Note: Descriptions are shown in the official language in which they were submitted.
CA 02290641 1999-11-18
WO 99/48096 ] PC'T/IB99/00439
Recordinglreproduction and/or editing of real time information on/from a disc
like record
carrier.
The invention relates to an apparatus for recording a real time information
signal, such as a digital video signal, on a disc like record carrier, to an
apparatus for editing
an information signal recorded earlier on said disc like record carrier, to
corresponding
methods for recording/editing information, to a reading apparatus for reading
the information
signal and to a record carrier. The record carrier may be of the magnetic or
the optical type.
An apparatus for recording a real time information signal, such as an MPEG
encoded video
information signal, on a record Garner is known from USP 5,579,183 (PHN
14818). The
record carrier in the said document is in longitudinal form.
Disc like record carriers have the advantage of a short access time. This
enables
the possibility of carrying out 'simultaneous' recording and reproduction of
information
signals on/from the record carrier. During recording and reproduction,
information should be
recorded on/reproduced from the record carrier such that an real time
information signal can
be recorded on the record carrier and 'at the same time' a real time
information signal
recorded earlier on the record carrier can be reproduced without any
interruption.
The invention aims at providing measures to enable the various requirements,
such as the ones described above. In accordance with the invention, the
apparatus for
recording a real time information signal, such as a digital video signal, on a
disc like record
carrier, a data recording portion of which is subdivided into fixed sized
fragment areas,
comprises
- input means for receiving the information signal,
- signal processing means for processing the information signal into a channel
signal for
recording the channel signal on the disc like record carrier,
- writing means for writing the channel signal on the record carrier,
the signal processing means being adapted to convert the information signal
into blocks of
information of the channel signal, the writing being adapted to write a block
of information of
CA 02290641 1999-11-18
WO 99/48096 2 PCT/IB99/00439
the channel signal in a fragment area on the record carrier, and wherein the
signal processing
is further adapted to convert the information signal into the blocks of
information of the
channel signal, such that the size of the blocks of information can be
variable and satisfies the
following relationship:
SFA/2 _< size of a block of the channel signal S SFA,
where SFA equals the fixed size of the fragment area.
Further, the apparatus for editing a real time information signal, such as a
digital video signal, recorded in an earlier recording step on a disc like
record carrier, a data
recording portion of which is subdivided into fixed sized fragment areas, the
information
signal being converted into a channel signal prior to recording and
subsequently recorded on
the record carrier, such that blocks of information of the channel signal are
recorded in
corresponding fragment areas on the record carrier, comprises:
- input means for receiving an exit position in a first information signal
recorded on the record
carrier and for receiving an entry position in a second information signal,
which may be the
first information signal, recorded on the record carrier,
- means for storing information relating to the said exit and entry position,
- bridging block generating means for generating at least one bridging block
of information,
which bridging block of information comprises information from at least one of
the first and
second information signals, which information is located before the exit
position in the first
information signal and/or after the entry position in the second information
signal, and where
the size of a bridging block of information can be variable and satisfies the
requirement:
SFA/2 <_ size of a bridging block of information <_ SFA,
where SFA equals the fixed size of the fragment areas,
- writing means for writing the at least one bridging block of information
into a corresponding
fragment area, and
- means for reproducing the edited stream of information from said record
carrier.
Further, the apparatus for reading a real time information signal, such as a
digital video signal, from a disc like record carrier, the information signal
being recorded in
channel encoded form in a data recording portion of the record carrier, the
data recording
CA 02290641 1999-11-18
WO 99/48096 3 PCTIIB99/00439
portion being subdivided into fixed size fragment area, blocks of information
of the channel
encoded information signal being recorded in corresponding fragment areas, the
size of the
blocks of information can be variable and satisfy the following relationship:
SFA/2 <_ size of a block of information of the channel signal _< SFA,
where SFA equals the fixed size of the fragment areas,
the apparatus comprises:
- means for reading the channel signal from the record carrier,
- signal processing means for processing the blocks of information of variable
size and read
from the fragment areas into portions of the information signal,
- means for outputting the information signal.
A further advantageous embodiment is characterized in that the blocks of
information of a consecutive sequence satisfy alternately the following
relationships:
SFA/2 <_ size of a block of the channel signal <_ SFA and
size of a block of the channel signed = SFA.
This leads to either a more efficient occupation of space or eases the
requirements of an
apparatus. Another advantageous embodiment with the same advantages as above
is
characterized in that the blocks of information of a consecutive sequence
satisfy the following
relationship:
2 SFA/3 <_ size of a block of the channel signal 5 SFA.
These and other aspects of the invention will be apparent from and elucidated
with reference to the embodiments hereafter in the figure description, in
which
Figure 1 shows an embodiment of the apparatus,
Figure 2 shows the recording of blocks of information in fragment areas on the
record carrier,
Figure 3 shows the principle of playback of a video information signal,
Figure 4 shows the principle of editing of video information signals,
Figure 5 shows the principle of 'simultaneous' play back and recording,
CA 02290641 1999-11-18
WO 99/48096 4 PCT/IB99/00439
Figure 6 shows a situation during editing when the generation and recording of
a bridging block of information is not required,
Figure 7 shows an example of the editing of a video information signal and the
generation of a bridging block of information, at the location of an exit
point from the
information signal,
Figure 8 shows another example of the editing of a video information signal
and the generation of a bridging block of information, at the same location of
the exit point as
in figure 7,
Figure 9 shows an example of the editing of a video information signal and the
generation of a bridging block of information, at the location of an entry
point to the
information signal,
Figure 10 shows an example of the editing of two information signals and the
generation of a bridging block of information,
Figure 11 shows an example of the editing of two information signals and the
generation of a bridging block of information, where the editing includes re-
encoding some of
the information of the two information signals,
Figure 12 shows a further elaboration of the apparatus,
Figure 13 shows sequences of fragments illustrating three embodiments of the
invention respectively satisfying the HF condition, the HFFF condition and the
2/3 condition.
