Direct access compact disc, writing and reading method and device for same

Abstract

A system that redefines how data is distributed on a conventional writable compact disc (CD-R/E). A rearrangement of the data on the disc provided during the writing operation preserves eight-to-fourteen channel frames and the control and display (C&D) channel and burst error mitigation while providing a direct access storage device (DASD) format and capability. The CD-DASD format is suitable for preformatting the CDs and has constant size sectors recorded contiguously along the spiral track. Each sector is independently addressable and synchronous with the C&D data word and ATIP channel words on the CD-R disc. The system uses the components of a conventional CD device and a mapping controller address translator to encode and decode the data bytes using a conventional CIRC encoder/decoder. A rectangular product code of C1 and C2 CIRC subcodes is provided that is interleaved to mitigate the effects of handling. The system provides for locking in on the changing data frequency that occurs when moving between spirals of the CD allowing reading and writing to occur while the CD is coming to the proper speed.

Claims

What is claimed is:

1. A compact disc, comprising:

a compact disc storage media; and

data stored on said media with a compact disc encoding and physical marking in a direct access storage device format comprising independently addressable sectors, wherein said disc is preformatted with sector headers which are produced separately and prior to any subsequent writing of information onto the disc using a CD-DASD drive.

2. A disc as recited in claim 1, wherein said data comprises user data encoded in a rectangular product code.

3. A disc as recited in claim 2, wherein said product code comprises C1 and C2 Reed-Solomon codes.

4. A disc as recited in claim 2, wherein the rectangular product code is interleaved at an interleave depth.

5. A disc as recited in claim 2, wherein user defined data includes bytes usable for one of additional error correction, synchronization, sector addresses, sector boundary location, sector mode type and disc type.

6. A disc as recited in claim 2, wherein all symbols of the code are stored contiguously within a single sector on said disc.

7. A disc as recited in claim 2, wherein the user data includes a preamble, a buffer, user defined data and parity.

8. A disc as recited in claim 7, wherein sectors are spliced together in the buffer.

9. A disc as recited in claim 1, wherein said format includes an eight-to-fourteen modulation.

10. A disc as recited in claim 1, wherein said format includes control and display data.

11. A disc as recited in claim 1, wherein the data contained in the sector headers is one of directly accessible upon reading of the disc and optionally accessible after only C1 decoding in performed.

12. A disc as recited in claim 1, wherein said format includes eight-to-fourteen modulation and the headers include interleaving of header data and zero value bytes within eight-to-fourteen modulation frames of the headers.

13. A disc as recited in claim 1, wherein the headers end with a mark and are followed by a preamble with a space bit sequence.

14. A disc as recited in claim 13, wherein the space bit sequence comprises eleven bits.

15. A disc as recited in claim 1, wherein the headers include a cyclic permutation of a variable frequency oscillator signal yielding a minimum digital sum variation channel data stream.

16. A disc as recited in claim 1, wherein said data is logically mapped between the compact disc encoding and marking and the direct access storage device format.

17. A disc as recited in claim 1, wherein a direct access storage device operating system process can access said data.

18. A disc as recited in claim 1, wherein said format includes a header area having eight-to-fourteen modulation frames and header data is repetitively encoded in separate C1 codewords each written in separate ones of the eight-to-fourteen modulation frames of the header area.

19. A disc as recited in claim 1, wherein each header on the disc is written with constant offset relative to synchronization bits of codewords of absolute time in a pregroove information stream encoded in a wobble pregroove of an unwritten compact disc.

20. A disc as recited in claim 19, further comprising a disc substrate having a spiral pregroove, wherein the headers on the disc are formed by molding, or embossing, pits and intervening land areas at regular length intervals along the spiral pregroove of the disc substrate.

21. A disc as recited in claim 1, wherein further comprising a header area storing plural instances of data which is read using a majority logic voting process.

22. A compact disc comprising:

a compact disc storage media; and

data stored on said media with a compact disc encoding and physical marking in a direct access storage device format comprising independently addressable sectors, wherein the data includes information stored in a preamble, such information being one of directly accessible upon reading the disc and optionally accessible after only C1 decoding is performed.

23. A disc as recited in claim 22, wherein the preamble includes a gap.

24. A disc as recited in claim 22, wherein the preamble includes a virtual field adapted to allow recording of eight-to-fourteen modulation frames of constant size.

25. A disc as recited in claim 22, wherein said disc includes a data synchronization field, wherein the relationship between the recovered instance of the data synchronization field and a replica of the data synchronization field which is stored in a register in a drive indicates the synchronization quality of the data recovered from the disc.

26. A compact disc, comprising:

a compact disc storage media; and

data stored on said media with a compact disc encoding and physical marking in a direct access storage device format comprising independently addressable sectors, wherein said format includes sectors with a pre-recorded header area, a preamble area, a data area including address/identification data, user defined data and reserved bytes together with corresponding parity data, and a buffer area.

27. A disc as recited in claim 26, wherein each of said sectors are recoverable distinctly from reading/writing any other of the sectors.

28. A compact disc, comprising:

a compact disc storage media; and

data stored on said media with a compact disc encoding and physical marking in a direct access storage device format comprising independently addressable sectors, wherein said data is recorded, using eight-to-fourteen modulation, as a succession of eight-to-fourteen modulation frames which each include the eight-to-fourteen modulation representation of a Compact Disc control and display byte.

29. A disc as recited in claim 28, wherein said eight-to-fourteen modulation frames include an eight-to-fourteen modulation channel bit synchronization field.

30. A method of writing data to a compact disc, comprising:

forming a rectangular product code using the data producing encoded data; and

writing the encoded data into contiguous locations on the compact disc in a single sector of plural independently addressable sectors of the disc.

31. A method as recited in claim 30, wherein said forming comprises:

encoding the data into Reed-Solomon C2 codewords;

forming a rectangular array of the C2 codewords; and

encoding the rectangular array into Reed-Solomon C1 codewords.

32. A method as recited in claim 31, said forming further comprising:

adding control and display information to the rectangular array C1 codewords; and

concatenating seven product codewords.

33. A method as recited in claim 32, further comprising interleaving the product code words at a depth of interleave.

34. An apparatus for writing data to a compact disc, comprising:

a system forming a rectangular product code; and

a compact disc writer writing the product code into contiguous locations of a disc sector among a plurality of independently addressable sectors of the disc, wherein said disc is preformatted with sector headers which are produced separately and prior to any subsequent writing of information onto the disc using a CD-DASD drive.

