear regulation of the operation of the decoding device and there is no guarantee of reliable operation. .

In view of the above problems, the present invention provides a data multiplexing method, a data transmission method, a multiple data transmission method, a multiple data decoding method, and multiple data for ensuring the operation of a multiplexed data decoding device in normal or special reproduction. A decoding device is provided.

[0015]

In order to solve the above problems, the data multiplexing method of the present invention is a first method in which a packet containing one or more kinds of digital data and a packet containing a reference time signal of a decoding device are multiplexed. Input packet sequence, and a second packet sequence in which a packet including one or more types of digital data and a packet including a reference time signal of a decoding device are input, and the first packet sequence and the second packet sequence are input. When the third packet sequence is output by multiplexing the second packet sequence, the packet including the reference time signal in the first packet sequence and the packet including the reference time signal in the second packet sequence are constant. If you enter within the time
The reference time signal is deleted from the packet including the reference time signal in the packet sequence of, or the data packet including one or more kinds of digital data and the table packet including the table indicating the packet content and the packet number are multiplexed. In the case of generating a multiplex packet by changing the contents of the table packet, before sending the data packet corresponding to the changed table packet, an instruction signal indicating that the table packet is changed and the changed table packet are In the case of multiplexing in a multiple packet, or in the case of multiplexing a packet containing one or more kinds of digital data and a packet containing the reference time signal of the decoding device, the contents of the digital data before the packet containing the digital data or the contents of the digital data are not If continuous, includes discontinuous digital data Packet of the reference time signal to be multiplexed before, the packet is assigned a reference time signal discontinuity flag indicating that the reference time signal is not continuous, the flag is also intended to multiplexing.

Further, the data multiplexing method of the present invention, when performing special reproduction using a plurality of packets containing audio data, image data and other digital data,
Alternatively, a packet including audio data constitutes a special reproduction packet only with data of the same frame, and the special reproduction packet is transmitted. In addition, the multiplex data decoding method of the present invention decodes a packet including digital image data compressed using intraframe or intrafield compression coding and interframe or interfield compression coding, and uses the decoded data to determine an arbitrary value. When stationary in a frame or field, at the time of receiving a stationary command signal, the image obtained by decoding the immediately before or immediately after the intraframe or intrafield compression-encoded image data is displayed as a still image, or compressed digital image data When a packet with a flag indicating the decoding or display time is decoded and the still data is stopped in an arbitrary frame or field using the decoded data, the time immediately before or immediately after the still command signal is received. Shows the decoding or display time of accessible image data Storing the lugs, at rest resume sending flag showing decoding or display time of the image data that can be the access multiplexer, is to resume decoding from the accessible image data.

(Operation) The present invention has the above-mentioned structure.
Since the interval of the reference signal PCR becomes a certain value or less, it does not hinder the clock reproduction. In addition, when the program is changed, the PMT, PCR, etc. are transmitted in advance of the changed data, so that the operation of the decoding device is ensured. Furthermore, at the time of trick play, the data of only the same picture is formed into a PES packet and is sent to the decoding device, so that the image is not interrupted on the way. It is possible to reproduce a continuous image even after the end of the still picture by transmitting the time information of the still picture to the transmission device and retransmitting it from the image data of the still picture even when the image display is still. Becomes

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION A data multiplexing method according to a first embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a data multiplexer in the first embodiment of the present invention. In FIG. 1, 1 is the first
The data multiplexing unit of the above comprises a clock generation circuit 11, a CPU 12, and a buffer 13. Reference numeral 2 is a second data multiplexing unit, and 3 is a second data multiplexing unit, both of which have the same configuration as the first data multiplexing unit. FIG. 2 is an explanatory diagram showing a part of the operation of the CPU 32 in the third data multiplexing unit.

A data multiplexing method in the data multiplexing apparatus configured as described above will be described with reference to FIGS. 1 and 2.