Figure 14 shows the general case of bridge creation without reallocation,
Figure 15 shows the worst case situation of creating a bridge assuming a HFFF
condition, with figures 16-21 illustrating the several allocation strategies
in this case
Figure 22 shows the result of bridge creation without reallocation in a
locally
FF stream, with figures 23-24 illustrating the several allocation strategies
in this case,
Figure 24A shows a bridge assuming a 2/3 condition containing only MPEG
data, with figures 24B-36 illustrating the several allocation strategies in
this case.
Figure 1 shows an embodiment of the apparatus in accordance with the
invention. The the following figure description, the attention.will be
focussed on the
recording, reproduction and editing of a video information signal. It should
however be noted
that other types of signal could equally well be processed, such as audio
signals, or data
signals.
CA 02290641 1999-11-18
WO 99/48096 PCT/IB99/00439
The apparatus comprises an input terminal 1 for receiving a video information
signal to be recorded on the disc like record Garner 3. Further, the apparatus
comprises an
output terminal 2 for supplying a video information signal reproduced from the
record carrier
3. The record carrier 3 is a disc like record carrier of the magnetic or
optical form.
The data area of the disc like record carrier 3 consists of a contiguous range
of
physical sectors, having corresponding sector addresses. This address space is
divided into
fragment areas. A fragment area is a contiguous sequence of sectors, with a
fixed length.
Preferably, this length corresponds to an integer number of ECC-blocks
included in the video
information signal to be recorded.
The apparatus shown in figure 1 is shown decomposed into two major system
parts, namely the disc subsystem 6 and the what is called 'video recorder
subsystem'8. The
following features characterise the two subsystems:
- The disc subsystem can be addressed transparently in terms of logical
addresses. It handles
defect management (involving the mapping of logical addresses onto physical
addresses)
autonomously.
- For real-time data, the disc subsystem is addressed on a fragment-related
basis. For data
addressed in this manner the disc subsystem can guarantee a maximum
sustainable bit rate for
reading and/or writing. In the case of simultaneous reading and writing, the
disc subsystem
handles the read/write scheduling and the associated buffering of stream data
from the
independent read and write channels.
- For non-real-time data, the disc subsystem may be addressed on a sector
basis. For data
addressed in this manner the disc subsystem cannot guarantee any sustainable
bit rate for
reading or writing.
- The video recorder subsystem takes care of the video application, as well as
file system
management. Hence, the disc subsystem does not interpret any of the data that
is recorded in
the data area of the disc.
In order to realize real time reproduction in all situations, the fragment
areas
introduced earlier need to have a specific size. Also in a situation where
simultaneous
recording and reproduction takes place, reproduction should be uninterrupted.
In the present
example, the fragment size is chosen to satisfy the following requirement:
fragment size = 4 MB = 22~ bytes
CA 02290641 1999-11-18
WO 99/48096 6 PCT/IB99/00439
Recording of a video information signal will briefly be discussed hereafter,
with
reference to figure 2. In the video recorder subsystem, the video information
signal, which is a
real time signal, is converted into a real time file, as shown in figure 2a. A
real-time file
consists of a sequence of signal blocks of information recorded in
corresponding fragment
areas. There is no constraint on the location of the fragment areas on the
disc and, hence, any
two consecutive fragment areas comprising portions of information of the
information signal
recorded may be anywhere in the logical address space, as shown in figure 2b.
Within each
fragment area, real-time data is allocated contiguously. Each real-time file
represents a single
AV stream. The data of the AV stream is obtained by concatenating the fragment
data in the
order of the file sequence.
Next, playback of a video information signal recorded on the record carrier
will
be briefly discussed hereafter, with reference to figure 3. Playback of a
video information
signal recorded on the record carrier is controlled by means of a what is
called 'playback-
control-program' {PBC program). In general, each PBC program defines a (new)
playback
sequence. This is a sequence of fragment areas with, for each fragment area, a
specification of
a data segment that has to be read from that fragment. Reference is made in
this respect to
figure 3, where playback is shown of only a portion of the first three
fragment areas in the
sequence of fragment areas in figure 3. A segment may be a complete fragment
area, but in
general it will be just a part of the fragment area. (The latter usually
occurs around the
transition from some part of an original recording to the next part of the
same or another
recording, as a result of editing.)
Note, that simple linear playback of an original recording can be considered
as
a special case of a PBC program: in this case the playback sequence is defined
as the sequence
of fragment areas in the real-time file, where each segment is a complete
fragment area except,
probably, for the segment in the last fragment area of the file. For the
fragment areas in a
playback sequence, there is no constraint on the location of the fragment
areas and, hence, any
two consecutive fragment areas may be anywhere in the logical address space.