35. An apparatus, comprising:

a CD cross interleaved Reed Solomon decoder including a demodulator demodulating eight-to-fourteen modulated channel data read from a compact disc, a Reed-Solomon C1 decoder and a Reed-Solomon C2 decoder;

a memory storing data transferred between the demodulator, the C1 decoder and the C2 decoder; and

an address translator remapping storage and retrieval addresses enabling a rectangular C1/C2 product code that is contiguously written on the disc to be decoded by the C1 and C2 decoders when only the information comprising the rectangular product codeword currently being decoded has been read from the disc.

36. An apparatus as recited in claim 35, wherein said remapping performs reorganization of a data sequence read from the disc into C1 and C2 code words at the input of the C1 and C2 decoders.

37. A method of reading data from a compact disc, comprising:

demodulating the data from the disc;

storing the data in a memory in data-0 out locations;

retrieving the data from data-1 in locations in the memory;

Reed-Solomon C1 decoding the data from the memory;

storing the data that has been C1-decoded in the memory in data-1 out locations;

retrieving the data that has been C1 decoded from data-2 in locations in the memory; and

Reed-Solomon C2 decoding the data retrieved from the data-2 in locations.

38. A computer system, comprising:

a computer requesting a direct access storage device data transfer; and

a compact disc system performing the direct access storage device transfer using a CD storage format by independently addressing disc sectors, wherein said disc is preformatted with sector headers which are produced separately and prior to any subsequent writing of information onto the disc using a CD-DASD drive.

39. A compact disc, comprising:

a storage media; and

data stored on the media in a direct access storage device compact disc encoding, modulation and physical marking format comprising independently addressable sectors, wherein said disc is preformatted with sector headers which are produced separately and prior to any subsequent writing of information onto the disc using a CD-DASD drive.

40. A compact disc, comprising:

a storage media; and

data stored on the media in a direct access storage device--compact disc (CD-DASD) format comprising independently addressable sectors, wherein said disc is preformatted with sector headers which are produced separately and prior to any subsequent writing of information onto the disc using a CD-DASD drive.

41. A compact disc system, comprising:

a compact disc storage media;

a compact disc writer forming independently addressable, constant size, contiguously stored sectors on said media, each sector including a header, a preamble with a virtual gap, user specifiable data and a buffer, the users data including an address, variable data, parity and system data, with adjacent sectors spliced together in the buffer, the user data being encoded into rectangular product codes using C1 and C2 codes with the codes being interleaved at an interleave depth, the media having control and display information and eight-to-fourteen modulation of frames; and

a compact disc reader including a CD CIRC (Cross Interleaved Reed Solomon Code) decoder and an address translator outputting the user data of said media, wherein said disc is preformatted with the sector headers which are produced separately and prior to any subsequent writing of information onto the disc using a CD-DASD drive.

42. A system, comprising:

a computer initiating a direct access storage device request; and

a compact disc drive connected to said computer, receiving the request and including a C1/C2 decoder, said drive accessing a compact disc, decoding contents of the disc using the decoder and providing decoded contents to said computer, the disc comprising data formatted using eight-to-fourteen modulation frames including control and display information, the data being divided into individually addressable sectors, each sector comprising a header, a preamble with a virtual gap, user specifiable data, parity and a buffer, with adjacent sectors spliced together in the buffer, each sector being parsed into logical sub-blocks encoded into contiguously stored, interleaved, C1/C2 rectangular product codes, the header comprising header data interleaved with zero value bytes, the header ending with a mark, and the preamble starting with an eleven bit space sequence.

43. A compact disc, comprising:

a compact disc storage media; and

data stored on said media with a compact disc encoding and physical marking in a direct access storage device format comprising independently addressable sectors, wherein said data is recorded, using eight-to-fourteen modulation, as a succession of eight-to-fourteen modulation frames which each include the eight-to-fourteen modulation representation of a Compact Disc control and display byte.

44. An apparatus for writing data to a compact disc, comprising:

a system forming a rectangular product code; and

a compact disc writer writing the product code into contiguous locations of a disc sector among a plurality of independently addressable sectors of the disc, wherein data is recorded, using eight-to-fourteen modulation, as a succession of eight-to-fourteen modulation frames which each include the eight-to-fourteen modulation representation of a Compact Disc control and display byte.

45. A computer system, comprising:

a computer requesting a direct access storage device data transfer; and

a compact disc system performing the direct access storage device transfer using a CD storage format by independently addressing disc sectors, wherein data is recorded, using eight-to-fourteen modulation, as a succession of eight-to-fourteen modulation frames which each include the eight-to-fourteen modulation representation of a Compact Disc control and display byte.

46. A compact disc, comprising:

a storage media; and

data stored on the media in a direct access storage device compact disc encoding, modulation and physical marking format comprising independently addressable sectors, wherein said data is recorded, using eight-to-fourteen modulation, as a succession of eight-to-fourteen modulation frames which each include the eight-to-fourteen modulation representation of a Compact Disc control and display byte.

47. A compact disc, comprising:

a storage media; and

data stored on the media in a direct access storage device--compact disc (CD-DASD) format comprising independently addressable sectors, wherein said data is recorded, using eight-to-fourteen modulation, as a succession of eight-to-fourteen modulation frames which each include the eight-to-fourteen modulation representation of a Compact Disc control and display byte.

48. A compact disc system, comprising:

a compact disc storage media;

a compact disc writer forming independently addressable, constant size, contiguously stored sectors on said media, each sector including a header, a preamble with a virtual gap, user specifiable data and a buffer, the users data including an address, variable data, parity and system data, with adjacent sectors spliced together in the buffer, the user data being encoded into rectangular product codes using C1 and C2 codes with the codes being interleaved at an interleave depth; and

a compact disc reader including a CD CIRC (Cross Interleaved Reed Solomon Code) decoder and an address translator outputting the user data of said media, wherein said data is recorded, using eight-to-fourteen modulation, as a succession of eight-to-fourteen modulation frames which each include the eight-to-fourteen modulation representation of a Compact Disc control and display byte.

Description

MICROFICHE APPENDIX

1 sheet of microfiche containing a total of 20 frames is included herein.

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention is directed to a compact disc (CD) system of the optical or optomagnetic type capable of reading discs recorded in the standard CD-Audio and CD-Read Only Memory (CD-ROM) formats, reading and writing discs in the CD-recordable (CD-R) format and/or the newly proposed CD-erasable (CD-E) format, as well as reading/writing in a direct access storage device (DASD) format, and, more particularly, to a system that uses the typical components of a CD-Audio/ROM system and low cost additional components to write/read data on a disc in both the CD-Audio/ROM and CD-DASD formats.

