Hey there, Stanley_4. Happy to help clear this up. There isn't much documentation on Group Coordinator as it only appears in the diagnostic information. This is, however, important to be aware of. In general, using a wired player as your Group Coordinator will reduce the amount of data needing to be transmitted from the source to players.
A handy way to determine what is the Group Coordinator is to look in the Rooms Tab and see what room is listed at the top of the group. If the Group Coordinator is removed from the group, Sonos will automatically handoff to another player in the group. The way it selects this player has changed a bit in the past (alphabetical by MAC address) and will likely continue to change as we make improvements to the software.
Let us know if you'd like us to build an FAQ or write up an in-depth post about the finer details of Group Coordinator. I'd be happy to get that ball rolling.
That answered my questions, I'll play with things a bit to get used to doing my groups this way.
A FAQ would be nice but this looks to be a rare issue and only impact large setups or ones with signal issues.
Is there a reason that the Sonos devices don't give priority to wired components over wifi for the coordinator job? Might be something to think about doing so folks with larger systems like the fellow in the post I linked to above automatically avoid this issue.
Alphabetic by MAC ordering is going to give priority to your older devices over newer ones, I can see doing that now as the oldest device is the most likely to be wired but in my setup it is going to hit ancient ZP80s here rather than newer Play 3s or 1s. Does the age of the Sonos hardware matter?
Thanks!
I think I'd like to see a FAQ on it... it could be helpful to point folks to if they have a SonosNet/Boost mode system with signal issues in groups.
Interesting that the GC handoff is based on MAC address. I would have thought that changing based on Root/Secondary/Tertiary nodes would be a better choice, going with a speaker of the highest node type, then the lowest path cost to the root node (if the root node isn't part of the group).
Obviously that doesn't work in WiFi setups... in that case, going with MAC address is probably as good as anything else... though I think I'd favor higher MAC addresses over lower ones, as those would probably be newer devices with the better wireless setups (more antennas, newer wireless cards, etc.).
Dropping the coordinator doesn't always result in the lowest MAC from the remaining nodes being elected as new coordinator. It looks like it might be the fastest to respond, since my PLAYBAR gets selected over a ZP80.
Second-guessing the optimum topology is a bit more complex than picking a node of the highest STP order, given than grouped nodes will directly connect peer-to-peer if they can. As a rule of thumb though, starting with a wired node as coordinator if possible makes a good start.
@Stanley_4 what you are seeing is same as what I see on my system:
https://en.community.sonos.com/advanced-setups-229000/sonosnet-dropouts-6798584
If you start playing on a wireless player, then add an ethernet player you double the wireless traffic, vs starting on ethernet player and add wireless to group. This was confirmed by Sonos support when they looked at diagnostics.
I always try to group starting with ethernet connected player first, to avoid dropouts using lossless formats.
The irony, playing 'lossless' has losses due to dropouts of sonosnet choosing an inefficient GC (wireless vs wired)
SonosNet doesn't choose the GC, at least at the outset. The user does, by starting the group from a certain room/zone as noted. (In any case SonosNet is a layer 2 entity, and can't choose anything as such application-wise.)
I guess I meant sonos software? rather than sonosnet. What I have to do when I want to group is first consciously select the wired sonos player, play music, then add wireless players to group. If I select multiple players before playing, sonos software does not automatically select the the wired player as the GC, hence double the sonosnet traffic and dropouts. It would be nice if the sonos software could choose the GC to minimise the wireless, rather than the user.
Okay, understood. There would be one or two usability implications though. If B was wired, A was wireless, and B was added to A, the ordering of the rooms in the controller would become B:A (not A:B) if the wired node was favoured as GC. This could be contrary to expectations. Also, since the queue would presumably have to be copied to B it could obliterate the previous queue there, which might be unwelcome.
As a 'user' I don't care about sonosnet or need to know what a GC is or how it works. As a 'techie' I now understand these functions a little more and can work around their design. However, it would be nice not to have to think about which player is wired when grouping zones etc. To get optimum performance for lossless formats (minimise occasional dropouts), I sometimes have to select a wired player, group, then turn volume right down on wired player if that room is not required to be part of the group.
As a 'user' I don't care about sonosnet or need to know what a GC is or how it works. [snip] It would be nice not to have to think about which player is wired when grouping zones etc. [snip] I sometimes have to select a wired player, group, then turn volume right down on wired player .
Very interesting. I may need to add some wire to my wireless setup so I can play FLAC files.
They make some very nice flat or 2mm diameter Ethernet cables these days, hiding either is a lot easier than fat Cat 5.
I have a follow on question with regard to the GC. I have extended my Sonos system, still using SonosNet, I have added a Sub and a stereo pair to a 'room'.
If I select that 'room', will the sonos software automatically use the wired device in that room as the GC? ie can I run a Ethernet cable to Sub (easiest) and sonos software will use the Sub as the GC for that room?
A SUB is never the GC. In a pair (with or without a SUB) the left unit is the GC of that bonded set. (A bonded set is a special type of group.)
Where a pair (again with or without a SUB) is part of a group of rooms it gets a little more complicated. In that case:
- if the room is the first one in the group, its left unit is the GC of the group
- if the room is not the first, then each channel (and the SUB) takes a feed from the GC elsewhere
In your case wiring the SUB will mean that the pair are most likely to make a SonosNet connection to it. Since it's presumably short range it should be a strong link. Choosing that room as the first in the group will make the left unit the GC. Whether it will make a connection to other rooms via the wired SUB or directly depends on the circumstances.
Where a pair (again with or without a SUB) is part of a group of rooms it gets a little more complicated. In that case:
- if the room is the first one in the group, its left unit is the GC of the group
- if the room is not the first, then each channel (and the SUB) takes a feed from the GC elsewhere
Huh, I always assumed that in the case where multiple speakers make up the zone, One of the speakers (left unit or playbar/playbase) sort of acted as the spokesman (spokeswoman..spokesit...whatever) for all the speakers in the zone. If the zone is not the CG, then the spokesman receives the audio (left and right audio) from the CG. It then forwards on the right channel audio to the right speaker, and the bass to the sub. If the zone is a 5.1 setup, the playbar/playbase will always be the one communicating to the other units in the zone. That especially made sense to me since the 5.1 zone units are communicating over 5.0, not 2.4.
Overally, I guess I assumed it was more object oriented in that manner. Zone A didn't really need to know anything about the speakers in Zone B or communicate with them. I can see where it would be more efficient to allows speakers to communicate directly with each other regardless of zone.
In the case of PLAYBAR/BASE I believe you're correct, since the PB applies band-pass filters and EQ for the surrounds. (Those areas of the internal diags are now inaccessible so I can't recheck.)
With everything else, both channels (and SUB) subscribe individually to the stream from the GC, then apply a channel map at their end.
Thankyou ratty. So In my situation, it would be best to cable the L speaker, rather than the Sub?
If I cable Sub it will flow:
Ethernet --> Sub -SonosNet-> L (GC) -SonosNet-> Sub
If I cable L speaker instead it will flow:
Ethernet --> L (GC) -SonosNet-> Sub
I have mostly Apple Lossless on NAS, and want to minimise the SonosNet traffic.
If wireless bandwidth is a concern and you're able to wire the L speaker then, yes, that would be the best option.