Next, editing of one or more video information signals recorded on the record
Garner will be briefly discussed hereafter, with reference to figure 4. Figure
4 shows two video
information signals recorded earlier on the record Garner 3, indicated by two
sequences of
fragments named 'file A' and 'file B'. For realizing an edited version of one
or more video
information signals recorded earlier, a new PBC program should be realized for
defining the
edited AV sequence. This new PBC program thus defines a new AV sequence
obtained by
concatenating parts from earlier AV recordings .in a new order. The parts may
be from the
*rB
CA 02290641 1999-11-18
WO 99/48096 ~ PCT/IB99/00439
same recording or from different recordings. Tn order to play back a PBC
program, data from
various parts of (one or more) real-time files has to be delivered to a
decoder.This implies a
new data stream that is obtained by concatenating parts of the streams
represented by each
real-time file. In the figure 4, this is illustrated for a PBC program that
uses three parts, one
from the file A and two from the file B.
Figure 4 shows that the edited version starts at a point P 1 in the fragment
area
f(i) in the sequence of fragment areas of figure A and continues until point
P2 in the new
fragment area f(i+1) of file A. Then reproduction jumps over to the point P3
in the fragment
area f(j) in file B and continues until point P4 in fragment area f(j+2) in
file B. Next
reproduction jumps over to the point PS in the same file B, which may be a
point earlier in the
sequence of fragment areas of file B than the point P3, or a point later in
the sequence than the
point P4.
Next, a condition for seamless playback during simultaneous recording will be
discussed. In general, seamless playback of PBC programs can only be realized
under certain
conditions. The most severe condition is required to guarantee seamless
playback while
simultaneous recording is performed. One simple condition for this purpose
will be
introduced. It is a constraint on the length of the data segments that occur
in the playback
sequences, as follows: In order to guarantee seamless simultaneous play of a
PBC program,
the playback sequence defined by the PBC program shall be such that the
segment length in all
fragments (except the first and the last fragment area) shall satisfy:
2 MB <_ segment length <_ 4 MB
The use of fragment areas allows one to consider worst-case performance
requirements in terms of fragment areas and segments (the signal blocks stored
in the fragment
ares) only, as will be described hereafter. This is based on the fact that
single logical fragments
areas, and hence data segments within fragment areas, are guaranteed to be
physically
contiguous on the disc, even after remapping because of defects. Between
fragment areas,
however, there is no such guarantee: logically consecutive fragment areas may
be arbitrarily
far away on the disc. As a result of this, the analysis of performance
requirements concentrates
on the following:
a. For playback, a data stream is considered that is read from a sequence
of.segments on the
disc. Each segment is contiguous and has an arbitrary length between 2 MB and
4 MB, but the
segments have arbitrary locations on the disc.
CA 02290641 1999-11-18
WO 99/48096 g PCT/IB99/00439
b. For recording, a data stream is considered that is to be written into a
sequence of 4 MB
fragment areas on the disc. The fragment areas have arbitrary locations on the
disc.
Note that for playback, the segment length is flexible. This corresponds to
the segment
condition for seamless play during simultaneous record. For record, however,
complete
segments areas with fixed length are written.
Given a data stream for record and playback, we will concentrate on the disc
subsystem during
simultaneous record and playback. It is assumed that the video recorder
subsystem delivers a
sequence of segment addresses for both the record and the playback stream well
in advance.
For simultaneous recording and playback, the disc subsystem has to be able to
interleave read
and write actions such that the record and playback channels can guarantee
sustained
performance at the peak rate without buffer overflow or underflow. In general,
different R/W
scheduling algorithms may be used to achieve this. There are, however, strong
reasons to do
scheduling in such a way that the R/W cycle time at peak rates is as short as
possible:
- Shorter cycle times imply smaller buffer sizes for the read and write
buffer, and hence for the
total memory in the disc subsystem.
- Shorter cycle times imply shorter response times to user actions. As an
example of response
time consider a situation where the user is doing simultaneous recording and
playback and
suddenly wants to start playback from a new position. In order to keep the
overall apparatus
response time (visible to the user on his screen) as short as possible, it is
important that the
disc subsystem is able to start delivering stream data from the new position
as soon as
possible. Of course, this must be done in such a way that, once delivery has
started, seamless
playback at peak rate is guaranteed. Also, writing must continue
uninterruptedly with
guaranteed performance.
For the analysis here, a scheduling approach is assumed, based on a cycle in
which one complete fragment area is written. For the analysis of drive
parameters below, it is
sufficient to consider the minimum cycle time under worst-case conditions.
Such a worst-case
cycle consists of a writing interval in which a 4 MB segment is written, and a
reading interval
in which at least 4 MB is read, divided over one or more segments. The cycle
includes at least
two jumps (to and from the writing location), and possibly more, because the
segment lengths
for reading are flexible and may be smaller than 4 MB. This may result in
additional jumps
from one read segment location to another. However, since read segments are no
smaller than
2 MB, no more than two additional jumps are needed to collect a total of 4 MB.
So, a worst-
case R/W .cycle has a total of four jumps, as illustrated in figure 5. In this
figure, x denotes the
last part of a read segment, y denoted a complete read segment, with length
between 2 MB and
CA 02290641 1999-11-18
WO 99/48096 9 PCT/IB99/00439
4 MB, and z denotes the first part of a read segment and the total size of x,
y and z is again 4
MB in the present example.