Image data, audio data, and other data to be multiplexed are compression-encoded by an image encoder 14, an audio encoder 15, and a data encoder 16, respectively, to form PES packets as shown in FIG. Input to the data multiplexing section of. The first and second data multiplexers 1 and 2 perform multiplexing while converting PES packets into TS packets so that image, audio and data synchronization and buffer overflow and underflow do not occur in the multiplex data decoding device. . The outputs of the data multiplexer 1 and the data multiplexer 2 are further multiplexed by the data multiplexer 3. This is because, for example, the output of the data multiplexers 1 and 2 is 4 Mbps, and both are 8
It is assumed that the data will be sent to the Mbps transmission line. As shown in FIG. 2, the data multiplexing unit 3 inputs TS packets from the data multiplexing unit 1 and the data multiplexing unit 2 at 210 and 211, and multiplexes the respective inputs alternately and outputs them. The problem then arises when both packets simultaneously contain the reference clock PCR required by the multiplex data decoder. In order to facilitate the design of the multiplex data decoder, it is desirable to suppress the fluctuation of the reference clock PCR to a certain value or less, and in the example cited in the conventional example, 30pp
It is less than m. However, when the TS packets of the multiplexing units 1 and 2 are input to the data multiplexing unit 3 at the same time, a shift of 1 TS packet occurs due to multiplexing, and as a result, PCR fluctuates. Therefore, in this embodiment, the reference clock PCR is frequently inserted in the multiplexed data in advance, and when a collision occurs, the PCR field included in the TS packet on the input side of the multiplexing unit 1 as shown in the flowcharts 212 and 215 of FIG. Will be erased. The PCR field is optional and deleting it has no effect. Although the interval of the packet including the PCR will be slightly extended, it does not hinder the operation of the multiplex data decoding device because it is frequently inserted in advance.

As described above, according to this embodiment, data multiplexing that does not adversely affect the reference clock PCR of the multiplex data decoding device can be realized.

FIG. 3 is an explanatory diagram of a data multiplexing method in the second embodiment of the present invention, and FIG. 4 is an explanatory diagram of a multiplexed data decoding method. The flowchart in FIG. 3 shows the operation of the CPU 12 of the first data multiplexing unit 1 in FIG. Further, FIG. 4 shows the CPU 141 of the multiplex data decoding device 141 in FIG.
FIG. 3 is an explanatory diagram showing the operation of 1. This will be described below with reference to FIGS.

In the data multiplexer 1 during normal operation, PMT,
Send PCR at the right time, and
TS conversion of data PES packet at appropriate timing,
The operation of multiplex transmission is repeated.

On the other hand, at the start of the program, since the multiplex data decoding device is not synchronized with the multiplex device, the reference time PC
It is necessary to send a packet containing R. When the PCR packet is sent, the decoding device attempts synchronization based on the PCR sent. PC that is completely unrelated to the decryption device when the program starts
Since R is sent, it takes time to synchronize. So Fig. 3
P indicating that the PCR is discontinuous like 319 in
Set the CR discontinuity flag to 1. When the decoding device detects that the discontinuity flag is 1, it does not attempt to synchronize with the sent PCR, but sets the sent PCR value and waits for the next PCR. Therefore, the time until the synchronization is shortened. Further, the above operation is effective not only at the start of the program but also when the reference time PCR changes due to a change in content.

Next, when starting the program,
It is necessary to send a PMT that represents the PID relationship. At that time, as in 321 in FIG. 3, the validity flag of the PMT is first set to 0, and then a table based on the new packet number PID is created (322). When the valid flag of PMT is 0, the multiplex data decoding apparatus invalidates the PMT and generally does not use the PMT. Next, when a PMT with a valid flag of 1 comes, decoding is performed using the PMT. In the MPEG2 standard of the conventional example, there is no clear regulation of the program change and the method of using the PMT valid flag. Therefore, it is possible to send different PMTs with the PMT valid flag set to 1. However, in general decoding devices, the PMT changes and the changed PMT
A certain period of time is required for a series of operations such as obtaining the PID of the TS packet required for decoding from, and starting extraction of the TS packet having the desired PID from the transmission data. As a result, the packet immediately after receiving the modified PMT cannot be extracted accurately and some data is lost,
Images and sounds may be interrupted for a moment. Therefore, in the present invention, before sending a new PMT, that is, when the program is started or when the content of the program is changed, the PMT valid flag 0 is always added and then the PMT corresponding to the new content is added.
Is sent.

The multiplex data decoding apparatus has the PM having the valid flag 0 as shown in the flow charts 46 and 47 of FIG.
When T is received, the required PID is obtained from the PMT, and at the start of the program, the packet having the PID necessary for decoding is prepared for extraction, that is, the PMT is updated, and then the valid flag 1 is set.
Extraction starts from the PID that follows the PMT with. Also, when the contents of the program change in the middle, when the PMT having the valid flag 0 is received, the extraction of the currently valid PID is continued, the PMT is updated with the new PMT, and then the valid flag 1 is provided. The extraction of the TS packet having the PID corresponding to the new content is started from the PID following the PMT.

As described above, the PMT having the valid flag 0
Is transmitted prior to actually valid PMT, even if there is a delay in the decoding device, there is no PID extraction omission, and video and audio can be played back without interruption.

