Audio streaming in mesh network. 1km and more

Hi Mikhail,

Using Z-Stack for voice over ZigBee is not a good idea. :)

(Even TI experts saying that Voice Over Zigbee, if achieved such ability, should

be limited to network depth <= 2 )

Please, read the following thread for more info on this subject.

Regarding cc110x, I'm not familiar with this device, nor with the applications it can carry out,

so it probably worth some further investigation.

Br,

Igor

Igor, thanks for advice.

>Using Z-Stack for voice over ZigBee is not a good idea. :)

Why not?

I am read clearest speech codecs usually work @ 64Kbps.

In XBee devices at 115200kbit/sec. real speed was about ~40kbit/sec.

On 250kbit/sec. real speed must be ~80kbit/sec... in theory :-)

I am never work with Z-Stack.

Z-Stack is to heavy? And easy to realize something like rs485 only on wireless ?

Regards,

Mikhail

Why not.. using Zstack for audio streaming will clog the network

(again, as TI guys saying it, you have the physical limitation of ZigBee network depth <= 2

for voice streaming and that is, I assume, theoretically speaking).

The a-law/u-law provides 64Kbit/sec, so I think some sort of additional compression

must be implemented here, like g726 (and I really doubt that cc2530 cpu can handle it).

It not just that Z-Stack heavy... Z-stack uses a simple (small) operation system (AKA OSAL)

to manage the running tasks (like NWK task, MAC task, ZDO, your application etc.),

the transition time between two tasks is about 2mSec and one more thing, since cc2530

based on 8051 core, it is not a parallel peace of hardware, therefore some heavy function

can postpone other tasks (just lets say for 20mSec, and even in the lowest sample frequency 

of 8Khz for voice sampling, you cannot afford such latency).

Additionally, once Z-stack running on cc2530/1 it leaves a little amount of RAM for the user

(with some tuning and no security enabled, for Routers and coordinator ~1.3K, End Devices ~4K ),

So, little available RAM for buffering is also an issue here.

And so on and on... :) 

One more peace of information, in these days I'm doing experiments with ZigBee coordinator,

12 routers and a single End Device, where the coordinator used like a data concentrator.

In some cases (when End device is roaming) it takes about ~20mSec for a  packet of data

to get to coordinator once it leaves the End device and hoping 3 routers on its way (Application

acknowledgement is ON in this case).

So, in the bottom line, I'm not saying it is not possible it just very difficult. :)

If you wish to proceed with your idea of Voice Over ZigBee (using Z-Stack) I'll be happy to

collaborate and hear about your progression, as one of the goals in the firm where I working

now Is to achieve such ability (even for intercom application).

Br,

Igor