In general, the required drive parameters to achieve a guaranteed performance
for simultaneous recording and playback depend on major design decisions such
as the
rotational mode etc. These decisions in turn depend on the media
characteristics.
The above formulated conditions for seamless play during simultaneous record
are derived such that they can be met by different designs with realistic
parameters. In order to
show this, we discuss the example of a CLV (constant linear velocity) drive
design here.
In the case of a CLV design, transfer rates for reading and writing are the
same
and independent of the physical location on the disc. Therefore, the worst-
case cycle described
above can be analyzed in terms of just two drive parameters: the transfer rate
R and the worst-
case all-in access time T. The worst-case access time i is the maximum time
between the end
of data transfer on one location and the begin of data transfer on another
location, for any pair
of locations in the data area of the disc. This time covers speed-up/down of
the disc, rotational
latency, possible retries etc., but not processing delays etc.
For the worst-case cycle described in the previous section, all jumps may be
worst-case jumps of duration T. This gives the following expression for the
worst-case cycle
time:
Tmax = 2F/Rt + 4.T
where F is the fragment size: F = 4 MB = 33.6 .106 bits.
In order to guarantee sustainable performance at peak user rate R, the
following should hold:
F >_ R.Tmax
This yields:
R <_ F/T",~ = Rt.F/2.(F + 2Rt.i)
As an example, with Rt = 35 Mbps and T = 500 ms, we would have: R <_ 8.57
Mbps.
Next, editing will be further described. Creating a new PBC program or editing
an existing PBC program, generally results in a new playback sequence. It is
the objective to
CA 02290641 1999-11-18
WO 99/48096 to PCT/IB99/00439
guarantee that the result is seamlessly playable under all circumstances, even
during
simultaneous recording. A series of examples will be discussed, where it is
assumed that the
intention of the user is to make a new AV stream out of one or two existing AV
streams. The
examples will be discussed in terms of two streams A and B, where the
intention of the user is
to make a transition from A to B. This is illustrated in figure 6, where a is
the intended exit
point from stream A and where b is the intended entry point into stream B.
Figure 6a shows the sequence of fragment areas ......, f(i-1), f(i), f(i+1),
f(i+2),
.... of the stream A and figure 6b shows the sequence of fragment areas
......, f(j-1), f(j), f(j+1),
f(j+2), .... of the stream B. The edited video information signal consists of
the portion of the
stream A preceding the exit point a in fragment area f(i+1 ), and the portion
of the stream B
starting from the entry point b in fragment area f~j).
This is a general case that covers all cut-and-paste-like editing, including
appending two streams etc. It also covers the special case where A and B are
equal. Depending
on the relative position of a and b, this special case corresponds to FBC
effects like skipping
part of a stream or repeating part of a stream.
The discussion of the examples focuses on achieving seamless playability
during simultaneous recording. The condition for seamless playability is the
segment length
condition on the length of the signal blocks of information stored in the
fragment areas, that
was discussed earlier. It will be shown below that, if streams A and B satisfy
the segment
length condition, then a new stream can be defined such that it also satisfies
the segment
length condition. Thus, seamlessly playable streams can be edited into new
seamlessly
playable streams. Since original recordings are seamlessly playable by
construction, this
implies that any edited stream will be seamlessly playable. As a result,
arbitrarily editing
earlier edited streams is also possible. Therefore streams A and B in the
discussion need not be
original recordings: they can be arbitrary results of earlier virtual editing
steps.
In a first example, a simplified assumption will be made about the AV encoding
format and the choice of the exit and entry points. It is assumed that the
points a and b are such
that, from the AV encoding format point of view, it would be possible to make
a
straightforward transition. In other words, it is assumed that straightforward
concatenation of
data from stream A (ending at the exit point a) and data from stream B
(starting from entry
point b) results in a valid stream, as far as the AV encoding format is
concerned.
The above assumption implies that in principle a new playback sequence can be
defined based
on the existing segments. However, for seamless playability at the transition
from A to B, we
have to make sure that all segments satisfy the segment length condition. Let
us concentrate on
CA 02290641 1999-11-18
WO 99/48096 PCT/IB99/00439
11
stream A and see how to ensure this. Consider the fragment area of stream A
that contains the
exit point a. Let s be the segment in this fragment area that ends at point a,
see figure 6a.
If 1(s), the length of s, is at least 2 MB, then we can use this segment in
the new
playback sequence and point a is the exit point that should be stored in the
PBC program.
However, if i(s) is less than 2 MB, then the resulting segment s does not
satisfy
the segment length condition. This is shown in figure 7. In this case a new
fragment area, the
so-called bridging fragment area f is created. In this fragment area, a
bridging segment
comprising a copy of s preceded by a copy of some preceding data in stream A,
is stored. For
this, consider the original segment r that preceded s in stream A, shown in
figure 7a. Now,
depending on the length of r, the segment stored in fragment area f(i), either
all or part of r is
copied into the new fragment area f
If I(r) + 1(s) _< 4 MB, then all of r is copied into f, and the original
segment r is
not used in the new playback sequence, as illustrated in figure 7a. More
specifically, the new
exit point is the point denoted a', and this new exit point a' is stored in
the PBC program, and
later on, after having terminated the editing step, recorded on the disc like
record carrier. Thus,
in response to this PBC program, during playback of the edited video
information stream, after
having read the information stored in the fragment area f(i-1), the program
jumps to the
bridging fragment area f , for reproducing the information stored in the
bridging fragment area
f , and next jumps to the entry point in the video stream B to reproduce the
portion of the B
stream, as schematically shown in figure 7b.
