Magnetic tape apparatus for recording and reproducing of signals

The invention is based on a magnetic tape device according to the preamble of claim 1.

It is known from a matrix head EP 89401125.3 to record large amounts of digital data on a magnetic tape with an extremely high bit rate, ie in a very short time. Thanks to the matrix head, which enables the recording of a large number of parallel tracks without turf, the same recording density can be achieved with the fixed head as with helical track recording.

Since the relative speed between the tape and the head is low in the case of longitudinal track recording with a large number of parallel tracks, the scanning must be carried out during playback with transducers which supply a speed-independent signal voltage. These are, for example, magnetoresistive heads, or heads that use the magneto-optical core effect. A multi-parallel track scanner in which a laser beam is modulated by the magnetization of the adjacent tracks and projected onto a CCD line in such a way that each pixel of the CCD line receives the light modulated by the magnetization changes of a single track is described in EP 89401125.3.

The number of pixels on the CCD line can also be greater than the number of tracks scanned at the same time, so that the adjacent tracks are oversampled with the CCD line. The signals of the individual tracks can then be separated by digital filtering without exact tracking being necessary. The coarse adjustment to a certain track area is made possible in that the CCD line is designed in such a way that it covers an enlarged track area and the desired track area is selected by electronic means.

The accuracy requirements for the head shifting mechanism required for the selection of different track areas can thereby be significantly reduced.

The signal voltage clocked out from the CCD line represents the binary values ​​of the bit cells lying next to each other in the parallel tracks. The originally recorded signal can be restored by means of suitable synchronous information by means of digital signal processing.

The invention has for its object to use these recording and playback options for a handy magnetic tape device. This object is achieved by the features of the invention specified in claim 1. Advantageous developments of the invention are specified in the subclaims.

The device according to the invention is intended for the storage of very large quantities of audio signals which are recorded as digital signals after data reduction. The device is to be used as a portable playback device and, with dimensions that are only marginally larger than a commercially available magnetic tape cassette, allows access to pieces of music that correspond, for example, to the content of approximately 200 CDs, without changing the tape. The advantage of the device according to the invention is the possibility of recording large amounts of data in a very short time, since, for example, 512 parallel tracks are recorded simultaneously with the matrix head, and a very high bit rate can be transmitted for each individual track.

However, this rapid recording can only be carried out with special external recording devices, since the portable magnetic tape device can neither generate the recording signal at the required bit rate nor is it able to supply the energy for the rapid tape transport.

It must therefore be ensured that the users of the magnetic tape devices according to the invention have the possibility, for example, of storing the contents of a selection of CD's in their device at a special recording station for a fee. In the event of a data reduction that does not yet result in a noticeable loss in quality, the content of a CD can be accommodated on a 1.5 m 1/4 '' tape, so that a tape length of approx. 300 m is required for 200 CDs.

At a tape speed of 5 to 6 cm / s, recording the tape length of approx. 1.5 m required for the content of a CD requires less than 30 s. The signal of the CD's to be recorded is stored in a mass memory after data reduction, division into signal sections and temporal rearrangement of these sections. The temporal rearrangement of the signal sections is necessary because the playback head only a part of all parallel tracks, ie 1/8 of 512 ie 64 Traces, scanned. As a result, the effort for the playback head is reduced and the data rate during playback is reduced. Such a simultaneously scanned track area is referred to later as a macro track in this text. Recording the contents of a CD in a limited band area enables quick access within a CD.

If the device according to the invention is also equipped with its own matrix head for the recording, besides the quick recording at the recording station, it is also possible to make own recordings in real time and recordings by copying from device to device. It makes sense to set up your own recording head for the same number of tracks as the playback head. If, for example, both heads for 64 tracks, ie 1/8 of the tracks on the tape are set up, then copying from device to device requires 8 times the fast recording time on the external recording device. The copying of a complete CD, including the return times required between the individual macro tracks, then takes about 5 minutes.

Since the tape speed cannot be reduced arbitrarily during playback, a start / stop operation of the tape is provided. The time-compressed signal sections sampled here are converted with the aid of a buffer memory into the continuous signal flow required for sound reproduction.

Real-time recording must also be done in start / stop mode. The buffer memory is used for time compression of the continuous audio signal, which is then recorded in time-separated sections.