2. Description of the Related Art

The Compact Disc.TM. (CD) optical data storage system was originally designed as a consumer product that would read (playback) digitized audio information in a sequential fashion, much like a tape, from unprotected plastic discs that would be extensively handled. Accordingly, the recording format (i.e., the precise manner in which the data stored on the disc is mapped to the trail of physical marks written on the disc's surface) of this system is optimized for the continuous retrieval of data from the disc and also to mitigate the affects of relatively large defects (such as scratches and fingerprints) on the reliability of the data recovered from the disc. The CD-Audio recording format therefore handles (during reads & writes) input and output data (i.e., user digital audio data) in small, contiguous 24-byte blocks called "frames" and further causes the data that comprises a single frame to be widely distributed on the surface of the disc when it is recorded. Moreover, there is no provision in the recording format for the precise addressing of an individual frame (i.e., allowing the CD playback device to determine the exact physical location of any of the constituent bytes in a frame on the disc). In fact, the only means of locating information on the disc is via the information carried by a separate control & display (C&D) channel that is multiplexed with the main (digital audio) data channel.

The specific item of information carried by the C&D channel that provides the vehicle for locating information on the disc is the "absolute-time-on-disc" which is the elapsed disc playing time relative to the beginning of the recorded disc information area. Absolute time information is updated with a granularity of 1/75th of a second. Since exactly ninety-eight 24-byte frames of audio data are played each 1/75th second, the C&D channel can be used to "segment" the contiguous audio data stream channel on the disc into data blocks that contain 98.times.24=2352 bytes. A main (audio) data channel block that consists of 98 contiguous frames, or 2352 bytes of digital audio data, is called a "C&D Section". However, a given 2352-byte C&D Section cannot be precisely located on a disc; this is due to the fact that the CD-Audio disc recording standards provide for a tolerance of .+-.1 second between the start of the C&D channel's absolute-time-on-disc information and the start of audio program data on the disc. (Note: The absolute time value is specified to be 0 minutes, 0 seconds and 0 seventy-fifth seconds at the start of the first data (audio) track of the disc, which immediately follows the disc's lead-in track. The lead in track is the first track in the disc's information area: absolute time increases from some negative value during the lead-in track such that it becomes zero exactly at the end of the lead-in track)

In 1984, or thereabouts, a new version of the CD system known as Compact Disc Read-Only-Memory (CD-ROM) was introduced. CD-ROM was designed as a playback-only computer peripheral and CD-ROM drives connected to a computer could be used to retrieve files of data from a prerecorded disc in response to commands from a requesting application program. To control the cost of the CD-ROM drives and to provide them with the capability to "play" CD-Audio discs, the recording format of the CD-Audio system was fully retained in the CD-ROM system. This enables CD-ROM discs, which each may hold over 600 Mbytes of data, to be produced on the same manufacturing line as CD-Audio discs and allows CD-ROM drives to share components with CD-Audio players. The CD-ROM system has proven to be a commercially successful, low cost means of distributing very large data sets and application programs to computer users.

Computer operating systems (i.e., the programs that, among other things, manage the storage and retrieval of data needed by application programs that are running on a computer) are designed to move data between the central processing unit (CPU) and the computer's storage peripherals in units, or blocks, called "data clusters". Clusters always contain 2.sup.n bytes, where n is an integer (usually n.gtoreq.10). Computer peripherals, such as hard disk drives, therefore, are designed to handle data in blocks called sectors that each contain 2.sup.m bytes of arbitrary-valued data that could be assigned to a specific cluster that belongs to some user data file (usually m is an integer .gtoreq.8). Because of the way that information on a compact disc is segmented by the timing information in the C&D channel, the CD-ROM system employs sectors that contain 2352 bytes and, in the most widely used embodiment of CD-ROM, each sector holds 2048 "user bytes", or arbitrarily valued bytes that could belong to a user data file.

The 2352 byte CD-ROM sectors are logically defined by exactly mapping them. i.e., assigning their contents to, 98 contiguous 24-byte frames. However, as was mentioned previously, the data in each of these frames is widely distributed along the disc's spiral data track. In fact, data stored on the disc data track is organized as contiguous 33-byte blocks called "eight-to-fourteen modulation (EFM) frames." Each EFM-frame contains one byte of (multiplexed) C&D channel information, eight bytes of error correction code (ECC) parity data and 24-bytes of user data. Each byte of user data in a given EFM-frame is obtained from twenty-four different 24-byte data frames that are distributed over 106 contiguous data frames. Thus, the 24 bytes of a given data frame are distributed over 106 consecutively recorded EFM-frames on the disc. But, in order to recover the 24 bytes of a single data frame from the disc, 111 consecutive EFM-frames have to be retrieved (the additional 5 EFM-frames contain all the ECC parity data needed to complete, and thereby render decodable, the ECC codewords that protect the specific 24-byte data frame).

Recall that the C&D channel's absolute-time-on-disc information segments the main data channel on the disc into 2352-byte C&D Sections (this is true for CD-ROM discs as well as CD-Audio discs because their low-level recording formats are exactly the same). Unfortunately, this segmentation cannot be used to precisely define where (on a CD-ROM disc) the boundaries, or start, of a given sector resides. This is due to the fact that the control & display (C&D) and main data channels are not aligned (as noted previously). Thus, since a sector may start in (that is, the first byte of the recorded sector may occur in) any arbitrary 33-byte EFM-frame on the disc, the "offset" between the boundaries of CD-ROM sectors and the C&D Sections on the disc will be <.+-.98 EFM-frames (or equivalently, <.+-.1/75 second since EFM-frames are synchronous with data-frames; one EFM-frame is formed for each data frame that is input to the CD-Audio/ROM encoder). To facilitate locating information on a CD-ROM disc each sector contains "address" data, which is used by the CD-ROM drive's controller to identify specific sectors (the computer operating system also uses a translation of this address data, together with the disc directory and file allocation tables, to identify how the user data in the sectors relates to the files on the disc). Thus, to retrieve a specific sector from a disc the CD-ROM drive must first read approximately 300 sequential 33-byte EFM-frames from the disc and then deliver the data contained in them to the drive's controller which "finds" the 98 sequential 24-byte data frames that comprise the sector and extracts the desired user data. Even if the offset between sectors and C&D Sections is zero, more than 200 contiguous EFM-frames still must be read to retrieve a single sector. This is because entire or complete error correction codewords must be recovered before decoding of the ECC words can be accomplished; the data needed to complete all of the error correction codewords that protect data that resides in the sector of interest is distributed over 208 contiguous EFM-frames. The underlying CD-Audio recording format specifies this wide scattering of the data that comprise individual codewords to enable the correction of long data error bursts that may be caused by large defects on the disc caused by handling.