If 1(r) + 1(s) > 4 MB, then some part p from the end of r is copied into f ,
where
the length of p is such that we have
2 MB <_ 1(r) -1(p) <_ 4 MB ~ 2 MB <_ I(p) + 1(s) <_ 4 MB
Reference is made to figure 8, where figure 8a shows the original A stream and
figure 8b shows the edited stream A with the bridging fragment area f . In the
new playback
sequence, only a smaller segment r' in the fragment area f(i) containing r is
now used. This
new segment r' is a subsegment of r, viz. the first part of r with length
1(r') =1(r) -1(p).
Further, a new exit point a' is required, indicating the position where the
original stream A
should be left, for a jump to the bridging fragment f . This new exit position
should therefore
be stored in the PBC program, and stored later on on the disc.
In the example given above, it was discussed how to create a bridging segment
(or: bridging block of information) for the fragment area f , in case the last
segment in stream
CA 02290641 1999-11-18
WO 99/48096 I2 PCT/1B99/00439
A (i.e. s) becomes too short. We will now concentrate on stream B. In stream
B, there is a
similar situation for the segment that contains the entry point b, see figure
9. Figure 9a shows
the original stream B and figure 9b shows the edited stream. Let t be the
segment comprising
the entry point b. If t becomes too short, a bridging segment g can be created
for storage in a
corresponding bridging fragment area. Analogous to the situation for the
bridging fragment
area f , g will consist of a copy of t plus a copy of some more data from
stream B. This data is
taken from the original segment a that succeeds t in the fragment area f(j+1)
in the stream B.
Depending on the length of u, either all or a part of a is copied into g. This
is analogous to the
situation for r described in the earlier example. We will not describe the
different cases in
detail here, but figure 9b gives the idea by illustrating the analogy of
figure 8, where a is split
into v and u'. This results in a new entry point b' in the B stream, to be
stored in the PBC
program and, later on, on the record carrier.
The next example, described with reference to figure 10, shows how a new
seamlessly playable sequence can be defined under all circumstances, by
creating at most two
1 S bridging fragments (f and g). It can be shown that, in fact, one bridging
fragment area is
sufficient, even if both s and t are too short. This is achieved if both s and
t are copied into a
single bridging fragment area (and, if necessary, some preceding data from
stream A and/or
some succeeding data from stream B). This will not be described extensively
here, but figure
I O shows the general result.
In examples described above, it was assumed that concatenation of stream data
at the exit and entry points a and b was sufficient to create a valid AV
stream. In general,
however, some re-encoding has to be done in order to create a valid AV stream.
This is usually
the case if the exit and entry points are not at GOP boundaries, when the
encoded video
information signal is an MPEG encoded video information signal. The re-
encoding will not be
discussed here, but the general result will be that some bridge sequence is
needed to go from
stream A to stream B. As a consequence, there will be a new exit point a' and
a new entry
point b', and the bridge sequence will contain re-encoded data that
corresponds with the
original pictures from a' to a followed by the original pictures from b to b'.
Not all the cases will be described in detail here, but the overall result is
like in the previous
examples: there will be one or two bridging fragments to cover the transition
from A to B. As
opposed to the previous examples, the data in the bridging fragments is now a
combination of
re-encoded data and some data copied from the original segments. Figure I 1
gives the general
flavour of this.
CA 02290641 1999-11-18
WO 99/48096 PCT/IB99/00439
13
As a final remark, note that one does not have to put any special constraints
on
the re-encoded data. The re-encoded stream data simply has to satisfy the same
bitrate
requirements
as the original stream data.
Figure 12 shows a schematic version of the apparatus in more detail. The
apparatus comprises a signal processing unit 100 which is incorporated in the
subsystem 8 of
figure 1. The signal processing unit 100 receives the video information signal
via the input
terminal 1 and processes the video information into a channel signal for
recording the channel
signal on the disc like record earner 3. Further, a read/write unit 102 is
available which is
incorporated in the disc subsystem 6. The read/write unit 102 comprises a
read/write head
104, which is in the present example an optical read/write head for
reading/writing the channel
signal on/from the record carrier 3. Further, positioning means 106 are
present for positioning
the head 104 in a radial direction across the record~carrier 3. A read/write
amplifier 108 is
present in order to amplify the signal to be recorded and amplifying the
signal read from the
record carrier 3. A motor 110 is available for rotating the record carrier 3
in response to a
motor control signal supplied by a motor control signal generator unit 112. A
microprocessor
114 is present for controlling all the circuits via control lines 116, 118 and
120.
The signal processing unit 110 is adapted to convert the video information
received via the input terminal 1 into blocks of information of the channel
signal having a
specific size. The size of the blocks of information (which is the segment
mentioned earlier)
can be variable, but the size is such that it satisfies the following
relationship:
SFA/2 <_ size of a block of the channel signal <_ SFA,
where SFA equals the fixed size of the fragment areas. In the example given
above, SFA = 4
MB. The write unit 102 is adapted to write a block of information of the
channel signal in a
fragment area on the record carrier.
in order to enable editing of video information recorded in an earlier
recording
step on the record carrier 3, the apparatus is further provided with an input
unit 130 for
receiving an exit position in a first video information signal recorded on the
record earner and
for receiving an entry position in a second video information signal recorded
on that same
record earner. The second information signal may be the same as the first
information signal.
