text stringlengths 1 112 |
|---|
bis subevent timings for the larger khz packets which wastes airtime if that causes a |
problem an alternative is to transmit the packets in separate bigs in this case that would |
mean one big for the khz sampled packets and another for the khz sampled ones the |
downside it that this adds complexity and doubles the am ount of advertising as each needs its |
own advertising set implementers need to decide which works best for their specific |
application |
robustness in a bis |
the way the parameters of a bis are defined is a little different as without a cknowledgments |
we need to employ some different strategies to maximise the chance of a transmission being |
received |
a bis still has a burst number bn and a number of subevent s nse defined in the same |
way as that for cises to compensate for th e lack of acknowledgements the core introduces |
the concept of a group count within isochronous intervals which is the ratio of nse to |
burst number nsebn group count is used to determine the way that retransmissions |
are arranged in a bis allocating them to groups which are used to add diversity into the |
transmission scheme these groups have nothing to do with broadcast isochronous groups |
it is an unfortunate use of the same word for two totally different concepts |
chapter isochronous streams |
figure the effect of group count |
figure shows two examples of a bis each of which contain s six subevent s so nse |
in the first case we have set a burst number of that means that each of the three payloads |
which are available for that bis event are sent twice in the second example on the right |
there are the same number of subevent s but with a burst number of so we have three |
retransmissions of the two payloads each individual subevent is given a group number g |
that goes from up to group count so in the left hand diagram we have a group count |
of comprising group and group in the right hand example there are three groups group |
group and group |
an important th ing to note is that although these two examples look the same they will have |
different isochronous intervals assuming that our frames are ms the example on the left |
will have an isochronous interval of ms as it contains two payloads p and p the |
example on the right will have an isochronous interval of ms as there are three payloads |
p p and p |
group count works with two other new parameters |
immediate repetition count irc which defines the number of groups which |
carry data associated with the current event and |
pretransmission offset pto the concept of pre transmission is to allow early |
transmission of packets in the hope th at a receiving device can receive audio data |
packets early allowing it to power down and then be ready to render them at the |
point that the final transmission opportunity occurs |
these parameters which are calculated by the scheduler in the controller replace flush |
timeout for a cis flush timeout is essentially an end stop terminating transmissions of a |
pdu in most cases its never needed because as soon as an initiator receives an |
acknowledgment that its data has been received it can close th e event and stop transmitting |
section broadcast isochronous streams |
a broadcaster cant do that it has to keep on transmitting therefore its advantageous to |
maximise the diversity of transmissions of packets to give every acceptor the best chance of |
finding a packet the sooner they c an do that the sooner they can go to sleep until the next |
one arrives |
pretransmission offset works by introducing the concept of subevent s associated with a |
current event as well as subevent s associated with future events this allocation of subevent s |
to pdus is even more complex than the scheme for cis so the best was to explain it is to go |
through a series of examples showing the effect of changing the values this is all worked out |
by the controller and is not accessible to an application but it he lps to have an understanding |
of it when you set the hci parameters to configure your streams |
to illustrate the concept of future subevent s figure demonstrates a bis event where we |
have an nse of eight subevent s the first five of which are associated with the current bis |
event the last three of which are used for data from future bis events |
figure the concept of future bis events |
we can now put everything together with some more examples |
these group numbers g which we first saw in figure along with the i mmediate |
repetition count i rc determine which payloads are sent in each transmission slot according |
to the following rules |
if g irc group g shall contain the da ta associated with the current bis event |
if g irc group g shall contain the data associated with the future bis event that |
is pto g irc bis events after the current bis event |
chapter isochronous streams |
well now look at a few examples which illustrate how these different parameters result in a |
variety of different retransmission schemes |
figure pretransmissions with nse bn irc and pto |
in figure we have four subevent s with an nse of the burst number of means |
one new payload is provided within each bis event so the isochronous interval for this |
example is ms the immediate repeat count is three which means that the data associated |
with each event is transmitted in the first three slots the conditions shown above show that |
the last transmission in that bis event comes from the next bis event with this scheme |
within every bis event there are always three transmissions of the current data plus one |
transmission of the data from the next event |
in case it seems like weve invented time travel we havent weve just ben t our definitions |
slightly event x cant happen until we have both the p and p pdus available so p is |
really the current data in event x this demonstrates that as soon as pto is gr eater than |
the latency starts to increase as the sync refence point cannot occur until after the last possible |
transmission of data which for p is in event x |
section broadcast isochronous streams |
figure pretransmissions with nse bn irc and pto |
figure shows what happens when we increase the number of subevent s to here the |
irc remains at but as nse is that provides two subevent s which are associated with data |
from future bis events these are used to pre transmit a packet from event x and event |
x that means that transmissions are being spread across more isochronous interval s with |
data coming from three consecutive bis e vents that helps to provide robustness against |
bursts of wide band interference which may last more than one isochronous interval but it is |
at the expense of greater latency which has grown by another m s |
figure pretransmissions with nse bn irc and pto |
chapter isochronous s treams |
figure looks at the case where we are spreading even further by increasing the pre |
transmission offset to this results in the inclusion of data from an x bis event and an |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.