In 1990, the Compact Disc-Recordable (CD-R) system was introduced. A CD-R "writer" can write digital audio data or logical CD-ROM sectored data to recordable discs that can subsequently be read in any CD-Audio player or CD-ROM drive (and in the CD-R writer as well). CD-R writers can write entire discs at once, or they can write a portion of a disc called a "session". In addition, the CD-R standards provide for the writing of small segments of data, e.g., a single CD-ROM sector, in one writing operation; this is called "packet writing". When appending any new information to a disc (i.e., when performing session or packet writing), however, a CD-R writer must always add the new information directly to the end of the already written portion of the spiral data track on the disc. Moreover, in packet writing, at least four "link sectors" (and usually seven to eight sectors, in practice) that contain useless (padding) data must be appended to the sectors of user data that one wants to record. These recording characteristics (i.e., sequential appending to the previously written portion of the data track and link sector overhead) result directly from the nature of the CD-ROM recording format and the underlying CD-Audio recording format.

High performance computer data storage peripherals, otherwise known as Direct Access Storage Devices (DASDs.), have recording formats that enable them to operate in a manner that is consistent with the way computer operating systems handle files. In particular, the recording formats used by DASDs cause all bytes that comprise a specific sector to be contiguously recorded along a continuous segment of the data track on the storage medium and further cause sectors to synchronously occur along the data track so that DASDs know the exact physical location of every sector recorded on their storage medium. Moreover, a DASD storage medium is subdivided into sectors prior to writing file data to it (this is done via a process known as "formatting"). Thus, a DASD can write, or read, a single sector as an independent unit and it can locate a sector anywhere on its storage medium, regardless of how much of, or what portion of, the medium is already written. These operational features allow fast file access (e.g., only a single sector might have to be rewritten if only a small part of a file is to be updated) and they are critical to overall data reliability (sectors that begin to experience data recovery errors, as reported by the DASD error correction sub-system, are retired and their contents rewritten to a new location on a portion of the storage medium that is known to be error-free).

The use of CD-writers to produce small numbers of discs that can be distributed to business and/or consumer computer users (who have a CD-ROM drive installed in their computers) is an important emerging application. However, the incorporation of CD-writers into personal computers and work stations is being impeded by the fact that they cannot perform DASD-like operations, i.e., the limited usefulness of a CD-writer makes it a very expensive peripheral from the perspective of a general user. One attempt at solving this problem is the Power Disc (PD) optical disc system recently introduced by Panasonic, which can read any compact disc (i.e., a disc that conforms to the standards for CD Audio/ROM discs) and which, in addition, will operate as a DASD. When operating as a DASD, the PD drive uses a proprietary recording format. Two drawbacks of the PD system is that it cannot use standard recordable CD discs when operating in the DASD mode and it cannot write compact discs that can be read on standard CD-Audio or CD-ROM players. The PD drive uses a proprietary disc and recording format when operating in DASD mode, i.e., it cannot write at all using a standard CD-R disc, nor can it write using the soon-to-be available CD-erasable, or CD-E disc.

An important problem to be solved, therefore, is to provide a CD-device that can write/read information in all standard CD recording formats and which has the additional capability of operating as a direct access storage device (DASD), and to do this using common CD components (i.e., conventional CD hardware and discs).

SUMMARY OF THE INVENTION

It is an object of the present invention to provide a compact disc (CD) with direct access storage device (DASD) capability (i.e., a CD that has information recorded using a CD-DASD recording format) and to provide the system for reading and writing such a CD.

It is another object of the present invention to provide a system compatible with prior CD standard formats.

It is also an object of the present invention to provide a system that uses existing standard CD-audio/CD-ROM components and a small number of additional low cost components in providing the CD-DASD capability.

It is a further object of the present invention to use the basic Reed Solomon (RS) error correction codes of the CD-Audio format in a DASD format.

It is also an object of the present invention to utilize the eight-to-fourteen (EFM) modulation scheme and the 588-channel bit frames utilized in the CD-Audio format in the CD-DASD format (that is, the low-level physical manifestation of the CD-Audio format is not altered).

It is another object of the present invention to accommodate variance in switching times between modes among different CD-DASD drives.

It is an additional object of the present invention to provide a system that improves the access time to recorded data by allowing reading or writing while the disc is changing speed.

It is an object of the present invention to preserve the Control and Display Subcode channel in CD-DASD format.

It is another object of the present invention to provide a means of decoding the RS error correction codes that will provide high recovered data reliability and enable fast access to certain recorded information fields (such as sector ID fields).

The above objects can be accomplished by a system that redefines how logical data is distributed on the compact disc (CD). The redistribution produces a DASD-like format that features a writable (or re-writable) CD that is formatted. The system uses the components of a conventional CD reader/writer (including conventional writable/re-writable CD discs) and a mapping or translation controller to alter the data byte interleaving employed in the conventional Cross Interleaved Reed Solomon Code (CIRC) coding used in the CD-Audio format. A rectangular product code is formed using the C1 and C2 CIRC subcodes. This product code can be interleaved to mitigate the effects of user handling of the disc. The system also provides synchronous voltage-frequency oscillator (VFO) fields for locking a write/read channel clock to the changing data frequency that may occur when radial disc seeks are performed. This feature will assist data reading and writing while the CD is acquiring proper rotational speed (assuming a constant linear velocity system).

These, together with other objects and advantages which will be subsequently apparent, reside in the details of construction and operation as more fully hereinafter described and claimed, reference being had to the accompanying drawings forming a part hereof, wherein like numerals refer to like parts throughout.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 depicts a typical system in which the present invention is used;

FIG. 2 illustrate conventional CIRC coding and interleaving;

FIG. 3 illustrates a conventional CIRC decoder;

FIG. 4 illustrates a product code according to the present invention;

FIG. 5 illustrates the format of the present invention without interleaving;

FIG. 6 illustrates one form of interleaving according to the present invention;

FIG. 7 depicts a structure of a CD-DASD sector;

FIG. 8 depicts user definable data and parity in a sector;

FIG. 9 depicts the steps of a read process according to the present invention;

FIG. 10 illustrates a write process;

FIG. 11 depicts a circuit for reading the data written by the process of FIG. 10;

FIGS. 12 and 13 depict the addressing scheme of the circuit of FIG. 3;

FIG. 14 depicts a circuit that separates write data from other data;

FIG. 15 depicts a C1 write--read scheme according to the present nvention;

FIGS. 16 and 17 depict a C2 write--read according to the present invention; and

FIG. 18 illustrates a circuit for realizing a bit clock which tracks rotation speed during reading and writing.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