Further, the apparatus comprises a memory 132, for storing information
relating to the said
exit and entry positions. Further the apparatus comprises a bridging block
generating unit 134,
CA 02290641 1999-11-18
WO 99/48096 PCT/IB99/00439
14
incorporated in the signal processing unit 100, for generating at least one
bridging block of
information (or bridging segment) of a specific size. As explained above, the
bridging block of
information comprises information from at least one of the first and second
video information
signals, which information is located before the exit position in the first
video information
signal and/or after the entry position in the second video information signal.
During editing, as
described above, one or more bridging segments are generated in the unit 134
and in the edit
step, the one or more bridging segments) is (are) recorded on the record
carrier 3 in a
corresponding fragment. The size of the at least one bridging block of
information also
satisfies the relationship:
SFA/2 <_ size of a bridging block of information <_ SFA.
Further, the PBC programs obtained in the edit step can be stored in a memory
incorporated in the microprocessor 114, or in another memory incorporated in
the apparatus.
The PBC program created in the edit step for the edited video information
signal will be
recorded on the record carrier, after the editing step has been terminated. In
this way, the
edited video information signal can be reproduced by a different reproduction
apparatus by
retrieving the PBC program from the record carrier and reproducing the edited
video
information signal using the PBC program corresponding to the edited video
information
signal.
In this way, an edited version can be obtained, without re-recording portions
of
the first andlor second video information signal, but simply by generating and
recording one or
more bridging segments into corresponding (bridging) fragment areas on the
record carrier.
In the above described embodiment, fragments are created on the disk that are
at least half full. This will be referred to as the HF condition while a
fragment is called FF if it
is completely full. As shown in the above, after editing a stream that
satisfies the HF
condition, it is possible to ensure that the resulting stream also satisfies
the HF condition. This
requires a single fragment to be allocated for the bridge. In the worst case
this may result in an
A/V stream consisting of all half filled fragments. Figure 13A schematically
illustrates such a
sequence of half filled fragments HF. This stream places severe requirements
on drive
performance.
Next a second and third embodiment will be described for recording and editing
of
video/audio streams on a disk. These embodiments ease the worst case situation
with respect
to occupation of space that may occur with the first embodiment and which is
illustrated in
Figure 13A. The worst case stream of fragments resulting from a second
embodiment is shown
CA 02290641 1999-11-18
WO 99/48096 15 PCT/IB99/00439
in Figure 13B. This stream satisfies a HFFF condition implying that at least
every second
fragment is fully filled. The worst case stream of fragments resulting from a
third embodiment
is shown in Figure 13C. It is remarked that the second and third embodiments
also ease the
requirements of an apparatus. This stream satisfies a 2/3F condition implying
that the
minimum fullness of a fragment is greater than 2/3. Although this case will be
considered in
detail, other values for the fullness are also possible.
It will be shown that to achieve either of these conditions, more than one
fragment may be
required for bridges. First bridge creation is considered in case of the first
embodiment
satisfying the HF condition. Fig. 14 shows the general case of creating a
bridge, the details
being discussed in detail here before. Note that the fragments before and
after the bridge may
originally have been full fragments but due to the choice of edit points the
result is that they
will be partially filled in the edit sequence. The only assumption is that the
fragments before
the bridge, the bridge fragment and the fragment after the bridge are at least
half full.
Next will be shown how to create a bridge in case of the second embodiment
satisfying the
1 S HFFF condition. Figure 15 shows an edited sequence illustrating the worst
case situation in
this case. In the original sequences both the last fragment before the bridge
and the first
fragment after the bridge must be full since it is assumed that the original
streams satisfy the
FFHF condition. First is tried to preserve the FFHF condition by reallocating
the fragment
before the bridge, the bridge fragment and the fragment after the bridge
(three fragment
reallocation). In general the following assumption can be made about the size
of these
fragments:
1.5 S size (3 *HF) <_ 3 ( 1 ]
where the units are the fragment size. This condition implies the following
possibilities for
reallocating the three fragments
1.5 < size (3*HF) < 2 [2]
2 <_ size (3*HF) S 2.5 [3]
2.5 S size (3*HF) <_ 3 [4]
Possibility [1] can be reallocated as FF + HF, [2] as FF + HF +FF and [3] as
FF + FF +~ HF.
CA 02290641 1999-11-18
WO 99/48096 16 PCT/IB99/00439
In the last case, by reallocating the three fragments as FF HF and FF, it is
possible to maintain
the FFHF condition. However, the result is that the bridge requires three
fragments instead of
one, as is illustrated in Figure 16. In the other cases it is not possible to
preserve the FFHF
condition by reallocating the three fragments. Therefore, resulting form [ 1 J
and [2] the
following condition holds:
1.5 <_ size (3*HF) S 2.5 [5]
Adding a fourth HF fragment (four fragment reallocation), this condition
becomes:
2 <_ size (4*HF) <_ 3.5 [6]
This condition implies the following possibilities for reallocating the four
fragments:
2 = size (4*HF) [7]
1 S 2 < size (4*HF) < 2.5 [8]
2.5 <_ size (4*HF) < 3 [9]
3 < size (4*HF) < 3.5 [10]
Possibility [7] may be reallocated as FF + HF, [8] as FF + HF + HF, [9] as FF
+ FF +HF and
[10] as FF + FF + HF +HF. In the second case [8] it is not possible to satisfy
the FFHF
condition. The third case [9] is shown in Figure 17. The fourth case [8]
requires four
fragments for the bridge and is shown in Figure 18. However, possibility [8]
implies that the
first HF fragment can be made 0.5 and it can still be guaranteed that the FF
segments can be
filled. But this means that just the first half of the original HF fragment
can be taken. This
results in a three fragment bridge which is shown in Figure 18.
