JP2005323158A - Video- and audio-signal encoding device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a video- and audio-signal encoding device capable of generating encoded data having higher picture quality and higher sound quality than data encoded with encoding parameters set by a user or effectively using a recording medium having small recording capacity. <P>SOLUTION: The video- and audio-signal encoding device generates fixed-length packets from an encoded signal generated by encoding a video input signal and an audio input signal by a given encoding system. The video- and audio-signal encoding device finds a stuffing length needed for fixed-length packeting and a predictive code length when encoding is performed by the encoding system by using an encoding parameter group needed for the encoding, and changes at least one encoding parameter of the encoding parameter group for encoding to match a 2nd predictive encoding length different from the 1st predictive encoding length by taking the stuffing length into consideration for the 1st predictive encoding length. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a video / audio signal encoding apparatus, for example, an invention related to a method of packetizing and multiplexing a video signal and an audio signal.

  MPEG is a common system for encoding video signals and audio signals. Among them, in MPEG2 used in broadcasting or the like, for example, as described in Patent Document 1 below, a video encoded signal (hereinafter also referred to as a video elementary stream) output from a video signal encoding circuit. An audio signal encoded signal (hereinafter also referred to as an audio elementary stream) output from the audio signal encoding circuit is multiplexed with TS packets called a transport stream.

  Here, the TS packet will be described. A TS packet is composed of a fixed-length packet of 188 bytes, and a packet called a PES packet with a separator header added to each elementary stream obtained by encoding a video signal and an audio signal (hereinafter referred to as an audio PES packet, a video PES). Called a packet). One PES packet describes the encoding information and bit length of each elementary stream. Each PES packet is divided into 184 bytes, generates a TS packet with a 4-byte header added, and is multiplexed with a system packet to form a stream.

  Here, audio multiplexing is taken as an example. MPEG1 layer 2 (commonly referred to as MP2) is a typical encoding method for audio recorded on digital video cameras and DVDs.

  This MP2 is a method of encoding 1152 samples as a unit of one frame, in other words, it is equivalent to forming an elementary stream using 1152 samples as a processing unit of encoding. For this reason, it is known that a technique for generating a PES packet in units of one frame is appropriate in consideration of random accessibility and ease of handling time stamp management.

In this way, an audio PES packet with encoding information, bit length, time information, etc. added to an audio elementary stream encoded in units of one audio frame is generated, and the audio PES packet has a fixed length of 188 bytes. Audio TS packets are generated and multiplexed with audio TS packets and system TS packets.
JP 2001-44957 A

  However, the method for generating TS packets from PES packets in units of one frame as described above causes the following problems.

In MP2, the unit of one frame is 1152 samples, and a coding bit rate of 256 kbps and a sampling frequency of 48 kHz are generally used. At this time, the bit length of the audio elementary stream generated in units of one frame is
Number of samples per frame * (bit rate / sampling frequency)
Therefore, the audio elementary stream under the above conditions is calculated as 768 bytes.

  An audio PES packet is formed by adding a PES header of about 20 bytes to the audio elementary stream. At this time, the packet length of the audio PES packet is about 790 bytes.

  When a 188-byte TS packet is generated from the audio PES packet, five TS packets are generated. Of these, four TS packets are composed only of audio data, but it is necessary to pack bits with stuffing data of about 130 bytes in order to make the last one TS packet a fixed-length packet of 188 bytes. This stuffing data is so-called useless data, and is not efficient in terms of bit usage. Also, if the recording medium has a large capacity such as a magnetic tape, the above-mentioned stuffing byte data is of a size that does not matter so much, but if the recording medium has a small capacity such as an optical disk or a memory card, It is hard to say that the capacity that was allocated is being used efficiently.

  The present invention has been made in view of such a problem, and can generate encoded data with higher image quality and higher sound quality than data encoded with an encoding parameter set by a user, or has a recording capacity. An object of the present invention is to provide a video / audio signal encoding apparatus capable of effectively using a small number of recording media.

  In order to solve the above problems, the present invention has the following configuration.