The invention is explained below with reference to the drawings. In it show:<dl id="dl0001" compact="compact"><dt>Fig. 1</dt><dd>rapid recording at the recording station,</dd><dt>Fig. 2</dt><dd>a block diagram for recording in real time,</dd><dt>Fig. 3</dt><dd>a block diagram for the quick copy function</dd><dt>Fig. 4-11</dt><dd>Scanning options for different recording formats.</dd></dl>

1 shows a representation of the data transmission from an external recording device to the magnetic tape device MP according to the invention. The magnetic tape device MP has a read head RH, a magnetic tape T and evaluation electronics AE. The recording device A contains a recording electronics RE, a mass memory M, a read head RH ', a magnetic tape T' and a write head SK.

The magnetic tapes are located on the MT or MT 'tape discs. The magnetic tape T or T 'moves in the longitudinal direction to the write head SK. To record the magnetic tape device MP, this is inserted into the recording device A. When the magnetic tape T 'is played, the data passes via the read head RH' to the write head SK, which records the data on the tape T moved in the longitudinal direction. Recording electronics RE controls the data to be transferred and the amount of data. The user can select the desired data beforehand via an input unit EA. After this process, the magnetic tape device MP can be removed. It is now possible to read out the data of the magnetic tape T by means of an evaluation electronics AE with the aid of the reading head.

2 shows the magnetic tape device MP according to the invention with its own recording electronics and recording head for recording in real time. The input signal, which comes from the digital output of a CD player, for example, is reduced in the encoder EN according to the MPEG process. The error coding in the ECC block and data interleaving in the memory ME enable the later error correction. The signal is then fed to the matrix head SK used for recording via the head driver D. During playback, the core head RH scans the parallel tracks of the tape. Laser light is modulated in the tracks depending on the magnetization and projected onto a CCD line. The signal clocked out of the CCD line is converted back into the data-reduced recording signal after A / D conversion in block AP by digital signal processing in blocks DP and ME and error correction and deinterleaving in blocks ECC and ME. With the help of the decoding circuit DE, the original input data stream is restored.

Fig. 3 shows the basic block diagram for the transfer from one magnetic tape device MPI to the second magnetic tape device MPII with the same recording format.

There are various recording formats for recording on the magnetic tape device MP, which are described below.

When playing back signals recorded on magnetic tape, it may be necessary to scan a coherent recording in short, time-separated sections.

This is the case, for example, when the relative movement between the tape and the heads must not fall below a certain speed value when scanning a longitudinal track recording with one or more parallel heads. The sampling-related bit rate can then be significantly higher than the bit rate required for signal processing and signal evaluation.

The signal is delivered time-compressed, so to speak, and must be expanded for processing with the aid of a buffer memory.

Because of the limited memory size, signals with a longer duration must be broken down into short, time-separated sections, the pause between the signal sections being determined by the processing speed.

For the scanning of the individual sections of a digital recording, the following two requirements have to be met, for example:<ul id="ul0001" list-style="none"><li>1. The successive sections should overlap one another, since the signal sections which can be evaluated must start and end with complete sync blocks for the sake of the signal assignment and lead-in areas are required for the settling of the signal processing circuit.</li><li>2nd The sync blocks must be numbered consecutively so that double evaluation of sync blocks due to the signal overlap can be avoided.</li><li>3rd The relative speed between the head and the tape is said to be constant in order to generate a constant sampling bit rate during the scanning of a section.</li></ul>

Fig. 4 shows the relative movements between the tape and the scanning head, which are to be carried out in order to meet the requirements. 4a shows the division of the tape recording into sections to be scanned one after the other. The overlaps are denoted by d. The arrows in lines b and c indicate the direction of the relative movements. The solid lines mark areas of the belt that are fed at a constant speed. The dashed lines represent relative movements with variable speed as a result of acceleration or deceleration. Waiting times can be inserted at the reversal points of the direction of movement u in order to adapt the scanning processes to the data flow required for signal processing.

The scanning process illustrated by FIG. 4 has the following disadvantages compared to continuous tape scanning:<ul id="ul0002" list-style="none" compact="compact"><li>1. Increased wear and tear and thus wear on the strip material and the mechanics used to generate the relative movement between the strip and the scanning head.</li><li>2nd Increased energy consumption</li><li>3rd Increased noise</li></ul>

As FIG. 5 shows, the alternating scanning of successive sections in the opposite direction halves the number of required back and forth movements between the tape and the head. This also halves the impact of the disadvantages mentioned.

The signals from the sections scanned in the reverse direction can be brought into the correct order with the aid of a memory in which writing and reading take place in the opposite direction. This rearrangement is advantageously carried out in the playback circuit. The sorting could also be done before the recording. However, this would mean that the type of scanning as well as the position and size of the sections is fixed.

On the other hand, normal continuous recording can be divided into sections in any way during playback.