The present invention is directed to methods and apparatus for using a CD-writer/reader as a direct access storage device (DASD) within the confines and context of the physical (recorded marks and lands) and logical (eight-to-fourteen modulation, CIRC error correction coding, multiplexed data/control and display channels, 588-channel bit EFM frames, etc.) data recording format that is used in all currently defined CD systems (e.g., CD-Audio, CD-ROM, CD-i, CD-R). The present invention defines a recording format that (i) enables full support of DASD operation and (ii) which is realized by redefining how the logical 33-byte EFM-frames (i.e., the frames that are created by the CIRC block encoding process that is defined in the CD-Audio recording format) are distributed on the disc. The invention redefines which physical marks on the disc will represent each of the data in these logical 33-byte EFM-frames. For convenience, this recording format is referred to as the CD-DASD recording format, or simply the CD-DASD format. The present invention allows reading of a CD containing CD-DASD formatted information via a CD-ROM player that has had only minor modifications made to. The present invention also allows recording of information using the CD-DASD format on conventional writable-CD discs (CD-R or CD-E discs) that are formatted prior to their use. During the formatting operation sector headers are written globally over the entire disc (or over an annular portion that is allocated to CD-DASD use). Formatting a disc is accomplished using a CD-DASD-writer; this writer writes headers in alignment with the "absolute-time-in-pregroove" (ATIP) information channel that is extant on standard CD-R discs. Alternatively, the sector headers could be embossed (or otherwise formed) on the disc during its manufacture. The physical marks on the disc that constitute the sector headers are formed using channel data sequences that maintain the 2,10 run-length constraints that characterize the physical marks of the low level CD recording format. The present invention organizes the data into the CD-DASD format and incluaes remapping of the 33-byte EFM-frames to physical marks on the disc, reorganization of the C1 and C2 subcodes that comprise the CIRC code into a rectangular product code that can be interleaved to depth seven or less, specification of a 4,704-byte CD-DASD sector and the identification of the various data fields contained in the CD-DASD sector.

The present invention is applicable to the typical computer system, as illustrated in FIG. 1, that includes a processor 10 having the appropriate disk and RAM storage, a display 12 and an input/output device 14, such as a keyboard, although all of these components may not be necessary in a particular application. In such typical systems one of the important mass storage components is a compact disc (CD) drive 16 that is capable of reading (CD-ROM) and/or writing (CD-E, CD-R) data on an optical or optomagnetic compact disc 18. The present invention is involved in the operation of the disc drive 16.

Current compact disc systems, such as the conventional CD-Audio or CD-ROM systems, use a Cross Interleaved Reed-Solomon Code (CIRC) to encode the user data bytes. This error-correcting code employs two distance 5, Reed-Solomon codes: C1(n1, k1) and C2(n2, k2) with n1=32, k1=28, n2=28, k2=24 bytes. The encoding process creates 33-byte Eight-to-Fourteen Modulation (EFM) frames which each have the format of:

F.sub.0,D.sub.1,D.sub.2,D.sub.9,D.sub.10,D.sub.17,D.sub.18,D.sub.3,D.sub.4, D.sub.11,D.sub.12,C2.sub.0,C2.sub.1,C2.sub.2,C2.sub.3,D.sub.5,D.sub.6, D.sub.13,D.sub.14,D.sub.21,D.sub.22,D.sub.7,D.sub.8,D.sub.15,D.sub.16,D.su b.23,D.sub.24,C1.sub.0,C1.sub.1,C1.sub.2,C1 (where F is a Control and Display byte, D are user data bytes, C2 are C2 code parity bytes and C1 are C1 code parity bytes. The conventional encoding and recording process performs the following steps (cf., FIG. 2):

step 1: User information (i.e., input data to be recorded) is first parsed into 24-byte blocks or user-data frames.

step 2: The 24-byte user data frames are scrambled and then 24 bytes (comprised of a first group of twelve contiguous bytes and a second group of twelve additional contiguous bytes that occur 48 bytes later in the sequential stream of scrambled data) are C2-encoded, that is, 4 parity bytes are calculated and provided to each block of 24 scrambled and delayed bytes to form a 28-byte C2 codeword 30.

step 3: The individual bytes of every C2 codeword are delayed for a variable number of words. These variable length delay lines provide the "cross-interleave" feature of the encoding.

step 4: Next, 28-byte groups (one byte from each of 28 different C2 words) are sent to a C1 encoder which generates 4 additional parity bytes and appends them to the 28-byte groups. The result is 32-byte C1 codewords 32 at the output of the C1 encoder.

step 5: The odd bytes of every C1 codeword are delayed for one C1 codeword to produce an additional interleave of depth 2. Next, one byte 34 of Control and Display (C&D) information is added to every 32-byte group appearing at the output of the depth-two interleaver to form the 33-byte EFM (Eight-to-Fourteen modulation) frames.

step 6: Channel data of the Compact Disc must obey the (2,10) Run-Length constraints, that is, there must be at least 2 and at most 10 zeroes between two consecutive 1's in the stream of channel data bits. EFM modulation coding accomplishes this by converting each of the EFM frame bytes into 14 channel bits that conform to the (2,10) RLL constraints. In addition, 3 link bits are added between pairs of such 14-bit words before they are merged. These link bits are necessary to ensure that the run-length conditions continue to be satisfied and to keep the DC content of the NRZ pulse read/write waveform, formed from the channel data stream, as low as possible.

step 7: Finally, 27 synchronization bits are added to the beginning of each EFM-encoded EFM-frame before the channel bit stream is recorded on the disc sequentially, frame by frame, as shown in the dataflow 36. This modulation scheme converts every 33-byte EFM-frame into 588 channel bits: [(33 bytes/EFM frame.times.17 channel bits/byte)+27 Sync. bits]=588 channel bits/EFM frame

step 8: The 588 channel bit representations of the EFM-frames are sequentially recorded on the disc. This is accomplished by using the nonreturn to zero (NRZ) pulse waveform that corresponds to the channel data stream to turn the writing laser on/off, thereby causing the sequence of marks/spaces. (or pits/lands) which comprise the disc data track to be formed.

The CIRC encoding process is shown in FIG. 2. In this encoding scheme, contiguous user data frames are not organized into groups of bytes that are stored together on the disc. In fact, the 24 bytes of a specific single user data frame are distributed over 106 sequential EFM-frames. Moreover, the CD-ROM uses 2352 byte sectors that consist of 98 contiguous 24-byte user data frames; CIRC encoding of 98 contiguous 24-byte user data frames causes these to be dispersed over 208 consecutive EFM-frames. In addition, individual C1 and C2 codewords can be comprised of data that belong to different CD-ROM sectors.

