JP3257988B2 - Image signal encoding method and recording medium recording this method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image signal encoding method, and more particularly to an encoding method for controlling a generated code amount of a frame to be encoded based on an estimation of an occupation amount of a reception data buffer on a decoding side.

[0002]

2. Description of the Related Art ISO / IEC13818-2 (MPE)
In G-2), a mechanism for estimating the occupation amount of the reception data buffer on the decoding side during encoding is adopted in consideration of the size of the reception data buffer on the decoding side. FIG. 4 shows a change in the occupation amount of the reception data buffer. In this figure, it is assumed that the encoded data of the decoded frame is immediately removed from the received data buffer at the time of decoding.

Now, a transmission rate R (bits / sec),
Consider a case where encoding is performed under the conditions of the maximum reception data buffer size Bmax (bits) and the inter-frame time T (sec). Decoded frame f _t generated code amount G _t, the decoded frame f _t B the received data buffer fullness immediately before decoding the _t
Then, the maximum generated code amount Lmax _{t + 1} of the next decoded frame f _{t + 1} is represented by the following equation.

[0004]

Lmax _{t + 1} = B _{t + 1} = B _t -G _t + RT (1) Further, the minimum generated code amount of the decoded frame f _{t + 1} is Lmin _{t + 1.}
Is represented by the following equation.

[0005]

[Equation 2] Lmin_{t + 1}= B_t-G_t+ 2R · T−Bmax (2) = B_{t + 1}+ R · T−Bmax (3) From these results, the decoded frame f_{t + 1}Generated code amount G of
_{t + 1}Is Lmin_{t + 1}≤G_{t + 1}≤Lmax_{t + 1}Within the range of
Control to prevent overflow or lack of received data.
Can be.

[0006] The following method is used as a method for controlling so as to fall within this range.

(1) Prevention of overflow of received data: Overflow of received data occurs when the code amount of a decoded frame falls below the minimum generated code amount. Therefore, the amount of generated code is increased by adding extra encoded data. For example, in MPEG-2, "0" of an arbitrary length can be added at a predetermined location. This is called zero stacking, and the added “0” is ignored on the decoding side.

(2) Prevention of lack of received data: The lack of received data occurs when the code amount of the decoded frame exceeds the maximum generated code amount. Therefore, at the time of encoding, processing for suppressing the generated code amount is performed. As a method of suppressing the generated code amount, there are a method of reducing the encoding frame rate, a method of increasing the quantization width, and a method of not encoding small blocks until the condition is satisfied.

[0009]

In a moving picture coding method such as MPEG-2, coding efficiency is improved by performing motion compensation prediction. In particular, MPEG-2
Then, there are three types of encoded frame types depending on the type of the motion compensation prediction. Intra-frame coding type (I frame) not using motion compensated prediction, unidirectional prediction coding type (P frame) for performing prediction from past image frames, bidirectional prediction coding type for performing prediction from past and future image frames (B frame).
These frame types have the characteristics shown in Table 1 below.

[0010]

[Table 1]

FIG. 5 shows an example of a reference relationship between the respective coded frame types. As described above, since the reference relationship differs depending on the coded frame type, the influence of the deterioration of the image quality of the coded frame on other coded frames is different. For example, if the image quality deteriorates in a certain P frame,
Image quality is degraded for all encoded frames that reference the frame. Conversely, a decrease in image quality in a B frame does not affect other encoded frames.

In the control of the generated code amount, the limitation by the maximum generated code amount involves a process of reducing the generated code amount, and thus involves a reduction in image quality. In the conventional method, the minimum generated code amount is determined by considering only the occupation amount of the reception data buffer. For this reason, when a process of reducing the generated code amount by the maximum generated code amount is performed on a coded frame referred to from another coded frame, the image quality of all the coded frames referencing this coded frame is reduced. There is a problem that decreases.

An object of the present invention is to provide an image signal encoding method capable of minimizing deterioration of image quality while preventing overflow or shortage of a reception data buffer.

[0014]

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and is characterized by the following encoding method.

[0015] determining a (first invention) image signal when coding, the processing priority of suppressing the amount of generated code for each encoded frame, the
Uncoded frames whose priority is lower than the frame to be coded
And the detected uncoded frame
Estimating the received data buffer occupancy when decoding the data, determining the maximum generated code amount of the encoding target frame from the estimated received data buffer occupancy, and generating the encoding target frame. Performing a process of suppressing the generated code amount of the encoding target frame when it is determined that the code amount exceeds the maximum generated code amount.

According to the present invention, the priority of the processing for suppressing the generated code amount is determined for each encoded frame. This priority is
The priority is determined based on the reference relationship between the respective encoded frames. Also, bidirectional predictive coding frames (B frames),
The priority is set higher in the order of the unidirectional predictive coded frame (P frame) and the intra-coded frame (I frame).

