CN1783302A - Treating play back bit stream - Google Patents

Abstract

The equipment is used to reproduce bit stream signal from CD and the bit stream is so controlled that only the required bit stream data are coupled for MPEG decoding. The converter is re-located to obtain required bit stream data before completing MPEG image decoding. The bit stream data are read out before buffer storing, so that required data are selected and stored while refusing unwanted data. The buffer memory is re-allocated for trick playing operation and randomly accessed for image selection in trick playing. The MPEG image decoding and storing is controlled, so as to complete frame decoding in one field period and store and read out decoded image simultaneously within a field period.

Description

Replay Bitstream Processing

This case is a divisional application of an invention patent application with an application date of May 6, 1998, an application number of 98814153.1, and an invention title of "Processing of Replay Bit Stream".

technical field

This invention relates to the reproduction of digitally encoded signals from media, and more particularly to the identification and processing of playback data in a multiplexed format.

Background technique

The introduction of optical discs, for example using digitally compressed audio and video signals using the MPEG compression protocol, has provided consumers with sound and picture quality essentially indistinguishable from the original material. However, consumer users expect such digital video discs or DVDs to provide features similar to those of their analog video recorders or VCRs. For example, a VCR can reproduce in the forward or reverse direction at a speed different from the recording speed. This non-standard speed playback feature is also known as trick play mode. MPEG-encoded video signals cannot readily provide trick-play features due to the hierarchical nature of compression that divides pictures into groups with different levels of compression. These groups are called groups of pictures or GOPs and require sequential decoding. A detailed description of the MPEG 2 standard is published in ISO/IEC standard 13818-2. Simply put, an MPEG 2 signal stream can include three picture types with different levels of content compression. Intra-coded frames, or I-frames, have the least compression of the three types and can be coded without reference to any other frames. Predicted or P-frames are compressed with reference to previous I- or P-frames and achieve a higher level of compression than intra-coded frames. A third type of MPEG frame, called a bidirectionally coded frame or B-frame, can be compressed based on a prediction of either a previous frame or a subsequent frame. Bidirectionally encoded frames have the highest level of compression. The three types of MPEG frames are arranged in groups of pictures, or GOPs. A GOP may, for example, contain 12 frames arranged as illustrated in FIG. 1A. Since only intra-coded frames can be decoded without reference to any other frames, each GOP can only be decoded after the I-frame is decoded. The first predicted frame or P-frame can be decoded and stored as a modification of the stored previous I-frame. Subsequent P frames can be predicted from stored previous P frames. The prediction of the P frame is indicated by the curved solid arrow in Fig. 1A. Finally, bidirectionally coded frames or B frames can be decoded by means of predictions of previous and/or subsequent frames, such as stored I and P frames. Decoding of B frames from prediction of adjacent stored frames is depicted in FIG. 1A by curved dashed arrows.

Patent application EP-A-0 696 798 discloses a method and device for recording data, a data medium and a method and device for reproducing data. Application EP-A-0 696 798 teaches the recording of MPEG signals on various types of disc media, and moreover discloses a recording format with additional information recorded separately from the data as subcodes in each sector of the recording medium. These subcodes provide information related to the payload data per sector and can be used to control data reproduction during playback. In US Patent 5,535,008, "skip mode" reproduction of MPEG recorded data is disclosed, for example using a CD-ROM. US Patent No. 5,535,008 teaches reproduction of a plurality of data set at predetermined intervals, and thus data to be reproduced next can be found by "subtracting a first fixed value from an integer multiple of the predetermined interval". US Patent 5,535,008 uses the average distance between I-frames to cause the transformer to move back and forth rapidly or to reproduce rapidly in reverse. In EPO application A-737 975 a method of MPEG recording for optical discs is disclosed. The recorded format includes a management area and a program area in which program data has a hierarchical structure. The disclosed format is somewhat similar to that employed by Digital Versatile Discs or DVDs. In US Patent 5,543,925 a digital playback device is disclosed for reproducing digitally processed photographic images. The digitized images are written on a compact disc and displayed on the screen according to stored data representing pre-recorded sequences or user-defined displays. US Patent 5,543,925 teaches that for each stored digitized image, the image file contains multiple sub-files that define the same scanned image at different resolutions. It points out that these multi-resolution versions of the same image advantageously reduce the latency of image display. High speed playback of digitally encoded images is disclosed in EPO application 0651391. EPO application 0651 391 discloses inter alia the use of two frame memories and three frame memories for storing multiple GOPs during "frame return reproduction". The output signal selection alternates between decoded image memories. Memory selection and output picture retention times are responsive to decode picture availability and time to retrieve the next desired picture.

The hierarchical nature of coded frames comprising MPEG groups of pictures requires that the I and P frames of each GOP be decoded in the forward direction. Thus, the reverse mode feature can be provided by effectively jumping back to an earlier, or previous, I frame and then decoding the GOP in the forward direction. The decoded frames are stored in frame buffer memory for later reverse readout to obtain the desired reverse program sequence. Figure 1B illustrates forward playback at normal speed and at a time prior to time t0, reverse, triple speed mode in trick mode selected. The trick-play mode starts at time t0, where I-frame I(25) is decoded and displayed. The next frame to be decoded is I frame I(13), so the converter is repositioned, as indicated by arrow J1, to acquire frame I(13). After recovering and decoding I frame I(13), the transformer traces as indicated by arrow J2 to obtain and decode frame P(16). This process is repeated as indicated by arrows J3, J4. After frame P(22) is captured and decoded, the transformer moves as depicted by arrow Jn to recover frame I(1). Decoding and displaying I, P and possibly B frames is required to smoothly depict scene motion. The jump-and-play process is repeated for the previous GOP, so forwarding is halted and the recording is rewinded while the program material is smoothly rendered on the video output in reverse sequence.

Providing a visually smooth reproduction during trick mode reproduction requires timely retrieval of the disc and access of specific images from memory. Although each digital disc is encoded with navigation data within each video object unit that provides entry points for images, they are limited in number and may inherently confuse image motion in time. To obtain a temporally smooth trick mode reproduction in forward and reverse directions at multiple speeds requires accessing and decoding all coded pictures. While this performance can be achieved at the expense of memory capacity, the bitstream analysis and buffer memory selection provide an opportunity to improve trick mode rendering through efficient memory utilization.

Contents of the invention

A device reproduces a bit stream signal from an optical disc. The bitstream is controlled to ensure that only required bitstream data is coupled for MPEG decoding. The converter is repositioned to obtain the required bitstream data before completing the previous MPEG image decoding. Bit stream data is read out before buffer storage to select data to be stored and to reject unnecessary data. Buffer storage is reallocated for trick-play operations and accessed randomly to facilitate trick-play picture selection. Controls MPEG picture decoding and storage for frame decoding within a field period. The decoded image is stored and read out substantially simultaneously within a field period.

In one arrangement of the invention, an apparatus reproduces a digitally encoded signal from a medium. The device includes: a converter for converting a digitally encoded signal and generating a bit stream therefrom; a processor coupled to receive the bit stream for controllably processing the bit stream; a memory coupled to the processor for storing the processing bit stream information; a controller, coupled to control the memory and a processor, for controlling identification of information within the bit stream, wherein the controller controls the processor to identify a particular sector type in the bit stream, and according to the specific sector For sector identification, the controller directs the memory to store the identified specific sector.