To recover a single 24-byte user data frame from the data stored on a conventional CD in the format shown in FIG. 2, a) 111 contiguous EFM frames must be read from the disc, b) the C1 words contained in the EFM frames must be decoded to obtain three consecutive C2 codewords, and c) two of these C2 codewords must be decoded to recover the 24-byte user data frame of interest. Once the first user data frame has been recovered, subsequent user frames may be obtained by reading additional EFM-frames one at a time; the recovery of every additional EFM-frame enables the recovery of one more C1 word, one more C2 word and one more user data frame. Thus, at least 111.times.33=3663 bytes of data, mostly unrelated to the user data frame of interest, must be read from the disc before the first 24-byte user data frame can be recovered. It should also be noted that if the user frame contains the initial 24 bytes of the recorded information, an additional 100-200 EFM frames of pad data (usually all zeroes) must be recorded immediately prior to the first EFM frame to "prime" the C1 and C2 decoders. This pad data is necessary to produce complete codewords at the input of the decoders when the disc is being read.

A conventional decoder 50, such as the Signetics SAA7310 Decoder, along with a conventional RAM 52, as illustrated in FIG. 3, is used to perform the CIRC (CD-Audio) decoding (i.e., invert the encoding process described above). The input to the EFM demodulator 54 consists of (2,10)-constrained RLL digital data in the form of 14-bit symbols grouped together as 33-symbol frames as previously described. (The 27 channel synchronization bits and thirty-three groups of 3 link bits contained in each 588-channel bit representation of an EFM-frame have been removed by an earlier processing step). These frames contain 32 information (user data) and parity symbols plus one Control and Display (C&D) symbol. After eight-to-fourteen (EFM) demodulation is performed, the subcode processors (not shown) strip off the C&D byte 34 to extract the C&D section timing/address information. The remaining 32 bytes of the frame plus erasure flag information are written to the RAM 52 during the "Write 1" cycle. The EFM-demodulator flags each output byte that occurs in correspondence with a 14-channel bit word that contains a (2,10) RLL-constraint violation; such flagged bytes are treated as being erroneous (i.e., they are erased) by the C1 decoder. The internal processor 56 of the decoder 50 provides the address locations as well as Read/Write timing control for the data written to the RAM 52. The C1 codewords are formed and fed into the C1 decoder 58 during the "Read 1" cycle. The internal processor 56 provides the address values for individual bytes that are retrieved from the RAM 52 during the "Read 1" process. These addresses are different from those that were accessed during the "Write 1" cycle and effectively, by writing the EFM frame bytes into one set of RAM locations and reading the C1 frame bytes in a different order, i.e.; from different locations, the required depth of 2 C1-word de-interleaving is accomplished. The C1 decoder 58 performs error correction/detection on the incoming 32-byte frames (C1 codewords) and discards 4 parity bytes before writing the remaining 28 bytes and new flag information to the external RAM during the "Write 2" cycle. The "new" flags are assigned by the C1-decoder to each byte at its output; the flags indicate the reliability of the decoding operation that the C1-decoder performed when the specific bytes were processed by it. This new flag information is subsequently utilized by the C2 decoder. The C2 codewords are then input to the C2 decoder 60 by reading 28 bytes from the external RAM during the "Read 2" cycle. The address values generated during the "Write 2" and "Read 2" cycles provide the cross de-interleaving that is necessary for extracting the C2 codewords.

There are two features of interest associated with the above decoding/de-interleaving architecture which are relevant to the present invention:

a) C1 and C2 codeword bytes are written into an external RAM and the data values as well as locations (addresses) of the read/written bytes within the RAM can be monitored. Thus, it is possible to intercept and modify the address values as they appear across the RAM bus.

b) The external RAM is logically divided into two parts. One half of the RAM is exclusively used for C1 de-interleaving and the other half is dedicated to C2 de-interleaving. This makes it possible to modify the addressing scheme in one half of the RAM without affecting the performance of the other half.

One aim of the present invention is to adapt the above de-interleaving architecture to provide the proper block retrieval of the data recorded in the CD-DASD format described below.

The present invention, as previously mentioned, provides a CD recording format that organizes related user data frames into groups (sectors) and stores them sequentially and contiguously on the disc 18, exactly as in a DASD recording format. This format is realized by restructuring the CIRC coding scheme (i.e., by changing its interleaving and scrambling scheme) such that the C1 and C2 codewords form a distance 25, rectangular product code. One such product codeword 68 is shown schematically in FIG. 4. In this figure, C1 codewords 70 are shown in columns and C2 words 72 occupy the rows. The product codeword shown in FIG. 4 belongs to a 28.times.28 product code type such that the twenty-eight columns are comprised of twenty-eight individual 32-byte C1 codewords while the upper twenty-eight rows are comprised of twenty-eight 28-byte C2 codewords. We note that the bottom four rows, which contain only the parity check bytes of the twenty-eight C1 codewords are not encoded as C2-words in FIG. 4. Thus, only the upper 28 rows of FIG. 4 contain actual C2-parity bytes.

The CD-DASD format of the present invention enables data to be written/read in blocks of fixed size and encoded/decoded accordingly. Much like the current magnetic disk recording formats, each block (sector) can have a pre-recorded sector header and defect management techniques such as, sector retirement/relocation, can be used to enhance the reliability and prolong the life of the storage media. The present invention essentially uses the same circuitry to write/read either the CD-DASD or conventional CIRC-CD recording formats on a given disc 18. It is also possible to make the format session-specific when a multi-session disc is being used. That is, part of the disc 18 can be written in one format and a different part in a different format. Since the same C1 and C2 distance 5, Reed-Solomon codes are used in both the CD-Audio/ROM and the CD-DASD recording formats, at least some decoders (chipsets) that exist in current CD drives, (cf., the previous discussion), can be modified via external logic to accomplish the decoding of data recorded in the CD-DASD format.

It is the intention of the present invention to take advantage of the standard read/write equipment that is implemented in the current CD drive systems. This means the format of the physical information written on the disc 18 will remain the same. More specifically, the same 33-byte structure of the EFM frames is implemented in the new block encoding format and each EFM frame is represented by 588 channel bits. The present invention does all that is necessary to realize a product error correction code (ECC) based, DASD recording format via logical remapping of (repositioning) the 33-bytes that comprise each of the EFM frames during writing and reading of the disc. At a higher logical level, the content of some of the user data which occur in one or more of the 24-byte input data frames will also be defined. For example, some of these data bytes may carry synchronization or sector address information.