In the process of encoding the encoding target frame, first, the occupancy of the reception data buffer is estimated, and the maximum generated code amount of the encoding target frame is calculated based on the priority determined in advance. In addition, when the encoding target frame is divided into small blocks and the encoding process is performed for each small block, if it is determined that the generated code amount of the encoding target frame exceeds the maximum generated code amount, the encoding target frame is Is performed to suppress the generated code amount.

[0018]

Further, when determining the maximum amount of generated code,
An unencoded frame having a priority lower than the encoding target frame is detected, and the occupation amount of the received data buffer when decoding the detected unencoded frame is estimated. From the estimated buffer occupancy, the maximum generated code amount of the encoding target frame is determined. For example, the priorities of three types of coded frame types (I, P, B) are P _I , P _P , P _B (P _I > P _P)
> P _B, i.e., B, if you P, performs processing to suppress preferentially generated code amount in the order of I) and, the maximum amount of generated code G _B B frames, as well as B frames uncoded,
The buffer occupancy of the uncoded I frame and P frame is also estimated, and the maximum generated code amount of the encoding target B frame is determined.

Further, in the future, it is possible to leave an image having a high degree of importance to be encoded without performing the process of suppressing the generated code amount. For example, when the encoding target frame is a B frame and an unencoded frame includes a P frame, the received data buffer may not have enough room to skip when processing the unencoded frame later. Even if the received data buffer seems to have a margin at first glance, it performs processing to suppress the code amount for the B frame that is the current processing target.

(Embodiment 2 ) A step of estimating a generated code amount of the uncoded frame when estimating a received data buffer occupancy when decoding the uncoded frame, And estimating the occupation amount of the reception data buffer in the uncoded frame using

According to the present invention, the generated code amount of each uncoded frame is estimated, and the occupation amount of the received data buffer of each uncoded frame is estimated from the estimated generated code amount.

( Third invention) When determining the maximum amount of generated code of the frame to be coded, the maximum amount of generated code in each uncoded frame is determined from the estimated occupancy of the received data buffer for each uncoded frame. And calculating a maximum generated code amount of the encoding target frame from the calculated maximum generated code amount of each uncoded frame.

[0024]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a flowchart showing an encoding method according to an embodiment of the present invention. In the present embodiment, it is assumed that there are three types of encoded frame types (I, P, B), a method of dividing an image signal into blocks, and performing encoding for each block.

As shown in FIG. 2, the coding structure of the coded frame is such that one I frame per N frame, I or P
The number of B frames is M between frames, and is fixed. The priority of each coded frame type (I, P, B) is P
_{Let I} , P _P , P _B (P _I > P _P > P _B ).

The detection of the uncoded frame is performed between the frame to be coded and the coded frame within the uncoded frame. As the estimated generated code amount of an uncoded frame, the generated code amount of a coded frame of the same frame type that has been coded in the past is used.

Hereinafter, the processing according to steps S1 to S6 in FIG. 1 will be described in detail.

(Procedure S1) Check of encoding frame type Encoding target frame f _t (t = 0, 1, 2,..., N−
Check coding frame type T _i 1).

(Procedure S2) Detection of uncoded frame From the frame to be coded _ft to the next I frame,
Among the encoded frames, an unencoded frame whose priority is equal to or lower than the encoding target frame is detected. This detection is as follows: (a) When the encoding target frame _ft is a B frame, all uncoded frames up to the next I frame are detected.

(B) If the encoding target frame _ft is a P frame, uncoded frames of I and P frames up to the next I frame are detected.

(C) If the encoding target frame f _t is an I frame, there is no unencoded frame to be detected, so nothing is detected.

(Procedure S3) Estimating the occupation amount of the reception buffer in the uncoded frame The code amount generated of the coded frame of the same frame type coded immediately before is used as the estimated code amount of the uncoded frame. Using this estimated generated code amount, the buffer occupancy Bs in each detected uncoded frame fs (s = t + 1, t + 2,..., N−1) is estimated. _{Assume that} the estimated generated code amount of the encoding target frame f _t is Gt ′, the received data buffer occupancy amount is Bt, the transmission rate is R, and the time between frames is T. At this time, the buffer occupancy Bs when decoding the detected uncoded frame is obtained by the following equation.

[0033]

(Equation 3)

Here, G _j ′ is the estimated generated code amount of the uncoded frame. However, uncoded frames not detected in step S2 are set to G _j ′ = 0.

(Procedure S4) Estimating the Maximum Generated Code Amount in Each Uncoded Frame Assuming that the estimated generated code amount of the uncoded frame s detected in the step S2 is G _s ′, estimation is performed up to the uncoded frame s. Generated code amount G of the encoding target frame when used for
_t and _s can be calculated from the following equations.

[0036]

(Equation 4)

(Step S5) Detecting the minimum value of the maximum generated code amount The minimum value of the maximum generated code amount G _t , _s calculated in step S4 is
The maximum amount of generated codes G _t of the encoding target frame f _t.

[0038]

G _t = min (G _t , _s ) (7) (Procedure S6) The encoding target frame is divided into small blocks, and the encoding process is performed for each block. The encoding target frame is divided into small blocks. performs coding processing in block units, when the generated code amount of the encoding target frame exceeds the maximum amount of generated code G _t, performs a process of suppressing the amount of code generated for the encoding target block.