In an inventive method, a start code is obtained from a plurality of start codes in a data stream arranged in a plurality of sectors switched by the digital disc device during reproduction. The method includes the step of searching the data stream to locate a particular sector type of the plurality of sectors. Search for that particular sector type to locate a start code from multiple start codes. Test the start code to determine if it is incomplete. The data stream is searched to locate a second sector of a particular sector type among the plurality of sectors. The second sector of the particular sector type is searched to locate the second start code from the start code type of the plurality of start codes. Determine if the second start code is the remainder of the incomplete start code. Combine the incomplete start code value and the remaining start code value to form the complete start code.

Another inventive apparatus for reproducing a digitally encoded signal from a medium comprises: a converter for converting the digitally encoded signal and generating a bit stream therefrom; a first memory coupled to the converter for storing the bit stream; a second memory of data, controllably coupled from the first memory; a controller, coupled to control the first and second memory, for controlling identification of information within the bit stream, wherein the controller controls the first memory to read from a particular sector The sector address outputs the bit stream therein, and the controller controls the second memory to store the first part of the bit stream output from the specific sector address.

Another inventive apparatus for reproducing a digitally encoded signal from a medium comprises: a converter for converting the digitally encoded signal and generating a bit stream therefrom; a memory coupled to the converter for storing the bit stream; a processor coupled to the memory , used to process the stored bitstream to identify the MPEG start codes contained within it. The processor searches the stored bitstream to identify an MPEG start code in the stored bitstream, and based on the MPEG start code identification, the processor indicates the identification and stores the sector address in the identified MPEG start code.

Another method for decoding and displaying images in a device for reproducing digital discs. The method includes the step of converting a digitally encoded signal from an optical disc. The digitally encoded signal is stored in a first memory. The digitally encoded signal is decoded to generate an image. The image is stored in a second memory. The image from the second memory is coupled for display. Controlling storage and coupling of the second memory for display occurs substantially simultaneously.

In a further inventive method, a field of an image is stored in a first field of a second memory and a previous image from the second field of said second memory is coupled for display.

In another inventive method, a field of an image is stored in a first field of said second memory and a previous image from the first field of the second memory is coupled for display. The storage and coupling of the first field of the second memory for display is controlled to occur sequentially within a field period.

In another configuration of the invention, an apparatus for reproducing digitally encoded signals from an optical disc medium includes: a converter for converting the digitally encoded signals; a memory coupled to the converter for storing the digitally encoded signals; a decoder , to decode an image therefrom according to the digitally encoded signal; the controller is used for the memory, wherein, in the first operation mode, the controller controls the memory to read out the stored digitally encoded signal from the memory according to the first sequence, in the In the second operation mode, the controller controls the memory to read out the stored digital coded signal from the memory in a second sequence.

In another configuration of the invention, an apparatus for reproducing a digitally encoded signal from an optical disc medium and comprising: a source of a bitstream representing the digitally encoded signal; a processor, coupled to the bitstream, for processing the bitstream to extract the Representing at least first and second data types; a memory, controllably coupled to the processor to store one of the first and second data types; a controller, coupled to control allocation of the memory, wherein in the first rendering mode , the controller allocates the memory to store the first data type, and in the second reproduction mode, the controller allocates the memory to store the second data type.

In another configuration of the invention, an apparatus for reproducing a digitally encoded signal from an optical disc medium includes: a converter for converting the digitally encoded signal; a processor coupled to receive the digitally encoded signal for processing and extracting the digitally encoded signal from generating an image; a memory coupled to the processor for storing the image; a controller coupled to control the memory and the processor, wherein in a first mode the image is stored in the memory and in a second mode the image is stored subsampled and stored in memory.

In another configuration of the invention, unnecessary processing of unneeded sector data is avoided. A method for controlling reproduction of data in sectors by an optical disc player using optical readout includes the step of switching sector groups including sectors requiring processing and sectors not requiring processing. The desired sectors other than the unnecessary sectors are provided to the data processor for processing, and the desired sectors of data are processed to extract data representing video information therein.

A digital disc player for reproducing sectors containing MPEG compatible data, comprising: a converter for converting groups of sectors from an optical disc including sectors having requests for MPEG compatible data required for processing area, and an unrequested sector having MPEG compatible data not required for processing; a first data processor, coupled to the converter, for processing and coupling except for the sector having MPEG compatible data not required for processing said requested sector with MPEG compatible data other than unrequested sectors; and a second processor, coupled to said first processor, for receiving said requested sector of data and extracting The requested MPEG compatible data.

In a method of the present invention, MPEG compatible data reproduced in sectors is controlled, the method comprising the steps of: switching groups of sectors comprising sectors having requests for MPEG compatible data required for processing, and having unrequested sectors of MPEG compatible data not required for processing; transmitting said requested sectors other than said unrequested sectors to a data processor for processing; and processing said requested sectors to extract said requested MPEG compatible data representing video information.

In another configuration of the invention, delays in the converted bitstream path are substantially eliminated from the process of controlling the transducer position. During playback of a digital optical disc device, a method includes the step of receiving first and second changer addresses for controlling the position of the changer in dependence on the playback address. The replayed address is compared with the first translator address to detect equality between them. Depending on the detected equality, the transducer is moved to a new position determined by the address of the second transducer.

Description of drawings

Figure 1A depicts an MPEG 2 group of pictures.

Figure 1B illustrates a group of pictures recorded during playback and triple speed reverse trick play.

Figure 2 is a block diagram of an exemplary digital disc player including the arrangement of the present invention.

Figure 3 shows parts of Figure 2 in more detail and depicts various inventive arrangements.

Fig. 4 shows the digital disc player of Fig. 2, including other advantageous configurations relative to Fig. 2 .

Figures 5A and 5B depict an exemplary bitstream prior to track buffering.

Figures 5C-5D depict exemplary data in the buffer memory.

Fig. 6 is a flowchart illustrating the arrangement of the present invention for recovery across start codes allocated on sector boundaries.

Figure 7 is a diagram illustrating the inventive sequence for reverse trick play mode at 3x playback speed (3x).

Detailed ways

Figure 2 depicts an exemplary block diagram of a digital video disc player. Block 10 describes the deck mechanism, which can receive the digital recording disc 14 rotated by the motor 12 . The digital signal is recorded on the optical disc 14 as a spiral track containing pits in response to each signal data bit having an 8/16 modulation code to determine the length of each pit. Recordings on the optical disc 14 are read by a pickup device 15 which collects reflected laser illumination. The reflected laser light is collected by a photodetector or optical pickup. Imaging means such as lenses or mirrors forming part of the pick-up means 15 are servo-controlled and driven by the motor 11 to follow the recorded track. Different parts of the recording can be accessed by quickly repositioning the imaging device. The servo-controlled motors 11 and 12 are driven by an integrated circuit driver amplifier 20 . The pick-up device 15 is coupled to an optical pre-amplifier block 30, which includes driver circuitry for the laser light emitting device, the pre-amplifier providing amplification and equalization of the reflected signal output from the optical pick-up device. The amplified and equalized playback signal from the optical preamplifier 30 is connected to the channel processor block 40 where the playback signal is used to synchronize the phase locked loop which is utilized to demodulate the 8:16 modulation used in the recording .