Next possibility [8] , which was not possible to reallocate, is added a fifth
HF
fragment:
2.5 < size (5*HF) < 3.5 [11]
This possibility implies the following possibilities for reallocating the five
fragments
2.5 < size (5*HF) <_ 3 [12]
CA 02290641 1999-11-18
WO 99/48096 1~ PCT/IB99/00439
3 < size (4*HF) S 3.5 [13]
In either case it is possible to reallocate the five fragments to satisfy the
HFFF condition.
Figure 19 shows the first case [ 12] requiring three fragments while Figure 20
shows the
second case [I3] requiring four fragments. However, as in the above, the
possibility [13]
implies that the first HF fragment can be made 0.5 and it can still be
guaranteed that the FF
fragments can be filled. But this means that just the first half segment from
the original HF
fragment can be taken. This results in a three fragment bridge which is shown
in Figure 21.
Concluding, it is possible to satisfy the HFFF condition with a worst case
bridge of three
fragments. Because this condition is weaker than the FF condition, the
previous analysis also
covers editing of FF streams as well. In most cases the streams being edited
will be locally FF
unless the user performs a number of very closed edits. The situation when
editing from a
locally FF stream should be better than the worst case described above. The
general case of
I S editing a stream that satisfies the FF condition is shown in Figure 22.
Considering the three HF
fragments, the following condition is valid:
1.5 <_ size (3*HF) < 3 [14]
This condition implies the following possibilities for reallocating the three
fragments
1.5 <_ size (3*HF) _< 2 [15]
2 <_ size (3*HF) <_ 3 [16]
Possibility [ 15] can be reallocated as FF + HF and [ 16] as FF + HF + HF. In
the first case [ 1 S]
it is possible to reallocate the three fragments as two fragments as shown in
Figure 23. It is
noted that in general it is not possible to use a single FF fragment for the
bridge and keep one
of the original HF fragments. In the second case [ 16] it is possible to use a
single fragment for
the bridge. Data can be copied from the other two fragments until the bridge
fragment is FF.
This allows the HFFF condition to be satisfied with a single bridge fragment.
This is shown in
Figure 24.
Finally it is remarked that a worst case will only occur where the user does a
number of close edits. In the normal case, where edits are a few seconds
apart, the bridge may
CA 02290641 1999-11-18
WO 99/48096 1g PCT/IB99/00439
require two fragments at most. The above discussion of the second embodiment
for a worst
case situation, eventually together with the detailed discussion of the first
embodiment, enable
a man skilled in the art to implement the bridge creation either in software
or hardware or a
mixture of both software and hardware for all cases.
S It is remarked that replacing the HF condition according to the first
embodiment with the
HFFF condition, will result in longer bridges and so virtual editing requires
more disk space
than with the HF condition. However, in the case of real editing, where the
original stream can
be discarded and only the edited stream is kept on disk, the second embodiment
will actually
save disk space. In a number of cases a group of partially filled fragments
will be reallocated
into a smaller number of fragments.
Next the third embodiment satisfying the 2/3F condition will be discussed. It
is
remarked that in general the larger the fullness required for fragments, the
more fragments are
required for bridge creation. For example, the HF condition could be replaced
with one where
the minimum fragment fullness was 0.75, that is, each at least partially
filled fragment PF
1 S satisfies
0.75 5 size (PF) <_ 1 [17~
This gives the same worst case average fragment fullness as the FFHF condition
of 0.75. To
maintain the 0.75 condition requires worst case bridges of four fragments,
even when the
original stream is FF. Therefore, this option is not as good as the HFFF
condition presented
above. A 2/3F condition gives a lower worst case average fragment fullness
than the HFFF
condition and consequently can be expected to require less reallocation than
the HFFF
condition. Figure 24 shows the starting point for creating a bridge assuming
that the original
2S streams satisfy the 2/3 F condition. Here D represents the MPEG part of the
bridge, that is, the
part of the streams that must be copied, recoded or remultiplexed to satisfy
the MPEG
requirements. No assumption can be made about the fullness of the fragment
directly
preceding the bridge or the fragment directly after the bridge because the
fullness of these
fragments will depend on the choice of edit points. A number of cases will be
considered
based on the fullness of the fragments before and after the bridge.
Figure 2S illustrates case 1. Here both the fragment preceding the bridge and
the fragment after the bridge are more than 2/3 full and the bridge contains
only the data
required to fulfil the MPEG requirements.
If
CA 02290641 1999-11-18
WO 99/48096 19 PCT/IB99/00439
B1 +D+B2 >_2/3 [18]
then there is no problem and the result is as shown if Figure 26.