Further, the frame frequency in the image and the sampling frequency in the audio are frequencies that determine the output frequency and operation of the decoding device, and the clock that is the basis of these is generally generated from the received reference clock PCR.
Emphasis characteristics in voice also change the output voice,
It is necessary to set the characteristics themselves, and changing these settings requires a certain amount of time. Therefore, if the data whose characteristics are changed is continuously sent, it becomes impossible to accurately decode the data of the packet received in the time required for the setting. Therefore, in the present invention, in FIG.
As shown by 4, 325, 326, 327, and 328, before sending the changing data, describe the frame frequency of the changed image in the area called the descriptor of the PMT, the sampling frequency or the emphasis characteristic in the audio. Send out. By transmitting the data as described above, the decoding device can change the clock or the characteristics before receiving the actual data, and thus the normal decoding can be performed even if the changed data is received. .

FIGS. 5 and 6 are explanatory views showing a data multiplexing method in the third embodiment of the present invention. This will be described below with reference to FIGS. 5 and 6.

The PES packet of an image may include a plurality of pictures in one packet as shown in FIG. 5 (a). Here, a picture refers to a frame or field in MPEG image compression. Whether it is a frame or a field depends on the compression encoding method. Figure 5
In (a), the data of two pictures are included in the same packet. When performing special reproduction such as fast-forward, slow-motion, still, etc., as shown in FIG. 13, a trick mode flag in the PES header for informing the data decoding device of the operation such as fast-forward is set and sent together with the image data. By reading these flags, the decoding device performs buffer control and screen control. Usually, these trick play operations are controlled in units of frames or fields, that is, pictures. However, when data of one picture or more is mixed in the PES packet for the case of FIG.
It is necessary to reconstruct the data for one picture from the S packet. Therefore, in the present invention, the PES packet as shown in FIG. 5 (a) is reconfigured into one picture / one packet before transmission as shown in FIG. 5 (b) and then transmitted. Figure 6
It is used as an algorithm of the CPU 12 of the first data multiplexing unit 1 in FIG. 1 in the flowchart for realizing the above operation. As for the operation of the CPU, when a signal requesting trick play is input, first check at 64 whether one picture consists of one PES packet, and if not, at 66, 67, 68 the position of the PES header. To be immediately before the picture header.

With the above operation, only one picture is included in one packet, the correspondence between the trick mode flag and the picture data is clarified, and at the same time, complicated operation of the decoding device or hardware is not required. .

Although only the image data has been described in the above embodiments, the same operation is effective for the audio data. However, in the case of audio data, it is not a picture but a frame indicating a unit of compression encoding.

FIG. 7 is an explanatory diagram showing a multiplex data decoding method in the fourth embodiment of the present invention, and FIG. 8 is an explanatory diagram showing the contents of multiplex data. This will be described below with reference to FIGS. 7 and 8.

Now, the image data of the multiplexed data is shown in FIG.
As described above, it is assumed that the frame is composed of an intra-frame coded picture (I picture), a forward prediction inter-frame coded picture (P picture), and a bidirectionally predicted inter-frame coded picture (B picture). The I picture can be decoded by itself, whereas the P picture cannot be decoded without the previously transmitted I picture and the B picture without the previously transmitted I and P pictures. In FIG. 8, I follows B picture
1, B0, P3, B2 are transmitted in this order. Decoding is done in transmission order, but the display is B0, I
The order is reversed like 1, B2, P3. It is also known that, when image compression encoding is performed by the above-described method, it is generally preferable that the image quality of I and P pictures is good and the image quality of B picture is slightly deteriorated.

When the image is stopped during special reproduction, an image with better image quality can be obtained by stopping the image other than the B picture for the above reason. Further, it is desirable to start from the same frame when releasing the stillness. When transmitting the above data,
It is common to put a sequence header in front of an I-picture that can be decoded by itself to enable random access. Therefore, access from the I picture, that is, resumption of stillness, is relatively easy.

FIG. 7 shows the above operation in a flowchart of the CPU 1412 of the data decoding apparatus in FIG. In FIG. 7, at 72, the DTS of the I picture (decoding time stamp: the decoding time is represented by the reference time PCR)
Always remember. When a stationary signal is input, I at 76
Check if the picture is displayed and stop updating the image display when the picture is displayed and stop the picture. After that, at 80, the last I picture stored, that is, the I that is currently displayed still
The DTS of the picture is sent to the data multiplexer. In the multiplexing device, the I picture currently displayed still is determined based on the DTS sent from the decoding device, and the transmission is started from the I picture corresponding to the DTS at the end of the stillness. In the decoding device, the B picture between the first P picture after the I picture is
Since decoding cannot be performed without an I or P picture before the I picture, the B picture is not decoded and displayed, but the I picture is decoded and displayed. By the above operation, it is possible to continuously reproduce the still I-pictures at the end of stillness without interruption.