The MPEG coded bitstream is coded for error detection and correction by means of Reed Solomon product coding applied in blocks of 16 sectors, where each sector contains 2048 bytes of payload data. Thus, the playback data stream after 8:16 demodulation is de-interleaved or reverse-shifted (de-shuffled) and error-corrected by means of the Reed Solomon product correction implemented in the ECC buffer memories 45 and 46 of FIG. Each buffer stores 16 sectors of the playback data stream arranged as an array to facilitate deinterleaving and to allow the required row and column product processing. The cascaded ECC buffer memories introduce approximately (2*16*1.4) milliseconds of delay to the reproduced sequential bitstream, where 2 denotes the ECC buffer pair, 16 denotes the sector to which the correction is applied, and 1.4 milliseconds at 1x rotational speed sector cycle. Thus, the rendered sequential bitstream is delayed by a minimum of approximately 45 milliseconds.

The error correction signal bitstream 41 is coupled to the bitstream or mechanical/track buffer memory 60A via the link processor. The track buffer comprises a DRAM memory type and is used to store playback data volumes so that lost data during repositioning of the changer or pick-up 15 will not produce any visible defects when decoded. In this way, the final output image stream will appear continuous or seamless to the viewer. Bitstream buffer memory 60A is part of an exemplary 16 megabit DRAM memory. Another exemplary 16 megabit SDRAM memory block can be partitioned to provide frame buffers 60C and 60D which provide storage for at least two decoded image frames, the compressed video bitstream being stored in buffer 60B prior to decoding , the audio bit stream is stored in buffer 60E and other data are stored in 60F, 60G, 60H. Channel processor 40 also includes timing control circuitry that controls writing over link 505 to bitstream buffer 60A. Due to changes in playback track addresses, e.g. due to user-defined playback video content such as "Directors cuts", basic (parental) guide selections, or even user-selectable alternate shot angles, Data can be written to the bitstream buffer intermittently. To facilitate more rapid access and retrieval of the recorded signal, the optical disc 14 can be rotated at an increased speed resulting in a higher bit rate converted bit stream and possibly provided intermittently.

As already described, the recorded data stream is arranged in ECC blocks of 16 sectors. Each sector has a unique sector identification address protected by error correction bits processed by the ECC block 47 of FIG. 4 . However, because the sector addresses are short and sector specific, any delay to the sector address signal 42 generated by the error correction processing block 47 is negligible. Sector address signal 42 is coupled to provide position information to servo control integrated circuit 50 . Integrated circuit 50 provides drive and control signals for servo motors 11 and 12 . Motor 12 rotates optical disk 14 and provides servo-controlled rotation at various speeds. The optical pick-up device or changer 15 is positioned and servo-controlled by the motor 11 according to the sector address signal 42, and additionally, can be controlled according to the sector address request to quickly reposition or jump to another sector address, or a position on the surface of the disc. The position is conveyed by the I ² C control bus 514 and is coupled via the component 54 of FIG. 4 .

The digital video disc player is controlled by the central processing unit or CPU of block 500, component 510, which receives the reproduced bit stream and error flags from the channel IC 40 and provides control instructions to the servo IC 50. In addition, CPU 510 receives user control commands from user interface 90, and MPEG decoder control functions from MPEG decoder component 530 of block 500. System buffer memory 80 is addressed and provides data to CPU 510. For example, buffer memory 80 may include RAM and PROM memory cells. RAM may be used to store various data extracted from bitstream 41 by CPU 510, such data may include descrambling or decryption information, bitstream and framebuffer memory management data, and navigation data, for example. The PROM may, for example, contain a beneficial converter jump algorithm that facilitates trick mode operation at a selected one of forward or reverse speeds.

In FIG. 3, the MPEG-encoded bitstream is coupled to Link Processor 505, which may act as a hardware demultiplexer to separate the MPEG-encoded audio, video, and control information from the DVD-formatted bitstream. On the other hand, bit stream demultiplexing can be accomplished by direct memory access or software control of DMA by CPU 510 of FIG. 3 to buffer 60A. The encoded bitstream before or within track buffer 60A is searched by microcontroller 510 to locate and read titles and extract navigation data. A beneficial bitstream search is discussed below with reference to FIG. 6 .

Microcontroller 510 is coupled to the front end via ^I2C control bus signal 514 to control or request transducer repositioning for the next sector requested by the trick play sequence. Translator positioning can be controlled by a beneficial store sequence or jump play mode that is retrieved with reference to the replayed sector address and GOP sector address read from the navigation pack data contained in each video object unit or VOBU . An exemplary sector address and VOBU navigation pack is depicted in FIG. 5A. However, after converter repositioning, the sectors initially retrieved from the front end may be recognized by the example microcontroller 510 as not requested by the jump instruction. Thus, microcontroller 510 advantageously overwrites this unwanted data in track buffer 60A and ensures that only requested data is present in the buffer.

After identifying the sector address or header, microcontroller 510 controls direct memory access to buffer 60A, which effectively separates the MPEG data from other DVD format data stored in the buffer. In this way, video DMA 515 separates the compressed video bits coupled for storage in exemplary video bit buffer 60B. Similarly, compressed audio bits are read from buffer 60A and stored in audio buffer 60E. Sub-picture data is also retrieved by DMA from track buffer 60A and stored in buffer 60F.

The compressed video bitstream in video bit buffer 60B is searched through start code detector 520 to locate a picture or higher level start code. Detected start code signal 512 is coupled to microcontroller 510 which then communicates via signal 511 to MPEG decoder 530 to indicate the next picture type, equalizer setting and start decoding. A decoder status signal 513 is coupled back to the microcontroller 510 to indicate that decoding is complete and that the image data is available for display or storage. As will be described, the compressed video bit buffer 60B can be thought of as a FIFO or ring buffer where the stored bit stream is accessed sequentially for MPEG decoding, however, random access through the buffer 60B can be beneficial to facilitate trick mode operation.

The video bitstream within the MPEG decoder 530 is processed by a variable length decoder 531 which searches the bitstream to locate slice and macroblock start codes. Certain decoded pictures from each group of pictures are written to frame buffers 60C and 60D for subsequent use as predictors when other pictures are acquired or constructed, such as the P and B pictures of a GOP. Frame buffers 60C and 60D have a storage capacity of at least two video frames. The separated audio packets are stored in audio bit buffer 60E, which is read out and coupled for audio decoding at block 110 . After MPEG or AC3 audio decoding, a digital audio signal is produced, which is coupled to audio post-processor 130 for digital-to-analog conversion and produces various baseband audio signal outputs. The digital video output signal of the decoded block read from reference frame buffer 60C/D is converted by display buffer 580 to raster scan format. However, during trick mode operation, the source of the output signal may be field memory that is beneficially reconfigured from memory that was not used during trick mode operation. In this way, the block-to-raster scan conversion within the display buffer 580 can be advantageously controlled according to trick mode operation. The display buffer is coupled to encoder 590, which provides digital-to-analog signal conversion and produces a baseband video component and an encoded video signal.