In the following, a further development is presented in which the segmentation during playback is predetermined by the recording format:

Device for the magnetic recording and reproduction of information represented as digital data on a flat, preferably tape-shaped, recording medium, in which the magnetic recording is applied in the form of a track in the longitudinal direction of the carrier on a plurality of N separate (macro) tracks, each of which in turn is composed of M. processed tracks exist, whereby from the tape speed v, the smallest usable wavelength λ, and the number K of the simultaneously available data rate to be determined written or read tracks is considerably higher than the data rate required for the application (eg digital stereo sound). <tables id="tabl0001" num="0001"><table frame="all"><tgroup cols="2" colsep="1" rowsep="0"><colspec colnum="1" colname="col1" colwidth="78.75mm" /><colspec colnum="2" colname="col2" colwidth="78.75mm" /><thead valign="top"><row rowsep="1"><entry namest="col1" nameend="col1" align="left">Example:</entry><entry namest="col2" nameend="col2" /></row></thead><tbody valign="top"><row><entry namest="col1" nameend="col1" align="left">Beam width a</entry><entry namest="col2" nameend="col2" align="left">8th mm</entry></row><row><entry namest="col1" nameend="col1" align="left">Belt speed v</entry><entry namest="col2" nameend="col2" align="left">60 mm / s</entry></row><row><entry namest="col1" nameend="col1" align="left">Bit length a</entry><entry namest="col2" nameend="col2" align="left">0.0003 mm</entry></row><row><entry namest="col1" nameend="col1" align="left">Track width b</entry><entry namest="col2" nameend="col2" align="left">0.01mm</entry></row><row><entry namest="col1" nameend="col1" align="left">Number of tracks K</entry><entry namest="col2" nameend="col2" align="left">64</entry></row><row rowsep="1"><entry namest="col1" nameend="col1" align="left">Number of macro tracks N</entry><entry namest="col2" nameend="col2" align="left">8</entry></row></tbody></tgroup></table></tables>

This gives us a gross data rate of usable over a longer period of time <maths id="math0001" num=""><math display="inline"><mrow><msub><mrow><mtext>R</mtext></mrow><mrow><mtext>0</mtext></mrow></msub><mtext> = K * v / a = 64 * 60 / 0.0003 = 12.8 Mbit / s</mtext></mrow></math><img file="EP0762387A2_D0001.tif" /></maths>. The data rate required for recording digital stereo sound, on the other hand, is only R (CD format)₁= 2 * 44100 * 16 = 1.41 Mbit / s, or even with data-reducing coding (e.g. according to ISO / IEC 11172-3, Layer II) only R₂= 256 kbit / s.

As a result of this discrepancy, the recording and playback cannot take place in the known form of the continuous 'serpentine method'. In this case, the recording or playback is carried out along a macro track from the beginning of the tape to the end of the tape, then the recording and reproducing means are suitably positioned on a further macro track and then processed in the opposite direction from the end to the beginning. It is also possible to rewind the tape in the meantime and to carry out the processing in the further macro track from the beginning to the end. The additional winding process, however, costs energy unnecessarily, which is a great disadvantage, for example, in the case of a portable device.

On the other hand, both the difference between gross data rate and user data rate and the simultaneous presence of N macro tracks result in the possibility of performing a recording or playback with high time compression. By continuously editing a macro track alone, a time compression of up to R₀/ R₂, using the N macro tracks up to <maths id="math0002" num=""><math display="inline"><mrow><msub><mrow><mtext>NO</mtext></mrow><mrow><mtext>0</mtext></mrow></msub><msub><mrow><mtext>/ R</mtext></mrow><mrow><mtext>2</mtext></mrow></msub></mrow></math><img file="EP0762387A2_D0002.tif" /></maths> can be achieved. With the above values ​​you can, for example, reduce the duration of a recording or playback of the music from a 'real time' 74 min CD to values ​​under 30 seconds (taking into account additional data for error protection and transition zones).

Slower speed

First of all, there is the possibility of reducing the tape speed in accordance with the desired useful data rate. This is not recommended, since on the one hand problems with friction and vibrations occur at low belt speeds, but on the other hand a certain amount of time is required at the end of the belt to reverse, while the data is not continuously available. This could be compensated for by buffering and temporarily increasing the belt speed.

