10 [0116]
Fig. 8 is a diagram explaining that the NP insertion
unit 63n in the reception device 12 (Fig. 4) restores the
synchronized stream on the channel ch#n as the NP-inserted
stream on the channel ch#n.
15 [0117]
As described with reference to Fig. 4, the NP insertion
unit 63n properly inserts NPs to the received post-FEC stream
(the NP-deleted stream (in the BB frame format)) by providing
and storing the received post-FEC stream on the channel ch#n
20 from the FEC unit 62n in the buffer 64n, and then outputting
the packets in the received post-FEC stream stored in the buffer
64n or NPs. This insertion restores the synchronized stream
on the channel ch#n as the NP-inserted stream on the channel
ch#n.
25 [0118]
In other words, the oldest packet among the packets that
the NP insertion unit 63n writes (stores) in the buffer 64n
and that is not read from the buffer 64n in the received post-FEC
stream on the channel ch#n from the FEC unit 62n is a notable
30 packet to be noted.
[0119]
26
The NP insertion unit 63n inserts NPs into the received
post-FEC stream on the channel ch#n from the FEC unit 62n by
outputting NPs as many as the number indicated by the DNP added
to the notable packet, and then outputting the notable packet,
5 and restores the synchronized stream on the channel ch#n as
the NP-inserted stream on the channel ch#n.
[0120]
Thus, after a packet in the received post-FEC stream
on the channel ch#n written in the buffer 64n is determined
10 as the notable packet, it is necessary to store the packet
in the buffer 64n until the completion of outputting NPs as
many as the number indicated by the DNP added to the notable
packet.
[0121]
15 The data input (provided) and written to the buffer 64n
is referred to as the buffer input data, and the data read
and output from the buffer 64n is referred to as the buffer
output data.
[0122]
20 Fig. 8 illustrates exemplary temporal variations in the
total amount of buffer input data and in the total amount of
buffer output data.
[0123]
The buffer input data is the received post-FEC stream
25 on the channel ch#n. To simplify the description, it is assumed
that the buffer input data is input at a constant data rate.
[0124]
In this example, the total data amount of the buffer
input data increases with a constant inclination as a solid
30 line illustrated in Fig. 8.
[0125]
27
On the other hand, the total data amount of the buffer
output data varies as a dotted line illustrated in Fig. 8,
depending on the NPs inserted into the received post-FEC stream
on the channel ch#n when the synchronized stream on the channel
ch#n 5 is restored as the NP-inserted stream on the channel ch#n.
[0126]
In other words, the total amount of the buffer output
data increases when the NP insertion unit 63n outputs the
packets in the received post-FEC stream stored in the buffer
10 64n as the buffer input data because the packets in the received
post-FEC stream are read from the buffer 64n while the NP
insertion unit 63n outputs the packets in the received post-FEC
stream stored in the buffer 64n and NPs.
[0127]
15 However, the total amount of the buffer output data does
not vary (increase) when the NP insertion unit 63n outputs
the NPs because the packets in the received post-FEC stream
are not read from the buffer 64n.
[0128]
20 When the NP insertion unit 63n sequentially inserts many
NPs into the synchronized stream on the channel ch#n restored
as the NP-inserted stream on the channel ch#n, namely, into
the divided stream on the channel ch#n provided by the splitter
21 in the transmission device 11 (Fig. 2), the total amount
25 of the buffer output data does not vary. This is because the
packets in the received post-FEC stream are not read from the
buffer 64n when the NP insertion unit 63n continues outputting
NPs in a period in which many of the NPs are sequential.
[0129]
30 The total data amount of the buffer input data increases
with a constant inclination as described above. Thus, when
28
the total amount of the buffer output data does not vary, the
difference between the total amount of the buffer input data
and the total amount of the buffer output data (hereinafter,
referred to also as the total amount difference) increases.
5 [0130]
The total amount difference is the data amount of the
data stored in the buffer 64n. Thus, the maximum total amount
difference is the buffer amount that the buffer 64n needs
having.