The implementation of the DASD recording format at the logical level(s) makes it possible to use the current writable CDs for DASD recording. In a conventional disc (which uses the ATIP time code for addressing), the minimum addressable length along a data track is 98 EFM frames (or, 98.times.588=57,624 channel bits). Each C1/C2 product codeword, as depicted in FIG. 4, occupies an equivalent of 28 EFM frames. Thus, 98/28=3.5 product codewords can be placed in a 98-EFM frame track segment of the disc 18 track and 7 product codewords can occupy exactly two contiguous 98-EFM frame track segments. It is, therefore, possible to use 7 C1/C2 product codewords to define the preferred read/write sector size. Such a sector contains 2.times.98=196 EFM frames that carry all sector synch, address, CRC, etc., fields, as well as user data. This information is logically mapped into 24.times.196=4,704 data bytes. If, in compliance with the logical sector layout of the CD-ROM Mode 01 recording format, 2.times.2048=4096 user data bytes are placed in such a sector (equivalent user data content of two logical Mode 01 CD sectors), 608 "extra" bytes will be available to carry the synch, address, etc., fields. In addition, a third level of ECC can be implemented and the parity bytes for such a code can also be incorporated in the remaining "extra" bytes. As an example, if the 4096 user data bytes are encoded as 32 interleaves of a (144,128) distance 17 Reed-Solomon code on GF(256), there will be [(144-128).times.32]=512 parity bytes generated for the resulting 32 codewords. This leaves 608-512=96 of the "extra" bytes for sector synch, address, etc. The implementation of such a third level ECC could provide the ultimate reliability for the data retrieved from the disc 18. Note: Only a few, if any, of the 608 "extra" bytes will be needed to implement sector resynch fields since resynch is already provided by the 27-channel bit synchronization fields that start each 588-channel bit EFM frame written on the disc.

Overhead is not increased by switching from the conventional CIRC recording format to the CD-DASD format. This can be shown by calculating the user-to-gross total byte utilization ratio's for the two formats: ##EQU1## Selection of the Block format sector layout in the above manner, dictates the above utilization ratio's to be always equal. The fact that the invention places 7 C1/C2 product codewords in one such sector, makes it possible to implement product code interleave depths of up to 7 product codewords.

The features of the CD-DASD recording format of the present invention are summarized below:

Physical recording format (marks written to the disc 18) is unchanged--588 channel bit structure used to represent each EFM frame.

Disc addressing structure (C&D channel and 98-frame C&D Blocks) is preserved--conventional means such as the "absolute time on disc" information contained in the control and display subcode q-channel may be used to physically locate sectors.

The entire contents of a sector are contiguously recorded on the disc track--DASD-like read/write, etc., operations are enabled.

Overhead is identical to that of CD-ROM.

Use of multiple decoding of C1/C2 subcodes and powerful third level RS ECC may provide increased data reliability compared to CD-ROM.

Conventional CD-R discs and write/read electronics (and perhaps unmodified CIRC decoding chips) can be used--modified CIRC decoders are necessary if multiple pass decoding is required.

The ATIP signals located on conventional writable CDs may be used to "format" (i.e., write sector headers onto) CD-DASD discs. Random writing of sectors is enabled if "formatted" CD-DASD discs are used.

FIG. 4, previously discussed, shows a product code 68 with distance 5.times.5=25 which contains 32.times.28=896 bytes (24.times.28=672 user bytes). In this figure, no interleaving has been indicated for the C1 words. In the case where various depths of interleave are utilized, the data block of interest would contain n.times.896 bytes where n=2, 3, 4, . . . is the possible depths of interleave for the C1 words. The depth of interleave is 32 bytes for the C2 words in the product code illustrated in FIG. 4 (33 bytes if the C&D byte is taken into consideration) when the 896 bytes that comprise a code word are written to the disc in a column by column fashion. The recording format which employs the FIG. 4 product code 68 is illustrated in FIGS. 5 and 6. In FIG. 5 the product code 80 is implemented without interleaving, or equivalently, with depth 1 interleaving. FIG. 6 shows interleaving of two product ECC codewords by alternating the recording of their columns, or equivalently, depth 2 interleaving. The result is a block of 56 EFM frames that contain the data from two product codewords. Adjacent EFM frames of this block contain the data from one complete column of each product codeword. Other schemes for interleaving the product codewords are possible.

The CIRC deinterleaving required by the present invention, as will be discussed in more detail later, is accomplished by writing the data bytes to an external RAM and reading them from RAM locations in a different sequential order than that used by the conventional decoding process previously discussed with respect to FIG. 3. It is also possible to bypass the RAM deinterleaving cycle and use a secondary RAM chip to read/write the bytes in the specific sequential order that is required for the block decoding (i.e., to construct product codewords) of the present invention.

The present invention is also suitable for multiple-pass decoding. Recall that one user data frame, in CIRC format, is spread over 106 EFM frames. This long depth of interleave reduces C2 decoding failures that are due to relatively long bursts of error. This protection against error bursts can be accomplished in the CD-DASD format by using product codeword interleaving and the ability to handle long error bursts can be further improved via multiple-pass decoding. In the multiple-pass decoding which is performed in the present invention, after the initial C1 and C2 decoding stages, C1 decoding is repeated. This may be followed by another C2 decoding and the cycle may continue until the decoding performance consistent with a distance 25, 32.times.28 product code is achieved. The data reliability achievable from such a code may be equivalent to, or greater than, that achievable via the conventional CIRC depending on the nature of the errors which contaminate the data. The cooperation between the C1 and C2 decoders is conventionally accomplished by passing information flags which are generated after each level of decoding. The multiple C1 and C2 decoding requires the implementation of decoding strategies which dictate the number of errors and erasure corrections in each decoding pass. Various conventional decoding/flagging strategies and product codeword interleave combinations can be used to optimize the decoding performance.

The error handling capability of conventional CIRC decoders is also enhanced by supplying erasure information from an outside source. Specifically, in many current compact disc read channel implementations, the EFM demodulator flags (i.e., erases) all output data bytes that are derived from channel data that violates the 2,10 run-length constraints that are inherent in the EFM modulation code. (The C1 decoder can correct 2.times. the number of erased erroneous bytes as non-erased errors, so long as the erasures are determined by a source external to the decoder.) This feature is fully retained by the decoder of the CD-DASD product code. However, due to the different interleaving structure of the product code that is defined herein for the CD-DASD recording format, and because decoding strategies different from those employed by the C1 and C2 decoders of conventional CIRC decoders may be used in the implementation of CD-DASD product code decoders (especially if multiple-pass decoding is used), the CD-DASD format may take advantage of other external (to the decoder) erasure-flagging mechanisms. As an example, if the 27-channel bit synchronization field of an EFM frame is detected to be skewed, or decentered relative to the channel synchronization field detection window, the 32 bytes corresponding to that EFM frame might be flagged as of low quality (and such flags may be different from those set by the EFM demodulator if "new" decoder circuits that recognize such differences are provided).