Intermittent recording / playback

The discrepancy between gross data rate and user data rate leads directly to intermittent operation with temporary storage, for example in a removable buffer or a continuous ring buffer. From the gross data rate calculated above and an appropriate size of such a buffer of, for example, 2 ... 4 Mbit, the length of a single block on the magnetic tape is, for example, 6 mm, corresponding to a time period of 0.1 sec or a data volume of 1.28 Mbit. This block can then correspond to the data for stereo sound with a real time of about 4 seconds . Intermittent operation means alternating between 0.1 sec active processing of a data block on the tape and about 3.9 sec time for positioning on the logically following block, which does not have to be the following spatially. Various suggestions for this logical arrangement can be offered for this. Each block should have a unique addressing, both the type of addressing and the playback function should be set up in such a way that the blocks can be identified even when the tape is running fast (rewinding). If reading and writing are provided in both tape running directions, then it may be expedient to provide the addressing data once at the beginning and once at the end of each block.

In the following, the various possible solutions are listed using figures.

<b>Version 1, serpentine mode</b>

Fig. 6 shows schematically a piece of tape with N = 4 macro tracks, each of which is divided into short blocks and between which transition zones are provided. The length of these transition zones is shown exaggerated. A block is also shown to be composed of M individual, parallel tracks in the longitudinal direction. Furthermore, the machining direction according to the serpentine principle is also indicated. It can be seen immediately that intermittent operation can only be implemented with great effort. If, for example, the tape run has to be stopped for 3.9 sec after reading block No. 1 for 0.1 sec, then either appropriately long transition zones must be provided within which the tape run can stop and start again (in Fig. 6 shown at macro track 2), or after each reading the tape must be rewound briefly so that the beginning of the following block can be read safely (shown at macro track 3 in FIG. 6). Of course, this also applies to a recording.

Version 2, zigzag fashion

Another possible embodiment is that the tape is rewound after each block so far that it is positioned again in front of the block, and then to be positioned (vertically) on the following macro track. This is shown in Figure 7.

Version 3, 'small' serpentine mode

The number of positions can be reduced if you use a 'small' serpentine mode, so to speak (see Fig. 8). It can be seen immediately that there can be several versions of the number of macro tracks and the treatment at the lower and upper reversal points. A version with an even number N is shown, in which the 'small' serpentine (blocks 5..8) following the first (blocks 1..4) is conveniently connected.

Version 4 continuous serpentine mode with intermittent operation

Another solution is to combine uniform tape motion with intermittent recording or playback. This is initially shown in FIG. 9. The band-head movement is again indicated by arrows, and the active blocks are marked by dots. A whole range of different versions are also possible here. It is important overall that either the entire available tape length can be processed accordingly, or that short pieces of tape can be viewed separately as a unit.

In Fig. 9, a piece of tape is considered, which corresponds to a length of 3 * 4 * 60mm = 720mm for the above parameters. The length of such a segment on the belt can, for example, be selected according to a specification for a maximum positioning time.

Description of the process for different operating modes:

Fast recording and / or playback:

Fast recording or playback can be achieved either across the entire width of the tape or within only one macro track if all the required data is available in a memory and if the blocks are scrambled within this memory. In this operating mode, the transition zones can be kept very small.

Real time recording on new or deleted tape:

It is characterized by a basically continuous tape run, a reversal of direction after a certain number of blocks, and the positioning of the next block to be written based on existing, already written blocks.

The recording begins at block 1. If the tape runs continuously, the position of block 2 is reached and written after the previously defined time (here, for example, 3.9 seconds). After writing block 3 (or a series of others), the tape direction is reversed and positioned on the next macro track. If the tape run is started again at the right time, then block 4 can be written approximately under block 3. This is repeated until the last block in the last macro track is reached (here block 12). After that, only the tape running direction is reversed and the following block (here 13) in the tape running direction<b>Behind</b> the next readable (here 12). This can be continued accordingly until the entire band surface is filled with blocks. Since each block is written after a block that has already been written, a longer transition zone is required due to the tolerances in the absolute tape speed and the distance between the reading device (for positioning) and the writing device.

Real time playback

The playback takes place accordingly at continuous tape speed and by detecting the logically following block. Here, too, the increasing distance between the logically following blocks at the reversal points has a disadvantageous effect. Ultimately, for example, the belt speed at the reversal points must be increased so that the next logical block can be read in time.

It can also be useful to have the first two blocks of a sound recording (not just 1 + 2, but also 44 + 45) follow each other more closely so that the data buffer is filled securely at the beginning of a playback and does not run empty. However, this can also be compensated for by a correspondingly higher belt speed, which is controlled by the buffer fill level.

Version 5

This is a further development of version 4, which avoids the time problems at the reversal points.

It is characteristic that when recording for the first time, not only the useful blocks, but also the blocks in between are written with empty data, but with the correct address. At the end, not all, but at least two more empty blocks are written, which can then be used for positioning in the following run. This means that positioning for writing at every logical position is possible for each subsequent run, without increased lengths of the transition areas being necessary to compensate for possible tolerances.