10 [0131]
The total amount difference increases in a period in
which NPs are sequential in the divided stream on the channel
ch#n provided by the splitter 21 as described above because
the total amount of the buffer output data does not vary.
15 [0132]
Thus, when NP can freely be inserted without prescribing
(defining) an NP insertion method in which the splitter 21
inserts NPs into the divided stream on the channel ch#n (for
example, the ratio of NPs to be inserted into each channel,
20 or the order in which the NPs are inserted into each channel),
the buffer 64n in the reception device 12 needs having a large
buffer amount. This can increase the cost of the reception
device 12.
[0133]
25 When the buffer amount of the buffer 64n in the reception
device 12 is smaller than the total amount difference, this
causes the overflow of the buffer 64n. This may causes the
NP insertion unit 63n to fail to properly process the received
post-FEC stream on the channel ch#n from the FEC unit 62n (to
30 restore the synchronized stream on the channel ch#n as the
NP-inserted stream on the channel ch#n).
29
[0134]
In light of the foregoing, an NP insertion method is
prescribed for the transmission system in Fig. 1 so that the
splitter 21 divides an input stream into the divided streams
5 of the channel ch#1 to ch#N including the packets of the input
stream at a predetermined density.
[0135]
According to the NP insertion method, NPs are smoothed
and inserted into the divided streams of the channel ch#1 to
10 ch#N so that the NPs are not inserted only in a part of the
streams.
[0136]
This can eliminate the need for a large buffer amount
of the buffer 64n in the reception device 12 as described above
15 due to the fact that many NPs are sequentially inserted in
a divided stream of the channel ch#n.
[0137]
The data rate of the NP-deleted stream on the channel
ch#n to be provided to the FEC unit 33n in the transmission
20 device 11 needs to be a predetermined data rate lower than
or equal to the processing speed (the FEC rate) of the FEC
unit 33n. This is because, when the data rate of the NP-deleted
stream on the channel ch#n exceeds the processing speed of
the FEC unit 33n, it is difficult for the FEC unit 33n to properly
25 process the NP-deleted stream on the channel ch#n.
[0138]
The data rate of the NP-deleted stream on the channel
ch#n to be provided to the FEC unit 33n in the transmission
device 11 is the throughput of the data field in the BB frame
30 generated by the FEC unit 33n (the data from which the BB header
in the BB frame is removed). The data rate is determined in
30
accordance with the symbol rate, a modulation scheme for the
quadrature modulation performed by the MOD unit 34n, the
error-corrected code used in the FEC unit 33n, or the on/off
of the pilot signal for the modulated signal provided by the
5 MOD unit 34n.
[0139]
On the assumption that the data rate of the NP-deleted
stream on the channel ch#n to be provided to the FEC unit 33n
in the transmission device 11 is previously determined as a
10 predetermined data rate lower than or equal to the processing
speed of FEC unit 33n, the NP insertion method can be prescribed
so that the ratios of the NPs that the splitter 21 inserts
into the channels ch#1 to ch#N are identical to the ratios
of the reciprocals of the predetermined data rates of the
15 NP-deleted stream on the channels ch#1 to ch#N.
[0140]
According to the prescribed NP insertion method, the
FEC unit 33n can be prevented from failing to properly process
the NP-deleted stream on the channel ch#n due to the fact that
20 many NPs are inserted in a channel ch#n’ and this decreases
the number of NPs inserted into the other channel ch#n, and
thus the data rate of the NP-deleted stream on the channel
ch#n becomes larger than the processing speed of the FEC unit
33n on the channel ch#n.
25 [0141]
In other words, when an NP insertion method is prescribed
so that the ratios of the NPs that the splitter 21 inserts
into the channels ch#1 to ch#N are identical to the ratios
of the reciprocals of the predetermined data rates of the
30 NP-deleted stream on the channels ch#1 to ch#N, the data rates
of the NP-deleted streams of the channels ch#1 to ch#N become
31
the data rates lower than or equal to the processing speeds
of the FEC units 33n previously determined for the channels
ch#1 to ch#N, respectively. This enables the FEC unit 33n
to properly process the NP-deleted stream on the channel ch#n.