Another feature of the CD-DASD recording format is its enablement of a "fast read" access to data. This feature is implemented by allowing the controller to access data bytes that appear at the output of the C1 decoder before any C2 decoding takes place. Referring to FIG. 4, fast read would allow access to the 24 data bytes in each column of the product codeword immediately after C1 decoding of the column. This feature would allow the CD-DASD drive controller to access information recorded in the sector header prior to reading or writing the remainder of the sector (this feature would allow determination/verification the sector ID, for example). The information contained in the non-header portion of the sector would pass through both C1 and C2 decoding (or through multiple C1/C2 decoding when implemented) before being passed to the controller.

Before describing the steps required record to information on a standard CD using the CD-DASD recording format, we must describe the structure of the CD-DASD data sector 88. This is necessary because the details of the encoding process that is used to write information to a CD-DASD disc will depend on this structure. It must be appreciated that the sector described below is a representative CD-DASD sector in the sense that it contains the various data fields needed to insure reliable data recovery under the constraint of maintaining the a high degree of compatibility with the logical and physical CD-Audio/ROM recording formats; the actual content of the some of the data fields within this sector may be somewhat changed in order expand functionality, or increase the appeal of the CD-DASD recording format as a standard for future CD systems.

We will first describe the physical channel structure. A typical example CD-DASD sector 88 is preferably comprised of four major sections or areas, namely a header 90, a preamble 92, data/ECC parity 94 and a buffer 96 areas as shown in FIG. 7. The total sector is physically recorded in a segment of disc track that holds 196 contiguous EFM frames. The number of EFM frames allocated to each sector area is also indicated in FIG. 7. The actual contents of each of the 196 EFM frames which physically represent (constitute) the recorded sector on the disc track are described below.

The four EFM frames which comprise the header 90 are written when the disc is formatted. Formatting is a separate process which prepares the disc for use in the CD-DASD storage system (in effect, the formatting process converts a standard CD-R/E disc into a CD-DASD disc). The header areas 90 of all sectors 88 of the disc (or in the annular region of the disc that is to be used for CD-DASD recording) are written and optionally verified during the formatting process. Some specific disc directory and file management information (e.g., the disc's Volume Descriptor field and Boot Record) is also written into the data/ECC parity areas 94 of appropriate disc sectors when the disc is formatted. The header portion of sectors 88 are never partially written; those sectors which have any information written into their data/ECC parity areas during the formatting operation are completely written when the disc is formatted, i.e., the entire header 90, preamble 92, data/ECC parity 94 and buffer 96 areas are written, according to the rules described in the sequel, when the disc is formatted. Note also that CD-DASD formatting could occur as a two stage process. Low level formatting would cause only sector headers and perhaps physical disc information such as a bad sector map and manufactures identification to be written. Subsequent high level formatting would cause information to be written into particular sectors that specializes the disc for use via a particular operating system.

When writing a header area 90 during the disc formatting process, the actual physical marking of the disc occurs in synchronism with the absolute-time-in-pregroove (ATIP) information that is carried in the spiral groove of a conventional writable-CD disc. That is, the location on the disc of the start of the 27-channel bit EFM frame sync pattern that begins the first EFM frame of every header will have a constant offset from the start of the sync pattern of the nearest ATIP word contained in the disc groove. The recording of the header 90 shall start by writing the second of the two EFM frames that comprise the buffer area 96 of the previous sector, that is, the sectors are spliced together.

We have defined that the splice between sectors occurs in the middle of the buffer area, i.e., at the point where the first EFM frame of the buffer 96 ends and the second frame of the buffer 96 begins. The splice at this point maximizes the tolerance that specifies where the splice must occur. For example, having the splice occur in the middle of the buffer area 96 means that the exact location of the splice can be in error by +/-0.95 EFM frames and the splice will still occur within the buffer (and thus will not contaminate the data/ECC parity area 90 of the sector 88 or the header area 90 of the next sector). With the location of the splice held to <+/-0.45 EFM frames, then we can define the position of the splice to be in the center of the first EFM frame of the buffer (without the danger of contaminating the data/ECC parity or header areas' data). Locating the splice in the center of the first EFM frames of the buffer 96 would prevent the splice from contaminating the EFM sync field of the second EFM frame of the buffer 96 (and thus will insure that all 196 EFM sync fields (one for each EFM frame) of the sector 88 will be found. That increases the robustness of the channel sync maintenance. In addition, by placing the splice in the center of the first EFM frames of the buffer, the CD-DASD drive will have 1.5 EFM frames (instead of 1.0) of sync field that is phased with the header to read prior to encountering the header. With the splice defined to occur in the center of the first buffer EFM frame, then the formatter writes 1.5 frames of the buffer 96 together with the header 90 of the next sector when the disc is formatted.

To maintain compatibility with the "incremental recording" linking rules described in the CD-Write-Once specifications (i.e. in the publicly available Sony/Phillips "Orange Book"), which state that the initial EFM frame that is written in any new instance of disc recording shall be the 26th EFM frame of a 98 frame C&D block, the specific C&D byte value to be recorded in the second buffer EFM frame shall be that of the 26th byte of the "current" 98-byte C&D channel word. The current C&D word is the one that contains the absolute-disc-time (ADT) value which is the same as the ADT value of the most recently read ATIP word (this will usually be the ADT of the previous ATIP word on the disc groove since the entire disc will generally be formatted in one sequential operation). The start locations of the second EFM frames of all buffer areas (i.e., the start of the 27-channel bit EFM frame sync patterns of these frames) on a CD-DASD disc shall be (TBD) EFM frames .+-.0.5 EFM frame from the end of the sync field of the nearest previous ATIP word (this normally is the ATIP word that has an ADT value that is 1/75th second higher than the ADT of the current C&D word). A description of the actual content of the second buffer EFM frame is given below.

The four EFM frames that constitute the header 90 shall begin with a 27-channel bit EFM sync pattern that is directly followed by the 17-channel bit sequences (including the 3 merging channel bits) that correspond respectively to the 27th, 28th, 29th and 30th Control & Display (C&D) bytes of the current C&D word. The contents of the remainder of each of the four header EFM frames shall be as follows:

       HEADER FRAME 0 (last 32 bytes/544 channel bits):
           14 replications of the 17-channel bit
    pattern...00100000000001001...;
            6 replications of the 3-byte pattern..47h; F2h; A8h.

    ....001001001001001001001001001001001001001001001001...
        .vertline.   .rarw. 47H .fwdarw.  .vertline.   .vertline.   .rarw.
    F2h .fwdarw.  .vertline.   .vertline.   .rarw. A8 .fwdarw.   .vertline.

    ______________________________________
    Write 2
    Input Address  Output Address
    ______________________________________
    E1; 0C, 2C, 4C 01; 00, 20, 40
    .              .
    .              .
    .              .
    01; 1C, 3C, 5C FB; 0C, 2C, 4C
    ______________________________________