(1) In a video / audio signal encoding apparatus that generates a fixed-length packet from an encoded signal obtained by encoding a video input signal and an audio input signal by a predetermined encoding method,
Using the encoding parameter group required for the encoding, a first prediction code length when encoded by the predetermined encoding method and a stuffing length required for the fixed-length packetization are obtained,
The video / audio signal code, wherein the code length is encoded by changing at least one encoding parameter of the encoding parameter group so as to conform to a second prediction code length in consideration of the stuffing length. Device.

  (2) The video / audio signal encoding device according to (1), wherein the second prediction code length is a code length obtained by adding the stuffing length to the first prediction code length.

  (3) The second predictive code length is a code length obtained by subtracting a code length obtained by subtracting the stuffing length from a packet length of the fixed-length packet from the first code length. The video / audio signal encoding device according to claim).

  (4) The method described in (2) and the method described in (3) for calculating the second prediction code length can be arbitrarily switched by a user of the recording apparatus. The video / audio signal encoding device according to any one of (1) to (3).

  (5) The video / audio signal code according to (1), wherein the coding parameter group is not changed when the stuffing length does not exist and when the stuffing length is less than a predetermined bit length. Device.

  (6) One of the encoding parameter groups is a bit rate for encoding the video input signal and the audio input signal by the predetermined encoding method. Video / audio signal encoding apparatus.

  (7) The video / audio signal code according to (1), wherein one of the encoding parameter groups is a sampling frequency when the audio input signal is encoded by the predetermined encoding method. Device.

  (8) One of the encoding parameter groups is a predetermined number of samples when the audio input signal is encoded by the predetermined encoding method. Signal encoding device.

  (9) The video / audio signal encoding device according to (1) or (8), wherein the predetermined number of samples is uniquely determined according to the determination of the encoding method.

  (10) The predetermined number of samples cannot be changed when changing at least one encoding parameter in the encoding parameter group. (1), (6) to (9) The video / audio signal encoding device according to claim 1.

  (11) When changing at least one encoding parameter in the encoding parameter group, a condition for changing can be selected by a user of the video / audio signal encoding device (1) The video / audio signal encoding device according to any one of (6) to (9).

  (12) The video / audio signal encoding apparatus according to (1), wherein the fixed-length packet is an MPEG2 transport stream packet.

  The present invention encodes the coding parameters given by the user of the coding apparatus so that stuffing byte data necessary for fixed-length packetization can be used as effective data such as a video signal and an audio signal. Since the encoding parameters can be changed, the encoded data having higher image quality and higher sound quality than the data encoded with the encoding parameters set by the user of the present encoding apparatus is created, or the recording medium having a smaller recording capacity Can be used effectively.

  Embodiments of the present invention will be described below with reference to the accompanying drawings.

  A first embodiment will be described with reference to FIGS. 1, 2, 3, and 5. FIG. 1 is an overall system configuration diagram according to the first embodiment, FIG. 2 is a flowchart for explaining the present embodiment, FIG. 3 is a circuit diagram of FIG. 2, and FIG. 5 is an encoding parameter changing process. Is illustrated.

  The first embodiment shows a preferred embodiment when an audio signal is encoded into a TS packet by MPEG1 layer 2 (MP2) which is a format adopted in the current digital video camera.

  In FIG. 1, reference numeral 101 denotes a video input signal from a CCD or a line input terminal (not shown), and 102 denotes a video signal encoding circuit that compresses and encodes the video input signal 101. The video signal encoding circuit 102 may be a high-efficiency encoding method such as MPEG2, but details of the video signal encoding circuit are not shown in this figure. Reference numeral 103 denotes a video encoded signal encoded by the video signal encoding circuit 102, which is called a video elementary stream.

  Reference numeral 104 denotes an audio input signal from a microphone or a line input terminal (not shown), and reference numeral 105 denotes an audio processor that compresses and encodes the audio input signal 104. Regardless of the type of CPU, DSP, etc., the processor 105 is not shown in detail in the processor. The function of the processor 105 can be realized by an HW circuit. Reference numeral 107 denotes an audio encoded signal encoded by the audio processor 105, which is called an audio elementary stream.

  Reference numeral 106 denotes a circuit for changing an encoding parameter when the audio signal is encoded by the audio processor 105, and details will be described later.