5 [0142]
Fig. 9 is a diagram of an exemplary stream that the
transmission system in Fig. 1 processes when the NP insertion
method as described above is used.
[0143]
10 Note that the number N of channels is two in Fig. 9,
similarly to Fig. 6.
[0144]
Furthermore, the illustration of ISCR is omitted in Fig.
9.
15 [0145]
The A of Fig. 9 is an exemplary input stream provided
to the splitter 21 in the transmission device 11 (Fig. 2).
[0146]
The A of Fig. 9 is similar to the A of Fig. 6.
20 [0147]
The B of Fig. 9 is exemplary divided streams provided
from the splitter 21 in the transmission device 11 (Fig. 2)
to the synchronization unit 31n via the buffer 22n.
[0148]
25 In accordance with the NP insertion method, the splitter
21 divides the input stream into the divided streams of the
two channels ch#1 and ch#2 so that the packets in the input
stream are included at a predetermined density by distributing
the TS packets in the input stream in the A of Fig. 9 to one
30 of two channels ch#1 and ch#2, and distributing NPs to the
other channel.
32
[0149]
In the B of Fig. 9, the ratios of the reciprocals of
the predetermined data rates of the NP-deleted streams on the
channels ch#1 and ch#2 is, for example, 1 to 1.
5 [0150]
Thus, the splitter 21 divides the input stream into the
divided streams of the channels ch#1 and ch#2 so that the ratios
of NPs inserted into the channels ch#1 and ch#2 is 1 to 1.
[0151]
10 In other words, in the B of Fig. 9, the splitter 21
distributes the TS packet #0 in the input stream to the channel
ch#1, and distributes an NP to the channel ch#2. After that,
the splitter 21 distributes the TS packet #2 in the input stream
to the channel ch#2, and distributes an NP to the channel ch#1.
15 The splitter 21 distributes the TS packets in the input stream
alternately to the channels ch#1 and ch#2, and distributes
NPs alternately to the channels ch#1 and ch#2. This divides
the input stream into the divided streams of the channels ch#1
and ch#2.
20 [0152]
As a result, the packets of the input stream are included
in both of the divided streams on the channels ch#1 and ch#2
at a constant density (the density at which 0.5 packet of the
input stream is included in a packet of the divided stream
25 in the B of Fig. 9).
[0153]
The C of Fig. 9 is exemplary NP-deleted streams provided
from the NP deletion unit 32n to the FEC unit 33n in the
transmission device 11 (Fig. 2).
30 [0154]
The synchronization unit 31n changes the divided stream
33
on the channel ch#n provided by the splitter 21 into the
synchronized stream on the channel ch#n.
[0155]
The NP deletion unit 32n changes the synchronized stream
5 on the channel ch#n into an NP-deleted stream on the channel
ch#n by deleting the NPs from the synchronized stream and
inserting DNPs.
[0156]
The C of Fig. 9 is the NP-deleted streams on the channels
10 ch#1 and ch#2 provided from the divided streams on the channels
ch#1 and ch#2 in the B of Fig. 9, respectively.
[0157]
The D of Fig. 9 is an exemplary transmitted post-FEC
stream on the channel ch#2 changed from the NP-deleted stream
15 on the channel ch#2 in the C of Fig. 9 by the FEC unit 332
in the transmission device 11 (Fig. 2).
[0158]
The transmitted post-FEC stream on the channel ch#2 is
a BB frame stream in which a BB header is added to one or more
20 packets of the NP-deleted stream on the channel ch#2 in the
C of Fig. 9.
[0159]
The E of Fig. 9 is an exemplary received post-FEC stream
on the channel ch#2 provided from the FEC unit 622 to the NP
25 insertion unit 632 in the reception device 12 (Fig. 4).
