Annex N: (void)

ETSI

Annex O (normative):

Interpretation of schedules for Alternative Frequency Signalling

The "Alternative frequency signalling: Schedule definition data entity - type 4" provides the functionality to restrict the availability of a list of alternative frequencies to certain time intervals based on a weekly schedule.

In every SDC data entity type 4 the following information can be signalled:

•With the Day Code field it can be indicated to which days of the week (Monday to Sunday) the following time range shall apply. Any day-combination from 1 to 7 days can be signalled.

•Using the Start Time and the Duration value, a time interval can be specified. This time interval applies to all specified days-of-the-week (using the Day Code).

The Start Time value indicates the minutes since midnight UTC (for every indicated day of the week), ranging from 00:00 to 23:59.

The Duration value specifies the number of minutes after (and including) the start time. It can potentially span more than one week. So for example it is possible to cover a full weekend using one single SDC data entity type 4.

•More than one time interval per day or different day-time-combinations can be specified by broadcasting multiple SDC data entities type 4 with the same Schedule Id (using the list mechanism for the version flag).

Every receiver has to evaluate these values in a consistent way. Therefore the following text defines how the receiver has to interpret the SDC "Alternative frequency signalling: Schedule description data entities - type 4". The function (presented in pseudo program code notation) checks whether the current time/date is within a scheduled time interval:

//input: time_in_week (minutes since last Monday 00:00 in UTC;

//the value is in the range 0 ≤ time_in_week < 60 × 24 × 7);

//schedule_id (id of the schedule to be checked;

//the value is in the range 0 ≤ schedule_id <= 15 )

//output: boolean value (true: time_in_week is inside schedule)

bool IsInsideSchedule (time_in_week, schedule_id)

{

// the schedule_id value 0 is fixed to 'always valid': if (schedule_id == 0)

{

return true

}

for every SDC entity with the given schedule_id

{

extract (day_code, start_time, duration) from SDC entity

for every day specified by day_code

{

//minutes_since_monday(day) returns the number of minutes

//of the start (00:00) of the indicated day since Monday 00:00

//(it is a multiple of 24 × 60)

//Wednesday -> 2 × 24 × 60 = 2 880, etc.

schedule_start = minutes_since_monday (day) + start_time schedule_end = schedule_start + duration

//the normal check (are we inside start and end?): if (time_in_week >= schedule_start AND

time_in_week <= schedule_end)

{

return true

}

//the wrap-around check:

minutes_per_week = 7 × 24 × 60

if (schedule_end > minutes_per_week)

{

ETSI

// our duration wraps into next Monday (or even later) if (time_in_week < (schedule_end - minutes_per_week))

{

return true

}

}

}

}

return false

}

The encoding for a certain time interval is not unique. A 48-hour interval starting on Wednesday 10:00 could be encoded as:

•"Wednesday and Thursday; start time 10:00; duration 24 hours".

•"Wednesday; start time 10:00; duration 48 hours".

•or using two SDC data entities with the same schedule id:

"Wednesday; start time 10:00; duration 24 hours" and "Thursday; start time 10:00; duration 24 hours".

•or using two SDC data entities with the same schedule id:

"Wednesday; start time 10:00; duration 10 hours" and "Wednesday; start time 20:00; duration 38 hours".

It is up to the encoding side to describe a certain schedule with as few SDC data entities as possible.

ETSI

Annex P (informative):

Transmit diversity

The DRM system is designed for various transmission environments with different delay spread and Doppler spread. Multipath environments with short and strong echoes, which typically occur in urban canyons, lead to a huge coherence bandwidth so that the channel can be described as flat instead of frequency-selective. Systems with a bandwidth smaller than the coherence bandwidth can accordingly suffer from flat fading. This is especially the case for robustness mode E. Time interleaving applied to the DRM system improves the performance of moving receivers in such circumstances.