FIG. 9 is an explanatory view showing a multiplex data decoding method in the fifth embodiment of the present invention. The difference from the fourth embodiment is that at 92, when the still signal is input, the display update is stopped as at 95 and the currently displayed image continues to be displayed as it is. At the time of the end of stillness, the currently received data is analyzed and decoding is restarted from the I picture.
According to this embodiment, it is impossible to reproduce from the same picture, but it is not necessary to send DTS to the data multiplexer,
The entire system can be constructed at low cost.

In the above embodiment, the image currently displayed is assumed to be still, but the present invention is not limited to this, and it is also possible to use an I picture as in the fourth embodiment.

FIG. 10 is an explanatory diagram showing a multiplex data decoding method in the sixth embodiment of the present invention. This example is first
The difference from the fourth and fifth embodiments is that the display is not stopped when the still signal is input, but the input is stopped as shown by 105. Generally, the image decoder continues to display the last displayable image as it is when there is no input. Therefore, by stopping the input of the decoding device, the image becomes still. Also, when the last input image is an I or P picture, it is desirable to use that picture as a still image because those pictures have good image quality.
B picture following P or P, that is I or P in display order
Since there is no B picture before the picture, the last I or P picture of the input is not displayed and the B picture sent before that is stopped. In this embodiment, 10 in FIG.
The last input picture like 6, 107, 108
In the case of I or P, the DTS indicating the decoding time and the PTS indicating the display time are ignored and the I or P picture is displayed and then stopped.

With the above operation, still special reproduction can be realized by changing only the input of the decoding device without adding new hardware to the display system.

[0042]

As described above, according to the present invention, the jitter of the reference time PCR is suppressed to a certain value or less, so that the operation of the decoding device is not hindered. When changing the program, PM
Since the data with changed contents such as T and PCR is transmitted prior to the data, the operation of the decoding device is ensured. Furthermore, at the time of trick play, the data of only the same picture is formed into a PES packet and is sent to the decoding device, so that the image is not interrupted on the way. It is possible to play back a continuous image even after the end of stillness by retransmitting from the image data of the still picture by sending the time information of the still picture to the transmission device by stillness in the intra-coded picture even when the image display is still Becomes

[Brief description of drawings]

FIG. 1 is a block diagram of a data multiplexing apparatus according to a first embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a part of the operation of the CPU 32 in the third data multiplexing unit in the first embodiment of the present invention.

FIG. 3 is an explanatory diagram of a data multiplexing method according to a second embodiment of the present invention.

FIG. 4 is an explanatory diagram of a second embodiment multiplex data decoding method of the present invention.

FIG. 5 is an explanatory diagram showing a data multiplexing method according to a third embodiment of the present invention.

FIG. 6 is an explanatory diagram showing a data multiplexing method according to a third embodiment of the present invention.

FIG. 7 is an explanatory diagram showing a multiplexed data decoding method according to a fourth embodiment of the present invention.

FIG. 8 is an explanatory diagram showing the contents of multiplexed data according to the fourth embodiment of the present invention.

FIG. 9 is an explanatory diagram showing a multiplexed data decoding method according to a fifth embodiment of the present invention.

FIG. 10 is an explanatory diagram showing a multiplex data decoding method according to the sixth embodiment of the present invention.

FIG. 11 is an explanatory diagram of a conventional multiplex data transmission method.

FIG. 12 is an explanatory diagram of a conventional multiplex data transmission method.

FIG. 13 is an explanatory diagram of a conventional multiplex data transmission method.

FIG. 14 is a block diagram of a multiplex data decoding device that decodes data multiplexed by a conventional data multiplexing method.

FIG. 15 is an explanatory diagram showing the operation of the CPU in the conventional multiplex data decoding device.

[Explanation of symbols]

1 First data multiplexer 2 Second data multiplexer 3 Third data multiplexing section 11 Clock generation circuit 12 CPU 13 buffers 14 image encoder 15 audio encoder 16 data encoder 21 Image encoder 22 voice encoder 23 Data encoder

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 5C059 PP05 PP06 PP07 RB02 RB09                       RB10 RC03 RC32 SS16 UA05                 5K028 AA03 AA11 AA14 EE02 EE03                       EE05 EE07 KK32

Claims (2)

[Claims] 1. When a packet including compressed digital image data is decoded, and the decoded data is used to stop at an arbitrary frame or field, input to a decoding device is stopped when a still command signal is received. A multiplex data decoding method characterized by: 2. When a packet including compressed digital image data is decoded, and the decoded data is used to stop at an arbitrary frame or field, the input to the decoding device is stopped when a still command signal is received. In this case, when the last input image data before the stop is decodable, the decodable image is not displayed and the decodable image is displayed as a still image.