[0160]
As described with reference to Fig. 4, the FEC unit 622
restores the NP-deleted stream in the BB frame format provided
by the FEC unit 332 in Fig. 2, and provides the restored stream
30 as the received post-FEC stream on the channel ch#2 to the
NP insertion unit 632.
34
[0161]
Thus, the received post-FEC stream on the channel ch#2
is identical to the NP-deleted stream in the BB frame format
on the channel ch#2, namely, the transmitted post-FEC stream
5 on the channel ch#2 in the D of Fig. 9.
[0162]
The F of Fig. 9 is an exemplary NP-inserted stream on
the channel ch#2 provided from the NP insertion unit 632 to
the merging unit 52 in the reception device 12 (Fig. 4).
10 [0163]
The NP-inserted stream on the channel ch#2 in the F of
Fig. 9 is the stream obtained by inserting NPs as many as the
number indicated with the DNPs included in the received
post-FEC stream on the channel ch#2 into the packets (sequence)
15 included in the BB frame of the received post-FEC stream on
the channel ch#2 in the E of Fig. 9, and deleting the DNPs.
[0164]
Fig. 10 is a diagram of exemplary temporal variations
in the total amount of the buffer input data and in the total
20 amount of the buffer output data when the NP insertion method
described above is used in the splitter 21.
[0165]
Similarly to Fig. 8, it is assumed in Fig. 10 that the
buffer input data that is the received post-FEC stream on the
25 channel ch#n is input at a constant data rate. In this example,
the total data amount of the buffer input data increases with
a constant inclination as a solid line illustrated in Fig.
10.
[0166]
30 On the other hand, using the NP insertion method divides
the input stream into the divided streams on the channels ch#1
35
to ch#N including the packets of the input stream at a constant
density.
[0167]
Thus, the packets of the input stream exist at a constant
5 density in the synchronized stream on the channel ch#n that
the NP insertion unit 63n in the reception device 12 restores
as the NP-inserted stream on the channel ch#n from the divided
streams of the channel ch#n.
[0168]
10 Thus, when the NP insertion unit 63n restores the
synchronized stream on the channel ch#n as the NP-inserted
stream on the channel ch#n, the packets stored in the buffer
64n are read at average intervals. This increases the total
amount of the buffer output data at average intervals, as a
15 dotted line illustrated in Fig. 10.
[0169]
This can eliminate the need for a large buffer amount
of the buffer 64n due to the fact that the NP insertion unit
63n sequentially outputs NPs and the packets are not read from
20 the buffer 64n when the NP insertion unit 63n restores the
synchronized stream on the channel ch#n as the NP-inserted
stream on the channel ch#n, and this increases the total amount
difference between the total data amount of the buffer input
data and the total data amount of the buffer output data.
25 [0170]
Note that Fig. 10 illustrates the total amount of the
buffer output data when the splitter 21 divides the input stream
into the divided streams on the channels ch#1 and ch#2
alternately including the NPs and packets of the input stream
30 one by one illustrated as the B of Fig. 9.
[0171]
36

[0278]
5 Fig. 17 illustrates the format of ISSY.
[0279]
ISSY includes ISCR, BUFS, and BUFSTAT.
[0280]
ISCR is the time information indicating the time when
10 the packet is transmitted as described above, and is two or
three-byte information.
[0281]
BUFS is (actually) two-byte information indicating the
buffer capacity (required buffer amount) of the buffer 64n
15 required to store the received post-FEC stream when the NP
insertion unit 63n restores the synchronized stream as the
NP-inserted stream by inserting NPs into the received post-FEC
stream.
[0282]
20 The fifth and sixth bits from the top of the bit string
that is two-byte (First byte and Second byte) BUFS are referred
to as BUFS_UNIT. The BUFS_UNIT indicates the unit of the buffer
capacity indicated by the BUFS. Ten bits from the seventh
bit to the last 16th bit indicate the value of the buffer
25 capacity.