The operation of the exemplary video player illustrated in FIG. 2 may be considered with reference to FIG. 1B, which illustrates a forward play and reverse trick play sequence. As noted above, the coding relationships that exist within each GOP require each GOP to be decoded starting with an I frame or picture in the forward direction. In this way, the reverse mode feature can be provided by effectively jumping back to switch the earlier or previous I-picture and then decoding the GOP in the forward direction. The decoded images are stored in frame buffer memory for subsequent readout in reverse order. However, sequences comprising B-pictures can further exploit the beneficial features that will be described. In FIG. 1B , assume that at some point before time t0 , eg, at I picture I(1), the exemplary video player assumes a forward play condition according to user command. Each GOP is decoded in the forward direction, as illustrated by the I, B, and P frames connected by arrowed lines in FIG. 1A. At a time before time t0, the reverse trick mode at three times playback speed is selected and started at time t0, where the I picture I (25) is decoded and displayed. As mentioned above, the next picture required for reverse trick-play decoding is the I picture I(13), so, as indicated by arrow J1, the converter is moved to obtain picture I(13). Then, signal recovery and decoding follow the playback sequence indicated in Fig. 1B, I (13) is obtained by arrow J1, P (16) by arrow J2, P (19) by arrow J3, P (22) by arrow J4 ) and obtained by the arrow Jn . . . As specifically required by each trick-play mode, the inserted B-picture shown in Figure 1B is converted, but may be discarded, for example by overwriting the buffer or disabled by the decoder. To avoid the previously described need for an additional reverse mode video buffer, various beneficial methods are used for the MPEG decoder, buffer memory control and allocation.

Determination of image data may be done in units of sectors referenced in the bit stream 41 or the track buffer 60A. However, because the MPEG picture start code is buried within the DVD data format and is not constrained to start concurrently with a sector boundary, the resulting location of the picture start code in units of sectors inevitably includes a previous, possibly non-video sector's part. Figure 5A shows an exemplary bitstream 41 portion comprising a video object unit containing audio video and sub-picture data sectors. Each sector contains 2048 payload bytes, with sector addresses shown shaded on sector boundaries. Video picture A is shown in FIG. 5B ending at sector 54 and immediately followed by the start code for video picture B. In FIG. However, the remainder of the video picture B start code is generated at sector 65, with intervening sectors 55-64 containing sub-picture and audio data. The determination or positioning of image data/video sectors in units of sectors is illustrated in Figure 5C, where the start code for the exemplary image A is shown in sector 2 and the start code for the next image B is shown in sector 9 generated. Equation 1 shows the image data position counted by sector, since image A starts at sector 2 and ends at sector 9, image A has a duration of 8 sectors. Figure 5C illustrates unwanted data fragmentation, where video data is described with reference to (video) sector numbers. However, such video sector numbers may be directly related to addresses or sector numbers in the reproduced bitstream. In FIG. 5C , the video bitstream is shown with the depicted exemplary picture A having a picture start code starting at byte 1000 of video sector 2 . Clearly, the first 999 bytes of sector 2 correspond to data from the previous image. More verbose processing may be used when image data is positioned in bytes. Byte-precise processing may require more sophisticated memory control than is required for sector-level precision. However, if byte-accurate processing is used, the entire image data can only be stored in the video bit buffer, thus these fragments are cleared and MPEG decoder 530 hangs are avoided. Byte-accurate picture determination for an exemplary picture A is shown in FIG. 5C , where the picture start code starts at byte 1000 of video sector 2 and the picture B start code starts at byte 500 of sector 9 . Therefore, by using Equation 2, the size of image A can be calculated to be 13,835 bytes. Thus, the byte-accurate image address allows the microprocessor 510 to point to the specific byte in the exemplary video bit buffer 60B from which the variable length decoder VLD 531 of FIG. 3 begins decoding.

If the picture data is specified in units of sectors, an MPEG decoder reading a picture from the video bit buffer must prevent being hung up due to discarded picture fragments occurring before or after the desired picture is decoded. Such picture fragmentation is depicted in the exemplary video bit buffer of Figure 5D, which shows multiple sectors containing P and B pictures, where unwanted data from previous or subsequent pictures is shaded diagonally Shows. Each Video Object Block Unit or VOBU includes navigation data identifying the ending sector address of the first I-picture and the last sector address of the two following reference or P-pictures of the first GOP of the VOBU. In addition, the navigation data includes sector addresses of I pictures in previous and subsequent VOBUs, so that only I pictures for trick mode can be easily provided. However, problems caused by image fragmentation can be avoided if the end byte of the desired image can be identified. Microprocessor 510/A, eg of type ST20, is advantageously configured as a hardware search engine which searches the data stored in the track buffer to locate the end byte of an I-picture stored in the end sector of buffer 60A. Thus, by identifying an I picture, it alone can be loaded into video bit buffer 60B, thus avoiding partial picture storage that could cause decoder lock-up problems. The exemplary microprocessor 510/A can be used to find the start code in I picture mode only, since the end sector is already known from the navigation data. However, for P, B, or multiple I-pictures, the exemplary microprocessor cannot provide a practical solution because a test must be done for each byte of the bitstream data, which represents an operationally intensive operation of the microprocessor 510. use.

Locating and determining the start code prior to decoding is conveniently accomplished by an arrangement that utilizes the link interface block 505 of FIG. 3 to search for the start code in the preceding bitstream of the track buffer 60A. This use of link interface 505 advantageously provides early preprocessing or analysis of images and/or audio headers that may be sent to microprocessor 510 . Thus, after identifying the header of the input bitstream prior to the track buffer, images and audio required by a particular trick mode can be stored in the exemplary track memory 60A, while unneeded images and/or other data are overwritten Deleted in cache.

In a first configuration, the start code is located by using a start code detector 520 which searches the bitstream in either the mechanical/track buffer 60A or the video bit buffer 60B. Although this approach has the advantage that the design of MPEG start code detectors is well known, the detectors however require continuous data. Therefore, only the data in the video bit buffer, remove DVD and transfer data structures can be searched. Thus, searching for MPEG data in the mechanical/track buffer may be difficult to implement, memory usage may not be optimal, and the exemplary microprocessor 510 may be overburdened by some interrupts, thus requiring the addition of a second microprocessor such as 510A Dedicated to implementing start code instrumentation.

In an advantageous arrangement, start code detection is facilitated by a start code detector which searches for a bitstream dedicated to the MPEG start code before or in the track buffer 60A. Thus, by advantageously providing early analysis of the MPEG video headers within the bitstream, trick-play image requirements can be anticipated and memory operations dedicated to trick-play operations can be accomplished. The same beneficial analysis can be applied to the video packet stream ahead of the video bit buffer during trick mode operation. For example, in reverse playback mode, this preprocessing allows trick-play specific selections between pictures to be buffered for decoding, and those not needed to be discarded before storage. Such picture selection, such as discarding B frames, may approximately double the number of I and P pictures stored in the exemplary video bit buffer 60B during trick-play operations. Thus, identifying wanted data from unwanted data is a direct result of pre-processing or analysis prior to buffer storage, which allows video bit buffer 60B to store only needed or trick-play specific images. Therefore, more trick-play specific video object units or VOBUs can be stored to facilitate smooth trick-play motion reproduction.

In one advantageous configuration, the storage capacity of track buffer 60A and video bit buffer 60B is increased during trick play mode by only selecting stored data for subsequent use. For example, in the exemplary trick play mode, B frames may not be decoded and therefore need not be stored in the track or video bit buffer. In this way, only required images are stored and unnecessary images or other data are discarded. To facilitate this beneficial selection between wanted and unwanted pictures requires that the bitstream or video packet stream be preprocessed, analyzed or searched to locate the sequence_header, picturegroup_header or picture_header prior to storage. In this way, analysis or pre-processing of the compressed bit stream allows determination of MPEG parameters such as time_code, closed_group of pictures and broken_link data for each group of pictures or GOP. Additionally, by preprocessing the packet stream, the picture_start_code can be located, which allows processing of the picture_header, which in turn allows determination of eg time_reference, picture_coding_type (I, P and B). However, as already described, such beneficial MPEG analysis is difficult due to the fact that DVD divides MPEG-type data into sectors of 2048 bytes. Also, because the MPEG start code (4 bytes) is not sector aligned, the exemplary picture start code can be allocated across sector boundaries. FIG. 5B illustrates the bit stream prior to track buffer 60A, where video picture A ends at sector 54 and is immediately followed by the start code for video picture B. FIG. However, the remainder of the video picture B start code is generated in sector 65, with intervening sectors 55-64 containing sub-picture and audio data. Figure 5C illustrates the demultiplexed video sector bitstream prior to video bit buffer 60B, where the start code for the exemplary picture A is shown in sector 2 and the start code for the next picture B is in sector 9 produce. The allocated start code for image C appears, which starts at the 2046th byte of sector 12 and continues at sector 13. Therefore, the part of the start code is in one video sector with the remainder of the next video sector.