  110 is a multiplexing circuit that multiplexes each of the video elementary stream 103 and the audio elementary stream 107 into TS packets, and 108 is a TS stream that is multiplexed by the multiplexing circuit 107, and is recorded on the recording medium 109. Is done.

  The encoding parameter changing circuit 106 will be described with reference to FIGS. The audio processor 105 performs encoding based on the encoding parameter originally given to the recording apparatus. The coding parameters here indicate the sampling frequency for converting the analog input signal into discrete data, the bit rate that affects the sound quality of the audio encoded signal, and the number of samples in one frame that is the unit of audio encoding. It can be set by the user who uses the device. In the MP2 format used in the current digital video camera for performing MPEG recording, one frame is 1152 samples, and the coding bit rate is generally 256 kbps and the sampling frequency is 48 kHz.

  At this time, when an audio elementary stream is generated in units of one frame and converted into TS packets, packets must be packed with stuffing bytes as described above. Therefore, the bit rate changing circuit 106 is used with reference to FIGS. The operation of will be described. The declaration in the text matches the subscript in the figure.

  In S201, the recording apparatus user gives an encoding format to the recording apparatus, whereby the number of samples 302 (FRAME) of one audio frame, which is an encoding parameter, is determined. Further, the bit rate 301 (BITRATE) and the sampling frequency 303 (SPF) may be either a form that can be determined by the user of the recording apparatus or a form in which the recording apparatus is determined to an arbitrary value in the initial state.

In S202, the bit length 306 (AU_SIZE) of the audio elementary stream, which is an output obtained by encoding the audio input signal of one frame, is calculated by the multiplier 305 with BITRATE 301, FRAME 302, and SPF 303 as inputs. The bit length calculation formula is
AU_SIZE = FRAME * (BITRATE / SPF)
It is. At this time, the sampling frequency 303 is converted to a reciprocal value by the reciprocator 304.

  In S203, the audio PES packet length (PES_SIZE) 309 is calculated by the adder 308 with the bit length 306 of the audio elementary stream and the bit length (PESH) 307 of the PES header as inputs.

In S204, the arithmetic unit 311 calculates the TS packet number (PACK_NUM) 312 and stuffing byte length (STAFF) 313 when TS packets are converted into the audio PES packet length 309 and the TS packet length (PACK_SIZE) 310 as inputs. The formula for calculating the number of TS packets and the stuffing byte length is as follows:
PACK_NUM = (PES_SIZE / PACK_SIZE) +1
STAFF = PACK_SIZE-PES_SIZE% PACK_SIZE
It is. Here,% indicates the remainder.

  In S205, the updated audio elementary stream length (AU_NEW) 315, which is the bit length when bits for stuffing bytes are used as audio data, is input using the stuffing byte length 313 and the audio elementary stream length 306 as inputs Calculate at 314.

  In S206, the updated bit rate (NEW_BITRATE) 318 with respect to the updated audio elementary stream length 315 is input to the multiplier 317 with the updated audio elementary stream length 315, the sampling frequency 303, and the number of samples 302 of one frame as inputs. Calculate with

  Note that the number of samples 302 in one frame is converted into a reciprocal value by the reciprocator 316.

  The updated bit rate is input to the audio processor 105, and audio encoding processing is performed using the updated bit rate.

  The above processing is shown in FIG. In FIG. 5, (a) is a code length (AU_SIZE) 306 generated when encoding is performed with the parameters set by the user of this encoding apparatus, and a packet header length (PESH) 307 is added to AU_SIZE 306. If the code length is converted into a fixed packet, a stuffing packet exists as shown in FIG. On the other hand, as shown in FIG. 5C, the sum NEW_AU_SIZE 315 of AU_SIZE 306 and PESH 307 resets the code length suitable for the packet length (PACK_SIZE) 310 of the fixed packet, and the changed code length is based on 315. The encoding parameter changing circuit 106 changes the bit rate and performs encoding in the processor 105.

  Moreover, it is possible to change the above processing as follows.

  The above method (FIG. 5C) was based on the idea of using stuffing byte data as meaningful data such as audio data. However, when the capacity of the recording medium is small, the packet that must originally be stuffed It is also possible to use a method that does not generate That is, as shown in FIG. 5D, a new audio elementary stream length is calculated so that a packet in which stuffing exists at the bit rate set by the user of this encoding apparatus is discarded, and the audio element is calculated. Coding is performed by updating the bit rate according to the length of the mental stream.