[0283]
For example, a storage area, which works as the buffer
64n having a buffer capacity indicated by the BUFS, is secured
in the NP insertion unit 63n of the reception device 12. The
30 received post-FEC stream is written to the buffer 64n while
the synchronized stream is restored as the NP-inserted stream.
55
[0284]
BUFSTAT is (actually) two-byte information indicating
the reading start time when the packet is read from the buffer
64n while the NP insertion unit 63n restores the synchronized
5 stream as the NP-inserted stream by reading the packets (of
the received post-FEC stream) stored in the buffer 64n.
[0285]
Note that the fifth and sixth bits from the top of the
bit string that is two-byte (First byte and Second byte) BUFSTAT
10 are referred to as BUFSTAT_UNIT. The BUFS_UNIT indicates the
unit of the reading start time indicated by the BUFSTAT. Ten
bits from the seventh bit to the last 16th bit indicate the
value of the reading start time. The 10 bits of the BUFSTAT
indicate the reading start time with the remaining amount of
15 data in the buffer 64n when the packets are read from the buffer
64n.
[0286]
The NP insertion unit 63n in the reception device 12
start reading the packets from the buffer 64n at the timing
20 (time) indicated by the BUFSTAT when the NP insertion unit
63n restores the synchronized stream as the NP-inserted stream.
[0287]
The ISCR in the ISSY described above is added to each
packet of the divided streams in the synchronization unit 31n
25 of the transmission device 11 so that the merging unit 52 in
the reception device 12 reconstructs the input stream.
[0288]
Thus, when only an ISSY can be added to each packet of
the divided streams and the ISCR is added to each packet, it
30 may be impossible to add the BUFS and BUFSTAT to each packet.
[0289]
56
Thus, the transmission device 11 needs transmitting the
signaling that is the BUFS and BUFSTAT in a method different
from the method for transmitting the ISCR.
[0290]
5 For example, similarly to the size identifier, the BUFS
and BUFSTAT can be included in the BB header of the BB frame
of which stream is generated by the FEC unit 33n.
[0291]
Fig. 18 is a diagram of the format of the BB header.
10 [0292]
For example, UPL or SYNC that is a field in which fixed
values are set in the BB header can be used as the BUFS and
BUFSTAT, similarly to the size identifier.
[0293]
15 In this example, when all of the size identifier, BUFS,
and BUFSTAT are included in the BB header, for example, the
BUFS and BUFSTAT can be included in UPL, and the size identifier
can be included in SYNC.
[0294]
20 Note that, for example, a one-bit size identifier can
be included in a one-byte SYNC in the BB header of each BB
frame.
[0295]
Each of the BUFS and BUFSTAT has two bytes. Thus, the
25 BUFS and BUFSTAT can periodically (regularly), for example,
alternately be included in a two-byte UPL in the BB header
of each BB frame.
[0296]

[0297]
57
Fig. 19 is an explanatory diagram of another method for
transmitting the size identifier, BUFS, and BUFSTAT.
[0298]
In the words, Fig. 19 is a diagram of the format of the
5 BB frame.
[0299]
As described above, the in-band signaling can be added
to the BB frame.
[0300]
10 There are two types of in-band signaling, IN-BAND type
A Signaling and IN-BAND type B Signaling in DVB-T2. In place
of the types of in-band signaling, new in-band signaling can
be used in the transmission system Fig. 1.
[0301]
15 As illustrated in Fig. 19, the new in-band signaling
includes a one-bit size identifier, a two-bite BUFSTAT_UNIT,
a 10-bit BUFSTAT (namely, the value of the transmission start
time), a two-bit BUFS_UNIT, and a 10-bit BUFS (namely, the
value of the buffer capacity).
20 [0302]
The new in-band signaling is added to a BB frame when
the FEC unit 33n in the transmission device 11 generates the
BB frame.
[0303]
25 TS is used as the input stream in the present embodiment.
Note that, however, a stream formed by a plurality of packets
can be used as the input stream instead of TS.
[0304]