In order to be able to analyze a bit stream with assigned start codes, FIG. 6 shows an exemplary method according to the invention. The exemplary method identifies and stores the sector type and address, and additionally identifies and stores the required start code. Allocated or partial start codes are identified and stored using the present invention's partial start code markers, which indicate start code occurrences. The remainder of the start code that occurs in the next video sector is recognized and restored to complete the start code. The inventive method of Figure 6 describes the search and MPEG analysis of the bitstream 41 before being applied to the track buffer. The bitstream is searched for the desired sector such as the video sector and then the assigned start code. Other non-video sectors may be separated from the assigned start code by containing eg audio, sub-pictures, navigation data etc. In this way, the bitstream is searched and subsequent video sectors are identified and processed, while intervening non-video sectors that are not currently needed, such as during a particular trick mode, are not processed and stored on the exemplary track buffer 60A. or can be discarded before overriding. Thus, after the next video sector is identified, the packet data is searched to locate the next start code. However, since the partial start code is set, the remainder of the partial start code is searched, and this remainder is combined with the code of the previous video sector to complete the start code.

Figure 6 is an exemplary diagram illustrating the method of the present invention for bitstream searching to identify required sector addresses, image types and addresses and start codes for detection and reassembly allocations. The method begins at step 10, where an error correction bitstream is searched to locate a particular desired sector from a plurality of sectors including navigation, audio video sectors, sub-image data sectors. The video sector is detected at step 100, where a "no" result forms a loop to continue the bitstream search. Similarly, an audio sector may be detected at step 105 and its sector address stored accordingly. If step 100 tests yes, a video sector is detected and the sector address is stored in step 101 . The detected video sector initiates another test at step 200 to detect the start code within the video sector. Step 200 depicts a picture start code, however, there may be various start codes such as sequence header, GOP header or picture header all present within a video sector, so any may be allocated across sector boundaries. A "No" at step 200 forms a loop that continues to search for the start code within the video sector. "Yes" at step 200 means checking the start code which initiates another test to check at step 250 the partial start code. As illustrated in Figures 5B and 5C, when interrupted due to the presence of sector boundaries and sector addresses, the determination between the partial and entire start codes described in steps 200 and 250 can be considered to occur simultaneously and sequentially, since any start code changes be partial or incomplete. A "NO" at step 250 forms a loop waiting for the partial start code to occur. In addition, "No" at step 250 also indicates the detection of the complete start code, which is tested at step 255 to determine if it is of the required type. The start code required test "yes" at step 255 causes the type and byte location within the sector address to be stored at step 260 .

The partial start code at step 250 detects "Yes", which causes the sequence to restart searching the bitstream to locate the next video sector by looping back to step 100. "YES" at step 250 also initiates the test of step 300 to determine if the partial start code flag is set. The partial start code flag is not set until the first assigned or partial start code is detected. Thus, "NO" at step 300 causes the partial start code flag to be set at step 350 and, additionally, the value of the partial start code at step 400 to be stored. A "YES" at step 300 indicates that the remainder of the assigned start code was detected and at step 500 results in a reset of the partial start code flag. "Yes" at step 300 also results in storing at step 450 the remainder of the detected start code. At step 550, the partial start code value from step 400 and the remainder of it from step 450 are combined to refine the assigned start code. Finally, at step 575, the modified start code type, byte and sector addresses are stored. Thus, the inventive method described identifies and stores specific sector types and addresses, identifies and stores start code types and byte addresses within sectors, and identifies and reassembles allocated start code segments. In this way, prior to buffering, a DVD format bitstream can be analyzed to determine, for example, a particular MPEG-encoded picture type.

The MPEG picture decoding order can be advantageously controlled based on the known positions in the video bit buffer where pictures start and stop. Thus, from a known picture position in video bit buffer 60B, such as that illustrated in FIG. 5C or determined by a bitstream search in FIG. Beneficially points to random access images needed eg during trick mode operation. Reverse operation at play speed and/or slow motion playback requires rendering of B frames. This reverse mode operation can be advantageously simplified by reversing the order in which adjacent B pictures are decoded, depending on buffer memory requirements. This reverse decoding order is advantageously obtained by setting the memory start pointer, enabling the decoding of the pictures required for the trick mode. Additionally, buffer memory size and control can be simplified during trick play mode by advantageously skipping or not reading pictures in the video bit buffer as required by the particular trick play algorithm. During trick-play buffering, memory size and control can be further optimized by advantageously enabling multiple picture decoding at once or as specifically required by the trick-play algorithm. Providing these beneficial features requires careful control of read/write functions and synchronization between them.

During trick mode operation, particularly during reverse play speed operation, the maximum picture buffer capacity is required to store groups of pictures for readout in reverse order. During this trick mode, certain player functions or features may not be required, may be useless or unavailable. Such functions or features include audio, multiple languages, sub-images, and on-screen displays and all utilize buffer memory capacity. In this way, buffer memory capacity not used by these functions or features can be reallocated to provide additional image storage during trick mode operation. However, during certain trick modes, such as fast play mode, there may be a beneficial requirement for the accompanying audio to be reproduced at high speed and the pitch corrected to aid scene positioning. Additionally, a limited on-screen display may be required to indicate trick-play speed and direction. In this way, unused buffer memory capacity can be dynamically reconfigured to advantageously facilitate caching compressed images, decoded frame predictors, and video display fields in trick modes.

In one inventive configuration, SDRAM buffer memory 60E-60H is functionally reallocated between forward and trick play mode operation. The memory capacity allocated to audio 60E, sub-image 60G during forward play mode can be used during trick play to provide additional compressed image storage, to augment video bit buffer 60B and provide additional predictor frames for decoding . Similarly, buffer memory may be reallocated, for example, and may not be required to store redundant compressed images for certain trick-play modes, such that unused or unrequired buffer memory capacity is reconfigured to provide output as depicted at 60H of FIG. Display buffer memory. The output display buffer memory can store frames or fields of video data for display. This dynamically allocated memory facilitates output images and is not used as a predictor, thus simplifying memory management during trick mode operation. As already described, memory reallocation may be initiated by user selection, however, dynamic allocation may be determined by favorably stored trick play sequence requirements and/or using favorable picture predictions identified from bitstream analysis of compressed picture types.