Tf
B1 +D+B2> 1 [19]
then not all of B 1 and B2 will be copied.
If
2/3 + B 1 + D <_ 1 [20]
or
2/3 +B2 + D <_ 1 [21 ]
then there is no problem and the result is as shown in Figure 27 and Figure 28
respectively.
Assuming that 0 < B 1 + D + B2 < 2/3 and adding the contents of the other two
fragments gives:
4/3 < 2/3 + B 1 + D + B2 +2/3 < 6/3 [22]
In this case it is possible to reallocate the data in two fragments of at
least 2/3 as shown in
Figure 29. When the original stream was locally FF or not, has no effect on
the result in this
case.
Figure 30 shows the starting point for case 2. Here both the fragment before
the
bridge and the fragment after the bridge are less then 2/3 full. If
2/3 <_ B 1 + D + B2 S 1 [23]
then there is no problem and the results is as shown in Figure 31. Now there
are two situations
to consider:
B 1 + D +b2 < 2/3 [24]
and
B1+D+B2>1 [25]
*rB
CA 02290641 1999-11-18
WO 99/48096 2~ PCT/IB99/00439
First the first situation [24] is discussed. Adding all or part of C 1 and C2
to make up the 2/3
fragment still results in a problem if C1 + C2 do not yield enough data. This
occurs when the
following condition holds:
C 1 + B 1 + D + B2 + C2 < 2/3 [26]
Adding the rest of the previous and next fragments gives:
4/3 < 2/3 + C 1 +B 1 + D + B2 + C2 + 2/3 .<6/3 [27]
This data can be reallocated as two fragments as show in Figure 32.
Now the second situation [25] is discussed. This will only cause a problem in
the case where:
1 <B1 +D+B2+<4/3 [28]
Again all or part of CI and C2 can be added to make the total at least 4/3 and
so there is a
problem if
1 < C1 + B1 + D + B2 + C2 < 4/3 [29]
In this case adding the previous or next fragment gives
SI3 < 2/3 + C1 + B1 + D + B2 + C2 < 6/3 [30]
The data in this case can be allocated as two fragments as shown in Figure 10.
If the original stream was locally FF then C1=1/3 and C2=1/3 and in the first
situation [24] a single fragment is sufficient for the bridge. In the second
situation [25], two
fragments are still required for the bridge.
Figure 34 shows the starting point for case 3. Here the fragment before the
bridge is less than 2/3 full and the fragment after the bridge is greater the
2/3 full.
If
2/3 5 B 1 + D + B2 [34]
and
CA 02290641 1999-11-18
WO 99/48096 21 PCT/IB99/00439
B1 +D 5 1 [35]
then there is no problem and a single fragment can be used for the bridge.
There are two cases to consider:
B 1 + D + B2 < 2/3 [36]
and
1 < B 1 + D < 4/3 [3~]
Now in the first case [36] adding the rest of the fragment after the bridge
gives:
2l3 <B1 +D+B2+2/3 <4/3 [38]
There is a problem if:
1<B1+D+B2+2/3 <4/3. [39]
Some or all of C1 and C2 can be added to ensure that the data can fill two
fragments, so there
is still a problem if:
1 < C 1 + B 1 + D + B2 +2/3 + C2 < 4/3 [40]
Adding the rest of the previous or next fragment gives
5/3 < 2/3 + C 1 + B 1 + D + B2 + 2/3 + C2 < 6/3 [41]
This data can be reallocated to two fragments as shown in Figure 35.
Next the second case [37] will be considered. Adding B2 is possible but there
is still a problem
if:
1 < B 1 + D + B2 < 4/3 [42]
Adding the rest of the fragment after the bridge gives
CA 02290641 1999-11-18
WO 99/48096 22 PCT/IB99/00439
5/3<B1+D+B2+2/3<b/3 X431
This data can be reallocated to two fragments as shown in Figure 36. The
bridge will still
require two fragments even if the original stream was locally FF.
Concluding, it is possible to replace the HF condition with a condition where
the minumum fragment fullness is 2/3 of a full fragment. This requires a
maximum of two
fragments for a bridge. Editing from a locally FF stream will still require
two fragments for a
bridge in some case.
Whilst the invention has been described with reference to preferred
embodiments thereof, it is to be understood that these are not limitative
examples. Thus,
various modifications may become apparent to those skilled in the art, without
departing from
the scope of the invention, as defined by the claims. The disclosed fragment
size of 4 MB is
characteristic of one specific embodiment. Another embodiments may use other
fragment
sizes, such as 6 MB for example. Further, in this respect, it should be noted
that first
generation apparatuses in accordance with the invention, capable of carrying
out recording and
reproduction of a real time information signal, may be capable of recording
signal blocks of
fixed size SFA in the fragment areas only, whilst they are already capable of
reproducing and
processing signal blocks of variable size from the fragment areas in order to
reproduce a real
time information signal from a record carrier that has signal blocks of
variable size stored in
the fragment areas. Second generation apparatuses that are moreover capable of
carrying out
an editing step, will be capable of recording signal blocks of variable size
in the fragment
areas.
Further, the invention lies in each and every novel feature or combination of
features. The invention can be implemented by means of both hardware and
software, and that
several "means " may be represented by the same item of hardware. Furthermore,
the word
"comprising" does not exclude the presence of other elements or steps than
those listed in the
claims.