A further method to overcome flat fading is antenna diversity, which means the application of more than one antenna at the receiver or transmitter. Antenna diversity at the receiver is effective but difficult to implement in small receiver boxes. For broadcast systems the use of transmit diversity is a good alternative or addition to receive diversity.

In the development of robustness mode E, different methods, such as space time coding and delay diversity, were evaluated. this work showed that delay diversity is the preferred choice because space time coding requires more overhead in the signal for channel estimation and it is more sensitive against the time-incoherence for fast fading channels.

The idea of delay diversity is quite simple. In addition to the original signal a delayed version of the same signal is transmitted from another spatially separated antenna. This method increases the channel delay spread by an additional echo with a comparable effect as with single frequency networks. The application of transmit delay diversity does not require any modifications at the receiver.

Figure P.1 shows how delay diversity can be implemented at the transmitter for an arbitrary number of antennas. After the OFDM modulation with an IFFT the signal path is split according to the number of antennas. Each signal path will be delayed by a chosen value δk before insertion of the guard interval.

Figure P.1: Transmit delay diversity scheme

The only parameters which have to be chosen are the delay δk of each path. Two requirements have to be considered:

•the delay δk should be large enough to increase the frequency selectivity of the composed channel that is the superposition of the channels for the transmit antennas;

•and it should be much less than the guard interval duration Tg in order to avoid inter-symbol-interference.

According to the first requirement the value δ should be at least 10 µs for a two antenna system in robustness mode E. This value also fulfils the second requirement because only 4 % of the guard interval duration gets lost. For further optimization the appropriate scientific literature is recommended.

The improvements which can be obtained with transmit delay diversity depend on the actual transmission channel. Simulations have been performed for the channel profiles described in clause B.2. They show that the SNR gain at a BER of 10-4 for the Terrain Obstructed profile at 60 km/h receiver speed is around 1 dB, for the Typical Urban profile at 60 km/h around 2 dB and the Typical Urban profile at 5 km/h more than 4 dB.

ETSI

Annex Q (informative):

Seamless reconfiguration

Clause 6.4.6 explains the mechanism used for reconfiguration, which can occur on a transmission super-frame boundary. Depending on the nature of the reconfiguration, the receiver may be able to follow the selected service without audio interruption. Table Q.1 indicates for which type of configuration this is possible.

Table Q.1: Cases of reconfiguration

In order for seamless reconfiguration to work, several parts of the chain need to behave correctly:

•The modulator or MDI generator should generate the appropriate FAC and SDC signalling to indicate the timing of the reconfiguration and the new parameters (see clause 6.4.6).

•The modulator should generate a continuous DRM signal through the reconfiguration, and should not clear the contents of any buffers or memories unnecessarily because of the change of parameters.

•The receiver should:

-Interpret correctly the signalling of the reconfiguration.

-Not clear the contents of any buffers or memories unnecessarily because of the change of parameters.

-Allow correctly for the delay through the de-interleaver when applying the new parameters.

One case where particular care should be taken concerns a change of MSC mode, i.e. the MSC constellation. Figure Q.1 shows the contents of the cell interleaver and de-interleaver following a change from 16-QAM to 64-QAM. For clarity, only the convolutional part of the interleaver, and not the pseudo-random part, is shown. The following points should be noted:

•Both the interleaver and de-interleaver contain a mixture of both types of constellation. The representation used in the interleaver and de-interleaver memories should therefore be able to deal with this mixture.

•In the signal on the air, the MSC cells in a given frame will contain a mixture of both types of constellation.

•For a given multiplex frame, the cells at the output of the de-interleaver are nevertheless all of the same constellation.

•The constellation type at the de-interleaver output will not change until the new constellation cells have worked their way through the de-interleaver, and so the change of parameters for the downstream processing should be delayed accordingly.

ETSI

•The number of cells in a multiplex frame has not changed, so the interleaver structure remains unchanged.

Figure Q.1: Contents of interleaver/de-interleaver system during change of MSC mode

ETSI