In another advantageous arrangement, the frame buffer memory capacity can be effectively doubled during trick play operation by horizontally subsampling the decoded image data prior to buffer storage. According to trick mode control commands from the controller 510 , horizontal subsampling, such as implemented by exemplary block 62 , horizontally averages the values of adjacent pairs of pixels. In FIG. 3, signal S1 represents the full bandwidth data coupled to subsampler 62, and signal S2 represents the resampled output data. Thus, a subsampled image contains approximately half the number of original pixels and, therefore, requires half the memory capacity to allow the image or video frame to be stored by the size of the field. In this way, by horizontally subsampling during trick-play operation, additional framebuffer memory can be used as required by the trick-play algorithm. Additionally, to increase trick mode memory capacity, the use of subsampling in the present invention advantageously reduces data and address bus manipulation by the memory manager during trick mode memory accesses. For example, only half the data is transferred in half the time, thus simplifying memory control and management.

The horizontally subsampled image is retrieved by subsampler 62 from, for example, 60C, 60D, or the reallocated buffer 60H of the present invention. In FIG. 3, signal S3 represents subsampled reference image data read out from memory for pixel count recovery. Subsampler 62 can address each subsampled memory cell twice, however, this action doubles the data and address bus utilization, which is beneficially reduced during memory processing. Thus, the subsampled image is recovered by duplicating each pixel value and output as signal S4 for direct coupling to the MPEG decoder 530 prior to MPEG decoding. While this approach doubles the buffer size and reduces data and address bus utilization, the horizontal spatial resolution is reduced. However, this reduction in horizontal resolution occurs during trick-play operations, but may be imperceptible to human psychovisual perception due to the increased image motion rate.

The block diagram of Fig. 4 shows the same functions and part numbers as described in Fig. 2, however, Fig. 4 includes an additional inventive configuration which will be explained.

The exemplary DVD player shown in Figures 2, 3 and 4 can be considered to comprise two components referred to as the front end and the back end. The front end controls the disc and the converter, and the back end provides MPEG decoding and overall control. This functional division may represent an obvious solution for coordinated, steady-state MPEG decoding. However, with this division of processing and control at the back end, the microprocessor may become overloaded, for example during trick mode operation and especially when playing in reverse.

As already described, the microcontroller 510 is required to manage the input bit stream 41 received from the front-end and to identify needed from unneeded data. In a first advantageous arrangement, a bit stream 41 is controllably coupled between the front end and the back end. As already described, in the exemplary player of Figure 2, the optical pickup or transducer 15 can be repositioned. The sector address obtained at the back end is transmitted to the front end servo system 50 via the I ² C control bus 514 to reposition the converter 15 . However, the optical pickup or converter 15 is servo-controlled according to the sector address, which is truncated to remove the least significant bit. This address truncation allows capturing of sectors in groups or blocks of 16 sectors. This grouping is required for error correction (ECC) during recording by means of ReedSolomon product coding and payload data interleaving applied over 16 sectors. Thus, this information is obtained from the ECC of the group of 16 sectors of the disc, and generally, the retrieved data containing the required sector address is the advance or previous address requested by the backend process. Alternatively, the changer is moved relative to the rotating disc using a radial or tangential motion to obtain tracks containing the ECC blocks of the sectors in which the desired sector addresses reside. Thus, after repositioning, the changer is refocused and the sectors are switched as the disc rotates towards the ECC sector block containing the desired or needed sector address. Therefore, if one considers the worst-case location of the converter and the address of the required sector, hundreds of unnecessary sectors may be translated. Because the number of sectors increases with increasing disc radius, the number of reproduced unnecessary sectors also increases. Additionally, capture of earlier or previous addresses may require a full disc rotation, which would result in unnecessary sector reproduction. In this way, a large amount of unnecessary data is generated before the required sector address occurs. This bit stream is depicted as signal 44 in FIG. 4 and contains both desired and unnecessary data coupled for error correction by ECC blocks 45 and 46 . The error corrected bit stream is output from the ECC process as signal 41, which is coupled to the backend where microcontroller 510 identifies wanted data from unwanted data.

Figure 4 shows an arrangement of the present invention where a data signal 44 is output from the 8:16 code demodulator and coupled to Reed Solomon error correction blocks 45 and 46 via control elements 45A such as transmission gates or logic functions. The control section 45A is controlled by a section 43 whose function is to obtain the recovered current playback sector address, which is error-corrected in block 47 and output as address signal 42, with the back end representing the next required data, such as the image type. Compared with the sector address 53A. This comparison is conveniently accomplished by a comparator or logic function. Thus, signal 43A allows the demodulated data output to be coupled to error correction buffer blocks ECC 45 and 46 when playback sector address 42 is equal to address 53A required by the backend. Because error correction is applied to groups of 16 sectors, a comparison of the requested address with the actual address is performed so that the ECC sector block containing the requested sector can be used for Reed Solomon correction. For example, addresses with truncated least significant bits facilitate sector address comparisons.

For example, since an MPEG picture of type B may occupy 3 sectors, where an MPEG picture of type I may require 30 or more sectors, the required sector address indicates the initial data sector of the desired picture type. Alternatively, signal 43A indicating that the required and playback sector addresses are substantially equal may be considered to represent a lockout function, where the logic state is held until the required address is changed, ie until another converter jump is required. Receipt of a new sector address changes the state of signal 43A, which blocks reproduced data until the new desired address is present in the reproduced signal and detected by comparator 43. In other words, signal 44 remains enabling error correction, ECC blocks 45 and 46 are enabled and output signal 41 is maintained, or simply, the disc continues to play until a different changer position is requested.

Detection of the required sectors for replay to occur can be performed by comparing the truncated sector addresses to ensure that the error correction buffers 45 and 46 are filled with the number of sectors required for RS correction. In another embodiment, the same detect replay occurrence may be employed using signal 45B to control or enable operation of error correction buffer memories 45 and 46 . In another configuration of the present invention, only required sectors are enabled through output control section 46A. The selection of component 46A differs from the control provided by components 45A and 45B because of the interleaved or shuffled data format, enabling the ECC block containing the desired sector. Detection of the desired replay sector can be accomplished by comparing the actual replay sector address with the desired or desired address. However, since this control function is basically performed after error correction and deshuffling with the buffer memory, the final output signal 41 is delayed by at least one ECC block time period. The error-corrected output signal thus corresponds to the group of sectors switched before the required data (address) as identified at the input of the ECC buffer. Clearly, since the buffering delay is known, for example by using the delay method described as t, this delay can be compensated in the control coupling of signal 43A to component 46A. Control element 46A is depicted as a series-type switch element capable of enabling or disabling the bit stream provided to the backend. As such, signal 43A, properly timed to compensate for processing and buffering delays, may be applied to selectively enable transmission of de-interleaved bitstream 41 to processing block 500 . The use of previous embodiments of the present invention allows only converted data from required sectors to be coupled to the back end for storage and decoding, thus reducing the workload on microprocessor 510.

As already described, the converted signal 31 is demodulated in block 40 to remove the 8:16 modulation, and output signals 44 and 44A are produced. Signal 44 is coupled for deinterleaving and error correction, and signal 44A is error corrected alone to generate playback sector addresses. Deinterleaving and error correction are performed in the ECC buffer memories 45 and 46 of FIG. Each buffer stores 16 sectors of the playback data stream arranged in an array to facilitate deinterleaving and/or enable the required row and column product processing. The cascaded ECC buffers introduce a delay to the reproduced serial bit stream at 1x rotational speed, which can be approximated by the following calculation (2*16*1.4) milliseconds, where 2 denotes the ECC buffers 45 and 46, 16 indicates the sector to which the correction is applied, and 1.4 milliseconds indicates the sector period at 1 rotation speed. Thus, the rendered sequential bitstream is delayed by a minimum of approximately 45 milliseconds.