  By realizing the above processing, recording is performed at a bit rate lower than the bit rate set by the user of the recording apparatus, but the recording capacity can be used effectively.

  If there is no stuffing byte in the encoding parameter set by the user of this encoding device, or if the user of this encoding device does not want to update the encoding parameter as described above, the user setting It is possible to record under the specified conditions.

  A second embodiment will be described with reference to FIGS. FIG. 4 is an overall system configuration diagram in the second embodiment, and FIG. 6 is a diagram illustrating target code amount change processing.

  The second embodiment shows a preferred embodiment relating to code amount control (quantization step control) when a video signal is encoded with MPEG2, which is a format employed in current digital video cameras and digital broadcasting.

  When encoding a video signal in MPEG2, it is necessary to control the quantization step for each macroblock so that an output buffer called a VBV buffer does not overflow and underflow. Since the algorithm is described in TM5 (Test Model Editing Committee: “Test Model 5”; ISO / IEC JTC / SC292 / WG11 / N0400), the details are omitted. In the quantization step control circuit, if the target code amount given from the free capacity of the VBV buffer can be related to the packet length of the fixed-length packet, a stream with less stuffing can be generated as the output stream.

  Since FIG. 4 is a simplified MPEG2 encoding circuit with the present invention block added, a detailed description of MPEG encoding will be omitted. Reference numeral 401 denotes a video input signal to the encoding apparatus and is stored in the frame buffer 402. When the picture type is an I picture, the data in the frame buffer 402 is sent to the DCT circuit 406 as data for each block. The DCT circuit 406 performs vertical and horizontal two-dimensional discrete cosine transform processing for each block. As a result, a signal in the time domain is converted into a signal in the frequency domain. The DCT coefficient from the DCT circuit 406 is sent to the quantizer 407 and quantized with a predetermined quantization step width. Thereafter, the coefficients rearranged by the zigzag scan (not shown) are sent to a variable length coding (VLC) unit 410 for Huffman coding. The encoded stream obtained as a result is temporarily stored in the output buffer 411 and then recorded on the recording medium 412 at a constant bit rate.

  When the picture type is a P picture or B picture, inter-frame coding is performed, and thus the encoded signal from the quantizer 407 is once decoded (local decoded) by the inverse quantizer 409 and the inverse DCT circuit 408. Then, the motion detection circuit 403 and the motion compensation circuit 404 perform motion detection and motion compensation with the input video signal, the subtractor 405 calculates a difference value with the input video signal, and DCT converts the difference value. The subsequent processing procedure is the same as that for the I picture.

  Here, in order to prevent the output buffer 411 from overflowing or underflowing, the VLC circuit 410 transmits the code amount 413 of the picture that is currently encoded to the target code amount setting circuit 414 and next. The target code amount of the same type of picture to be encoded is determined.

  The target code amount setting circuit 414 calculates the target code amount 415 of the next picture to be encoded. This target code amount 415 is sent to the target code amount changing circuit 416, and the packet length of the fixed-length packet and the packet length The target code amount is changed in consideration of the header length 417. The target code amount changing circuit 416 will be described later in detail. The changed target code amount 418 is quantized by the quantizer 407 on the DCT coefficient converted by the DCT circuit 406 by the quantization step control circuit 419 determining the quantization step.

  Here, the target code amount changing circuit 416 will be described with reference to FIG. 6A shows the target code amount 415 calculated from the target code amount setting circuit 414 and the header length 417 of the packet. If the code length obtained by adding the packet header length 417 to the target code amount 415 is converted into a fixed packet, there will be a stuffed packet as shown in FIG. On the other hand, the target code amount changing circuit 416 is such that the sum of the target code amount 415 and the header length 417 is an integral multiple of the packet length 420 of the fixed packet, as shown in FIG. In this way, the target code amount is reset. The changed target code amount is 418, and the quantization step control circuit 419 determines the quantization step based on the value.

