text stringlengths 1 112 |
|---|
x bis event effectively spreading the audio data transmissions over five events and adding |
a total of ms to the latency compared with what it would be with pto in this figure |
were only showing the arrows for group and group in that first bis event x otherwise |
the figure becomes very cluttered and difficult to read |
figure pretransmissions with nse bn irc and pto |
finally in figure weve set nse to burst number and irc to and the pre |
transmission offset is also with these settings each bis event includes the two packets |
for that current event transmitted twice one after the other as we saw in figure with |
the final two subevent s used for packets two events in the future because bn is the |
isochronous interval will be ms as the pto is p and p come from two isochronous |
intervals in the future so the latency is ms greater than the simple example of figure |
these bis parameters allow for some very flexible transmission schemes in order to cope with |
different requirements in terms of latency and robustness in the next chapter well see how |
they can be used for different broadcast situations |
in conclusion b oth flush timeout and pre transmission offset increase latency for a cis |
flush timeout helps to share battery savings between both initiator and acceptor because |
the use of acknowledgements means that events can be closed with a bis wher e there are |
no acknowledgements a broadcast transmitter has to transmit at every subevent pto |
effectively gives all of the power savings to the acceptor by providing a diversity of |
transmission that gives it the best chance of acquiring a packet |
section broadcast isochronous streams |
the c ontrol subevent |
the control subevent core vol b does not need to be included within every big |
event typically control subevent s are used for items such as channel map updates so only |
occur occasionally when that information needs to be provided to all of the devices that a re |
receiving the broadcast audio streams the advantage of using control subevent s is that |
devices no longer need to scan to find out the basic big information such as the hopping |
channels that are being used th is minimise s the amount of work that a rec eiver has to do to |
ensure that it stays synchronised with a big and the bises within it without them a |
broadcast sink would need to continue to receive the periodic advertising train and |
periodically examine the biginfo to discover any changes |
a contro l subevent is transmitted in six consecutive big events it may then be transmitted |
in any subsequent big event but only one control event can be transmitted at a time every |
bis header for a bis subevent in a big event which includes a control subevent must have |
the control subevent transmission flag cstf set to to signal the presence of a control |
subevent in that big event its control subevent sequence number cssn will tell a |
receiver if it is the same as one which they have already received control events use the same |
frequency hopping scheme as every other subevent taking the index of the first bis event in |
the same big event |
big synchronisation |
as with connected isochronous stream s individual receiving devices typically a pair of |
earbud s or hearing aids dont necessarily know about each others existence to keep their |
audio in synchronisation they need to use information thats included within the big to |
understand when they need to render it to enable them to do this a big synchroni sation |
point is defined which coincides with the end of the transmission of the audio data because |
we have no acknowledgments to take into account in broadcast that big synchronisation |
point is located exactly at the end of the last bis transmission of the last bis in a big |
normally that will be coincident with the end of the big event however for the case when |
there is a control subevent present the big synchronisation point remains at the end of the |
last bis subevent transmission whilst the big e vent extends to the end of the control |
subevent as shown in figure |
chapter isochronous streams |
figure big synchronisation |
at the big synchronisation point every device that is listening to any of the broadcast |
isochronous stream s within that big will know that every other device has received its data |
hence the big synchr onisation point is a fixed point in time at which they can apply the |
presentation delay this is defined higher up the stack and dictates the point at which audio |
needs to be rendered the important point here is that audio is n ot rendered at the big |
synchronisation point its rendered a t the end of the presentation delay which commences |
at the big synchronisation point as broadcasting consists only of transmissions from a |
broadcast s ource there is no concept of presentatio n delay applied to captured data coming |
back from an acceptor |
figure the application of presentation delay in a big |
the derivation of the big synchronisation point is also a little different for broadcast unlike |
unicast every bis has the same basic timing parameters thats because they need to be |
defined in the biginfo and there is not enough space to allow different setting s for individual |
bises this means that the big synchro nisation point is a fixed interval from the big anchor |
section broadcast isochronous streams |
point both of which are known by every acceptor as theyre in the biginfo in contrast |
each cis uses a cissyncdelay based on its own anchor point as it doesnt know the |
relative timings of any other cises and cant assume they are the same as its own |
hci commands for bises |
as with connected isochronous streams a host application can define the sdu interval the |
maximum sdu size and the maximum transport latency which is the maximum time |
allowed for the transmi ssion of a bis data pdu as with a cig it is up to the scheduler in |
the controller to use these to guide it in defining the actual link layer parameters ie nse |
bn irc and pto |
setting up a big is much sim pler than a cis largely because there is no communication with |
any of the receiving devices so only two hci commands are required lecreate big |
configures and creates a big with a n umbis number of bises within it and the |
leterminatebig command end s and removes it there are no options to add or remove |
bises everything is done in a single operation both commands apply only to an initiator |
which is transmitting one or more bigs |
there are two similar commands for an acceptor which want s to recei ve one or more bises |
from within a big a process which is called synchronising with a bis they are the |
lebigcreate sync command and lebig terminatesync command but before we |
get to those we need to look at how an acceptor can find a broadcaster and connect to it |
finding broadcast audio stream s |
when we looked at connected isochronous stream s we didnt talk about how the devices |
initiated the connection the reason is that devices using connected isochronous stream s |
connect in exactly the same way as every other bluetooth le device using the normal |
advertising and scanning procedures they pair bond set up acl links discover each others |
features and then get on with setting up the streams if they are a coordinated set cap |
procedures ensure that all members of that set have the same procedures applied to them |
with broadcast audio stre ams none of that pairing and negotiation process happens because |
there is no connection instead receiving devices need to find a way to discover what is being |
transmitted when it is being transmitted and work out how to synchronise to it |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.