Bitstream 44A is processed at ECC block 47 to correct sector identification addresses for errors. However, since sector addresses are short and sector specific, error correction block 47 introduces a negligible delay in replaying sector address signal 42 .

As already described, an error correcting bitstream undergoes an error correction delay. The bitstream 41 is received at the back end where the various MPEG packets are separated from the DVD data. The video packets are stored in exemplary buffer 60B for decoding by MPEG decoder 530 . As described above, the decoder 530 sends a signal 513 to the controller 510 to indicate the completion of each decoded picture, which then captures the next picture to be decoded. Thus, signal 513 is generated by the decoder at the end of a particular picture, such as the picture contained in the video sector labeled A in FIG. 5A. The exemplary next picture required for decoding must be retrieved from the disc, therefore, the converter 15 must be relocated to the address of the sector containing the desired picture. Figure 5A shows a portion of bitstream 41 coupled to buffer 60A, comprising a video object unit consisting of a plurality of sectors, each sector containing video, audio, sub-image and navigation data. The end of sector A may advantageously be determined in or before track buffer 60A as the next sector address occurs, or after MPEG decoding as indicated by signal 513 . Thus, the arrow labeled "Next" in FIG. 5A shows the approximate timing of the occurrence of the next sector address request from the microprocessor 510 to the front end. This address and jump request are communicated over the ^I2C control bus which delays the required sector request according to the interrupt priority.

In another advantageous configuration, the interrupt priorities of microcontroller 510 interrupts are reordered between operating modes. For example, in forward play mode, memory addressing and control requests are different than those required during trick mode operation and especially during reverse play speed operation. Certain features, and thus their memory and MPEG decoder control, are not required during trick mode operation. For example, audio decoding and sub-picture processing are not required during trick mode operation, so address, data, and control bus interrupt priorities can be assigned lower priority, while higher priority is assigned to slave tracks and video Bit registers to access images.

Timely capture of requested sectors is especially important during trick mode operation. However, as already described, the execution of required sector acquisition according to the back-end processing forms a control loop with multiple delay components. Figure 4 shows an inventive arrangement which reduces the delay in sector acquisition, simply by allowing detection replay of the last required sector to occur to initiate the changer movement to the previously received new sector address. Arrow B shown in FIG. 5A is positioned to indicate the approximate timing relationship between replayed bitstream 41 or track buffer 60A and issuance to the servo of the next/end sector address of the present invention. In FIG. 5A, arrow B is shown occurring shortly after the shaded navigation packet has been read from the bitstream. In image A, the arrow "next" is shown indicating that signal 513 occurs approximately seven sectors later (decoding complete). However, in reality, the I and P type images contain many more sectors than depicted in FIG. 5A, therefore, the arrow "next" issued corresponding to the address and jump request occurs much later than illustrated. Thus, the present invention next/end sector address is generated by microcontroller 510 after navigation packet capture and/or beneficial image/sector address determination and table construction. Identifying the required sector address using the next/end sector address can be separated in time from the jump converter instruction. The next/end sector address is effectively preloaded in the converter servo with a converter jump performed based on the first without requiring sector address reproduction. Because the sector addresses do not experience the lengthy ECC delay of the bitstream 41, the converter is shifted before the last unwanted sector emerges from the ECC blocks 45 and 46.

In FIG. 4, the control data is communicated via the ^I2C control bus 514, which communicates the next desired playback sector address to the servo control system 50. In FIG. The next required playback sector address is generated by microcontroller 510 which processes address data from stored trick play, specific speed sequences, playback and stored navigation data or from beneficially determined playback image data . The next address is read from the I ² C bus and stored in unit 53 . The ^I2C data also includes the end/last sector address of the present invention, or the first unneeded sector address. The end/last sector address can be obtained from the recovered and stored navigation data, however, this only provides a limited number of predetermined image addresses, so for trick modes it is advantageous to use the end of the image sector address. The ending/last sector address is read from the I ² C bus and stored in unit 52 . The last sector address can be modified before bus transfer or after reception to prevent loss of needed sectors, by for example adding a unit count to the sector address, thus ensuring addressing and detection of the first unwanted sector. The last sector address or modified address 52A is coupled for comparison with replay sector address signal 42 in exemplary comparator 51 . Thus, when playback sector address 42 is equal to address 52A, the first unnecessary sector will be switched and comparator 51 generates control signal 51A. Control signal 51A enables coupling from component 53, such as by loading or moving stored address data to the servo, or as described by the exemplary selector switch 54, which couples the next sector address to the servo and initiates a shift device 15 repositioning. As already described, the changer moves to the track containing the next desired picture and the data output signal 41 is advantageously enabled by means 43 when the desired picture is reproduced.

The changer continues to follow the tracks rendering the required sectors for processing by the backend. Based on the data recovered from these sectors, a new pair of next and end sector addresses are generated and transmitted via ^I2C . These new addresses are received and stored in components 52 and 53 as before. However, to avoid initiating converter jumps before the new ending sector address is replayed and detected by component 51, exemplary selector 54 is reset or opened, preventing premature initiating and capturing of the new sector address.

The inventive transducer control sequence described above initiates transducer motion by comparing the substantially undelayed replay sector address with the preloaded desired sector address, thus avoiding delays in capturing a new replay bitstream, Facilitates enhanced trick mode operation.

As is known, the MPEG picture decoding order is determined by the coded picture hierarchy, thus following the decoding sequence for forward mode operation. However, trick-play operation can be advantageously implemented by controlling the MPEG picture decoding order based on the sequence of pictures required by the predetermined trick-play algorithm and known picture starts and stops in the video bit buffer. Thus, knowing the picture position in video bit buffer 60B, such as calculated in FIG. 5C or determined by bitstream search in FIG. Random access to images required during trick mode operation. The exemplary video bit buffer shown in Figure 5D contains image tiles as described above. The start code detector memory pointer is depicted by arrow SCD which searches the exemplary video bit buffer to locate the MPEG start code. However, in the third sector of the first P picture, the start code detector memory pointer SCD1 indicates the detection of the start code from the next, but unwanted picture. Thus, by beneficially directing the start code memory pointer to a known byte-accurate memory location, as represented by arrow SCD2 of FIG. 5D , unwanted pictures and undesirable decoder hangs are avoided.

In another advantageous trick mode configuration, unneeded data from previous pictures is flushed into and out of the FIFO, start code detector (SCD) 520 and variable length decoder (VLD) 531 first-in-first-out registers. Signals 521/532 depicted in FIG. 3 clear or reset the corresponding FIFO to clear remaining data from previous decode operations. This clearing or flushing of the FIFO ensures that the SCD and VLD start the next decoding operation with new data from demonstration bit buffer 60B, thus eliminating another source of decoder misoperation caused by remaining previous data.

Reverse operation at playback speed requires rendering of B frames and, in another trick mode, buffer memory requests in reverse order from decoding adjacent B pictures, advantageously simplifying optimized reverse mode operation. By setting or controlling the memory start pointer to enable the decoding of the particular picture required by the trick mode, the decoding order is advantageously reversed. In another trick mode optimization, buffer memory size and control. Memory size and control can be further optimized during trick-play by advantageously enabling the decoding of multiple images immediately or as specifically required by the trick-play algorithm. The beneficial features provided require careful control of read/write functions and synchronization between them.