  At this time, it is necessary to pay attention to the overflow or underflow of the target code amount to be changed. That is, the target code amount setting circuit 414 must calculate a target code amount that does not cause the buffer to fail even when there is an increase or decrease of one packet.

  By performing the above processing, it is possible to perform control so as to reduce the stuffing bytes as much as possible when encoding according to the target code amount.

The block diagram which shows the structure of the encoding apparatus in 1st Embodiment. The flowchart which shows the process of the encoding parameter change circuit in 1st Embodiment. Circuit diagram of coding parameter changing circuit in the first embodiment The block diagram which shows the structure of the encoding apparatus in 2nd Embodiment. The figure which shows the mode of the voice data with code amount control The figure which shows the mode of the voice data with code amount control

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Video input signal 102 Video signal encoding circuit 103 Video elementary stream 104 Audio input signal 105 Audio processor 106 Coding parameter change circuit 107 Audio elementary stream 108 TS stream 109 Recording medium 110 Multiplexing circuit 301 Bit rate 302 Audio 1 frame Sample number 303 Sampling frequency 304 Inverse number 305 Multiplier 306 Audio elementary stream bit length 307 PES header bit length 308 Adder 309 Audio PES packet length 310 TS packet length 311 Calculator 312 TS packet number 313 Stuffing byte length 314 Adder 315 Updated audio elementary stream length 316 Inverse number 317 Multiplier 318 Bit rate 401 after update 401 Video input signal 402 Frame buffer 403 Motion detection circuit 404 Motion compensation circuit 405 Subtractor 406 DCT circuit 407 Quantizer 408 Inverse DCT circuit 409 Inverse quantizer 410 Variable length encoder 411 Output buffer 412 Recording Medium 413 Picture code amount 414 Target code amount setting circuit 415 Picture target code amount 416 Target code amount change circuit 417 Packet header length 418 Changed target code amount 419 Quantization step control circuit 420 Fixed packet header length

Claims (12)

In a video / audio signal encoding device that generates a fixed-length packet from an encoded signal obtained by encoding a video input signal and an audio input signal by a predetermined encoding method,
Using the encoding parameter group required for the encoding, a stuffing length required for the fixed-length packetization and a first prediction code length when encoded by the encoding method are obtained,
At least one coding parameter of the coding parameter group is adapted so that the first prediction code length is adapted to a second prediction code length different from the first prediction coding length considering the stuffing length. A video / audio signal encoding apparatus, wherein the encoding is performed by changing.   The video / audio signal encoding apparatus according to claim 1, wherein the second prediction code length is a code length obtained by adding the stuffing length to the first prediction code length.   2. The code length according to claim 1, wherein the second predicted code length is a code length obtained by subtracting a code length obtained by subtracting the stuffing length from a packet length of the fixed-length packet from the first code length. Video / audio encoding device.   The method according to claim 2 and the method according to claim 3 for calculating the second prediction code length can be arbitrarily switched by a user of the recording apparatus. 4. The video / audio signal encoding device according to any one of 3 above.   The video / audio signal encoding apparatus according to claim 1, wherein the encoding parameter group is not changed when the stuffing length does not exist and when the stuffing length is smaller than a predetermined bit length.   The video / audio according to claim 1, wherein one of the encoding parameter groups is a bit rate for encoding the video input signal and the audio input signal by the predetermined encoding method. Signal encoding device.   The video / audio encoding apparatus according to claim 1, wherein one of the encoding parameter groups is a sampling frequency when the audio input signal is encoded by the predetermined encoding method.   The video / audio encoding apparatus according to claim 1, wherein one of the encoding parameter groups is a predetermined number of samples when the audio input signal is encoded by the predetermined encoding method. .   The video / audio signal encoding device according to claim 1 or 8, wherein the predetermined number of samples is uniquely determined according to the determination of the encoding method.   The predetermined number of samples cannot be changed when changing at least one encoding parameter in the encoding parameter group. 10. The video / audio signal encoding device according to Item.   The change condition can be selected by a user of the video / audio signal encoding apparatus when changing at least one encoding parameter in the encoding parameter group. The video / audio signal encoding device according to claim 9.   The video / audio signal encoding apparatus according to claim 1, wherein the fixed-length packet is an MPEG2 transport stream packet.