In the present embodiment, only the shortage of the received data is considered, but when the overflow of the received data buffer is considered, the method of estimating the buffer occupancy in the uncoded frame in step S3 in FIG. What is necessary is just to change as shown by procedure S3 'in FIG.

In FIG. 3, the generated code amount of the immediately preceding encoded frame of the same frame type is assumed to be the estimated generated code amount of the uncoded frame. Using this estimated generated code amount, each detected uncoded frame f _s (s = t +
1, t + 2,..., N−1). The estimated generated code amount of the encoding target frame ft is represented by G
t ′, the occupation amount of the received data buffer is Bt, the transmission rate is R, and the time between frames is T. At this time, the buffer occupancy Bs at the time of decoding the detected uncoded frame
Is obtained from the following recurrence formula.

[0041]

B _k = B _k−1 −G ′ _k−1 + R · T (8) where G _j ′ is the estimated generated code amount of the uncoded frame. However, the uncoded frames not detected in step S2 are set to G _j ′ = 0. The calculation of this recurrence formula is repeated until k = s, and the buffer occupancy Bs is calculated. If the buffer occupancy B _k during the repetition of the calculation exceeds the maximum buffer size B _max , B _k = B _max
Continue the calculation.

By performing such processing, overflow of the reception data buffer can be considered when estimating the buffer occupancy.

This process is not for preventing the overflow of the reception data buffer, but for reflecting the fluctuation of the buffer occupancy due to the process of preventing the overflow of the reception data buffer in the estimation of the buffer occupancy in the uncoded frame. It is. Therefore, when overflow of the reception data buffer is detected during the encoding process, it is necessary to perform a process such as addition of extra data, such as zero stacking of MPEG-2.

A part or all of the method shown in FIG. 1 or FIG. 3 is described in a computer program so that the method can be executed, and the method is executed by a computer-readable recording medium such as a floppy disk (FD). And M
It can be recorded on O, ROM, memory card, CD, DVD, removable disk, etc., provided, and distributed.

[0045]

As described above, according to the present invention, it is possible to give a priority to the processing of suppressing the amount of generated codes according to the coding frame type. As a result, images having little effect on image quality are given priority. It is possible to perform processing for suppressing the amount of generated codes, and it is possible to minimize deterioration of image quality while preventing overflow or shortage of the reception data buffer.

[Brief description of the drawings]

FIG. 1 is a processing procedure of an encoding method according to an embodiment of the present invention.

FIG. 2 illustrates an encoded frame structure according to the embodiment.

FIG. 3 is a main processing procedure of an encoding method according to another embodiment of the present invention.

FIG. 4 is an example of estimating a reception data buffer occupancy.

FIG. 5 is a reference relationship between encoded frames.

────────────────────────────────────────────────── ─── Continuation of front page (56) References JP-A-5-227524 (JP, A) JP-A-6-350983 (JP, A) JP-A-6-205393 (JP, A) (58) Field (Int.Cl. ⁷ , DB name) H04N ^7/ 24-7/68

Claims (6)

(57) [Claims] When encoding an image signal, deciding a priority order of a process for suppressing a generated code amount for each encoded frame, and an unencoded frame whose priority order is equal to or less than a frame to be encoded.
Detecting the received data buffer occupancy when decoding the detected unencoded frame ; and determining the maximum generated code amount of the encoding target frame from the estimated received data buffer occupancy. Determining, and when it is determined that the generated code amount of the encoding target frame exceeds the maximum generated code amount, performing a process of suppressing the generated code amount of the encoding target frame. Encoding method of an image signal to be encoded. 2. The coding method of an image signal according to claim 1 , wherein when estimating an occupation amount of a received data buffer when decoding the uncoded frame, an estimated code amount of the uncoded frame is estimated. And estimating the occupancy of the received data buffer in the uncoded frame using the estimated generated code amount. 3. The image signal encoding method according to claim 1 , wherein, when determining a maximum generated code amount of the encoding target frame, the maximum occupied amount of a received data buffer for each unencoded frame is determined. Calculating the maximum generated code amount in the uncoded frame, and calculating the maximum generated code amount of the encoding target frame from the calculated maximum generated code amount in each uncoded frame. An encoding method of a featured image signal. 4. The image signal encoding method according to claim 1, wherein when the priority order for performing the process of suppressing the generated code amount is determined for each image frame, the priority order is determined based on a reference relationship between the image frames. Encoding a video signal. 5. The encoding method for an image signal according to claim 4 , wherein when determining the priority order for performing the process of suppressing the generated code amount for each image frame, a bidirectional predictive encoded frame or a unidirectional predictive encoded frame is determined. A higher priority order in the order of intra-frame coding. 6. The program for executing the processing procedure in the coding method for an image signal according to any one of claims 1 to 5 computer, that the computer is recorded in a recording medium readable A recording medium on which a characteristic image signal encoding method is recorded.