In another trick mode optimization, the controllability of decoders to facilitate audio-video synchronization or panning by skipping picture decoding is advantageously increased over the control range and exploited during trick mode operation to allow A selectable number of pictures between 1 and at least 6 is skipped or not decoded. This picture manipulation advantageously facilitates trick-play operation at six times playback speed by skipping B-pictures within each GOP.

In addition to the memory control and allocation requirements for trick mode operation, it can be beneficially optimized by substantially simultaneous operations such as decoding an I or P picture within a field period and writing the decoded result for display and/or memory storage MPEG decoding. Assume the ability to decode B-type pictures without using buffer memory. This B-type image decoding is called Bframes-on-the-fly (BOF) in the air. Additionally, trick-play operations can be beneficially enhanced by writing decoded fields to memory and simultaneously reading display fields from the same in-memory interleaving unit. Display fields can be from temporally separated images. Such substantially simultaneous read and write operations can be performed within a display field period. However, decoded fields do not have to overwrite or conflict with display field reads. B-pictures do not require such interleaved read and write operations because they can be decoded without a buffer memory.

In an exemplary player with reverse trick-play decoding, bitstream or track buffer 60A is used to store a compressed MPEG video bitstream retrieved from media. Track buffer 60A or compressed video bit buffer 60B may be used to facilitate access to multiple individual MPEG pictures. The decoded trick-play output signal must conform to the TV signal standard to allow display by normal TV receivers. The following example illustrates the inventive control sequence for MPEG decoding in a DVD player. Figure 7 is a diagram illustrating the arrangement of the present invention for use in reverse trick play mode at 3x playback speed (3x) in a video player. The exemplary diagram has columns representing MPEG-coded I-pictures and P-pictures, which include groups of pictures or GOPA, B, C, and D. Each GOP contains twelve pictures not obtained from a film source.

In this exemplary trick-play sequence, reverse decoding is conveniently accomplished with an advantageous configuration of an MPEG decoder and two frame buffers that provide reverse order decoding and display of the decoded video. In this example, only I-pictures and P-pictures are decoded, so only they are listed on the diagram. Figure 7 illustrates a sequence of 37 encoded images, with image numbers indicated in parentheses. The rightmost column is labeled "Output Field #" and represents the time axis incrementing in field periods. The first field, Output Field #1, marks the start of a trick-play reproduction. Each row in the graph shows the processing of the present invention that occurs during the corresponding field period. The following abbreviations are used in Figure 7. The frame buffers are numbered 1 and 2. A capital "D" denotes decoding the image/frame indicated at the top of a particular column. The process of decoding an image and storing its result is described by "D>1", where the number represents the destination frame buffer number, ie 1. A lowercase "d" indicates the display of a field from a frame of a particular column. An output field can be selected to hold the output signal interleaving sequence. In order to provide a continuous sequence of output fields, it is explicitly required that each row of the diagram contain a field display directive 'd'.

The sequence illustrated in Figure 7 begins at output field #1, where the I picture I (37) is decoded and stored in frame buffer 1 or 60C. A field such as the top field of an I frame (37) is displayed while the I picture (37) is being decoded. An advantageous decoder 530 is used to facilitate decoding and simultaneous display of the decoded video signal. During output field #2, MPEG picture 1 (25) is retrieved from bitstream buffer 60B, decoded and stored in frame buffer 2 or 60D. Simultaneously, another field such as the bottom field of I(37) read from frame buffer 1, 60C, is displayed.

During the output field #3 period, actions illustrating an aspect of the present invention occur. During field #3, the top field of exemplary I(37) is repeated by reading from frame buffer 1, ie 60C. Simultaneously with the readout of the repeated top field of I(37), the predicted picture P(28) is decoded with reference to I(25) and stored in frame buffer 1, ie 60C. With exactly synchronous timing, the decoded frame P(28) is written into frame buffer 1 or 60C. This simultaneous operation is achieved by sequentially decoding picture P (28) on a line-by-line basis after picture I (37) display field readout. Sequential reading and writing of the frame buffer 1 is another beneficial feature provided by this exemplary decoder and memory management system.

At the end of output field #3, pictures I (25) and P (28) of GOPC are stored in frame buffers 1 (60C) and 1 (60D), respectively. However, these frames represent earlier events in time and are required to be able to decode sequentially generated frames such as frames P(31) and P(34). Intra coded picture I (25) residing in memory 2 (60D) is used to decode frame P (28), but is not currently needed. Thus, to provide the display for output field #4, frame memory 2 is overwritten with frame I (37), reread from video buffer 60B and decoded. To maintain the output interleaved field sequence, the appropriate fields of frame 1 (37) are fetched from frame buffer 2 for display. In output field #5, the beneficial simultaneous processing done in field #3 is repeated. Output field #5 is obtained by reading out the field of image I (37) from frame buffer 2. Meanwhile, the picture P(31) is decoded with reference to the picture P(28) from the frame buffer 1, and the decoding result is stored in the buffer 2. Thus, the first five output fields of this exemplary triple reverse playback include still, or frozen, I-picture (37) pictures. However, at the end of output field #5, the trick play output signal is generated starting with pictures I(28) and P(31) stored in frame buffers 1 and 2, respectively.

In output field #6, the predicted picture P (34) is read from the bitstream buffer 60A or video bit buffer 60B, decoded and displayed without storage in the appropriate field. Thus, field #6 enables the display of 3x speed reverse motion. At output field #7, image P (34) is retrieved again, decoded and other selected fields are used for display. The picture P(31) previously decoded and stored in the frame buffer 2 is read out and output fields #8 and #9 are respectively provided.

At the end of output field #9, there is no longer a requirement to store picture P (31), so the intra-coded picture I (13) of the next preceding GOP B is acquired, decoded and stored in frame buffer 2. Output fields #10 and #11 are read from the frame buffer 1 containing the predicted picture P (28). Simultaneously with reading out the field #11, the predicted picture P(16) is obtained from the bit stream buffer 60B, decoded and sequentially stored in the frame buffer 1. Since both frame buffers contain the anchor frame of the next preceding GOP B, output fields #12 and #13 are obtained in the same manner as output fields #6 and #7. The predicted picture P(25) is read from the bitstream buffer 60B, decoded and if not stored, the appropriate fields are displayed.

Thus, processing as described for GOP C encompasses the next preceding GOP B of pictures I(13), P(16), P(19) and P(22).

Claims (3)

1. A method for controlling data reproduced in sectors by an optical disc player using optical readout, comprising the steps of: Convert groups of sectors, which include sectors required for processing and sectors not required for processing; providing said needed sectors other than said unneeded sectors to a data processor for processing, and The required sectors are processed to extract data representing video information therein. 2. A method of controlling MPEG compatible data reproduced in sectors, comprising the steps of: switching groups of sectors comprising requested sectors with MPEG compatible data required for processing and unrequested sectors with MPEG compatible data not required for processing; transferring said requested sectors other than said unrequested sectors to a data processor for processing; and The requested sectors are processed to extract the requested MPEG compatible data representing video information. 3. A digital disc player for reproducing sectors containing MPEG compatible data, comprising: a converter for converting groups of sectors from the optical disc comprising requested sectors with MPEG compatible data required for processing and unrequested sectors with MPEG compatible data not required for processing; a first data processor, coupled to said converter, for processing and coupling said requested sectors with MPEG compatible data other than said unrequested sectors with MPEG compatible data not required for processing; as well as A second processor, coupled to the first processor, for receiving the requested sector of data and extracting the requested MPEG-compatible data representing video information.

Images