How to adjust LMR VOX threshold

[Hans Johnson]

Hello,

I have a mostly functioan LMR setup going, using a VIC2-2E/M in a 3825.  I am hoping to use this to allow phone users to link to our radio network.  Unfortunately, the VOX as generated by the cisco router is at too low of a threshold, so the router keys the radio based even on minor noises.  I have the system setup as a simple dial-peer and voice port, and am trying ot use the "lmr ip-vad" function.  Is there a way I can set the threshold for ip-vad so that it will not key up the radio on background noise?

Here is my configuration:

voice-port 0/0/0

auto-cut-through

operation 4-wire

type 2

signal lmr

lmr e-lead voice

lmr duplex half

lmr ip-vad

lmr led-on

input gain 5

output attenuation -16

no echo-cancel enable

no comfort-noise

timeouts call-disconnect 3

timeouts ptt rcv 3

timing hangover 300

music-threshold -10

threshold noise -60

!

dial-peer voice 16 pots

destination-pattern 2505

port 0/0/0

What happens is that when you dial 2505, it connects and will key up the radio when you talk, but it will aso just randomly key up the radio causing all sorts of havoc.  What I am looking for is a way to modify the threshold so it takes more voice energy to key up the radio.

Thanks,

Hans

[Hermanus Janse van Vuuren]

Have a look at the following document from page 8 onwards:

http://www.cisco.com/en/US/docs/interoperability_systems/c_ipics/211/design/guide/sr211lmr.pdf

llowing shows a sample configuration for an LMR voice port that is configured for VAD operated

signaling.

In this example, type { 2 | 3 | 5 } typically is type 3, but refer to

Figure 3-3 on page 3-5

,

Figure 3-4 on

page 3-6

, and

Figure 3-5 on page 3-7

to select the type that best matches your radio requirements. Input

gain { -27 - 16 } typically is 10, but adjust this value as needed to best receive audio on Cisco IPICS

endpoints. Output attenuation { -16 - 27 } typically is 10, but adjust this value as needed to best receive

audio on radios. When connecting a radio to a voice port in an LMR gateway, you may need to make

adjustments to properly balance the audio levels. A

radio typically provides gain adjustments, and the

level of the signal from the radio to the voice port and the level of the signal from the voice port to the

radio may require some adjustments on the radio and

the voice port. When using a tone controlled radio,

it is important to note that the tones that are sent fr

om the LMR gateway to the radio also are affected by

the voice ports output attenuation settings. When opti

mizing these settings to achieve the desired audio

levels, take care to ensure that th

e voice port adjustments do not have

an adverse effect on the level and

quality of the tone signals.

voice class permanent 1

signal timing oos timeout disabled

signal keepalive disabled

signal sequence oos no-action

!

voice-port 0/2/1

voice-class permanent 1

auto-cut-through

operation 4-wire

type { 2 | 3 | 5 }

signal lmr

lmr e-lead voice

lmr duplex half

lmr led-on

input gain { -27 - 16 }

output attenuation { -16 - 27 }

no echo-cancel enable

no comfort-noise

timeouts call-disconnect 3

timeouts wait-release 3

timing hookflash-in 10

timing hangover 40

connection trunk 102

description VAD Operated Voice Port

threshold noise -40

!

dial-peer voice 102 voip

destination-pattern 102

session protocol multicast

session target ipv4:239.193.1.2:21000

codec g711ulaw

vad aggressive

[Hans Johnson]

Hi Hermanus,

I've read through the documentation that you cited prior to posting the question, but I'm not sure how it helps (perhaps it does, and I'm just missing it).  In the above config snippet you posted, lmr e-lead voice means that it will seize the e-lead when it receives voice packets from the network, relying on the far-end VAD to do the voice gating/VOX stuff.  This is problematic since I'd lke to be able to, say, conference the radios with a remote caller as well as someone local.

The configuration directive of "lmr ip-vad" is supposed to set it up so that the DSP watches the inbound voice packets from the network, and only cause the port to seize when there is actually voice energy in those packets.  The problem is that it's too sensitive.  I was/am hoping there is a way to adjust these thresholds.

Thanks for your time.