The sequence would be as follows (see Fig. 10): write blocks 1, 2, 3, 3 in the first pass in the first macro track, write blocks 24 .. 49 or 23 .. 50 in between with empty data and addresses, at the end write the two blocks 22 and 27 with empty data and addresses.

After reversing the direction and positioning the track on macro track 2, the two blocks 28 and 21 are empty and then block 4 is written with user data. If the tape run starts in good time, these blocks are approximately below 27, 22 and 3. This continues until block 12 at the end of macro track 4. After that, there is only a reversal of direction.

In all subsequent write operations, blocks that have already been written can now be used for positioning.

Block 12 is used for positioning for block 13, block 11 for block 14, block 10 for block 15. Then another empty block (39) is now written behind the already existing empty block 34, so that there are two again at the end. The already written empty block 33 is used for positioning for block 16.

In this way, there is always at least one block (empty or with user data, but with the correct address) available in each track and for each direction, which can be used for positioning for the next block to be written. With this proposal, the time available for reversing direction and switching lanes is additionally required for processing a maximum of 4 blocks. This means that time problems can no longer occur at the end of the track.

Version 6

This version is a further development of version 5, in which further energy savings are made by reducing the track positioning processes. It is characteristic that a macro track is filled completely before another track is used. This is shown in Fig. 11.

It is characteristic again that at the end of a track at least two empty blocks are written which, after reversing the direction, are used for positioning for the next writing process.

It is easy to see that an odd number of blocks (as shown) will result in the read / write device being at the end of the track when it is completely full. The next track can then be built up in the opposite direction (as shown). An even number of blocks leads to the read / write device being at the beginning again after a complete processing of a track. Then the next track can be built up in the same direction, whereupon it always returns to the beginning of the piece of tape.

If the number of blocks is odd, it is also important whether the number of macro tracks is even or odd.

If the number of tracks is odd, the read / write device is finally back at the end of the tape or piece of tape when the tape is completely full, which can be advantageous for further writing or reading operations on the following tape pieces, but means longer positioning times back to the beginning.

If the number of tracks is even, the read / write device is finally back at the beginning of the tape when the tape is completely full, which means a short positioning time at the beginning but a longer pause for further processing of a subsequent piece of tape.

The question of where the read / write device is after the complete processing of a track is of little consequence, however, if there is sufficient time for positioning and a rewinding process. For a continuous playback process of the music recorded on a tape section, it can be advantageous if there is also a continuous process mechanically. In the case of a real-time self-recording, the absolutely last pass could be omitted, whereby for a small loss of storage space one is given the opportunity to continue directly on the next piece of magnetic tape.

The version 6 described has the additional advantage that a 'moderate' fast recording, for example only using a 64-track head, can be done macro-wise, with only the data required for this track being available in the recording device in a memory for scrambling have to stand.

From the above statements it follows that version 6 described should be used expediently for SDCR audio album. The number of blocks and number of tracks are even here. In total, several formats can be defined:

Format for quick recording with data reduction

Here, the blocks are provided in the individual tracks according to FIG. 11. The distances between the blocks can be kept as small as is necessary for a direct complete recording and a sequential read operation.

According to the above parameters, 6x40 blocks with a length of 6 mm could be accommodated on a 6x24 cm long piece of tape. This would result in a playing time of 16 minutes per track and a total of N x 16 minutes.

Self-recording format with data reduction

With self-recording you would have to proceed as described for version 6. Since safety distances and also run-in sequences are required for the PLL for positioning, erasing and recording, fewer blocks can be accommodated per track. If the belt run was otherwise the same, for example, only 6x30 blocks would have to be accommodated over the same length of 6x24 cm. The self-recording would then lead to a reduction in the recording time per piece of tape, but in principle with the same tape run.

Self-recording format without data reduction

Without data reduction, ie in 16-bit CD format, the amount of data that has to be applied is about 6 times higher. This can in turn be achieved with basically the same tape run by denser sequence of logically successive blocks on the tape. If a new block has to be written or read every 4 seconds, it is every 4/6 seconds. That means that the logically related blocks are no longer 24 cm apart, but only 4 cm. With safety distances to the neighboring blocks, it is then no longer necessary to accommodate 30, but 5, for example.

If the tape run is otherwise the same, the recording time is shortened according to the data reduction factor.

Other formats

The proposed form of continuous tape movement, together with the intermittent read and write operation, can be easily adapted to other data formats and rates.