[Asterisk-Users] feature - VM gain adjust?

[Rich Adamson]

> At 5:00 PM -0600 on 7/11/04, Rich Adamson wrote:
> >I'm toying with adding a feature request to provide some sort of
> >gain setting for voicemail when accessed from "certain" interfaces.
> >Maybe something like voicemail=6.0 (db) within a specific channel
> >section of zapata.conf corresponding to a pstn line.
> >
> >Situation:
> >1. Someone calls into asterisk and leaves a voicemail. The sound
> >is recorded at some volume well below 0 db, and is directly related
> >to the distance asterisk is from the central office (pstn cable
> >loss) plus whatever distance the user placing the call is from
> >his/her central office.
> >2. I receive a text message that a voicemail was left.
> >3. I call into asterisk remotely (assume from a cell phone) and
> >retreive the voicemail. My location creates another xx db of loss
> >between myself and asterisk, and voicemail can hardly be heard.
> >
> >Actual Measured Values:
> >1. Asterisk is 5.6 db from the central office. Called from one
> >pstn line, through the central office, to asterisk and sending a
> >1004 hz tone at 0db. Recorded the tone into voicemail. (Tone should
> >have been recorded at about 11.2db, two times the cable loss)
> >2. Called into asterisk again, this time to retreive the voicemail
> >and measured the 1004 hz tone from voicemail. It was -36db "actual".
> >This retreival added another 11.2db of loss due to pstn interfaces
> >and plant loss.
> >3. The calls were through a TDM FXO module with rx and tx gains
> >set to 0. (Changing rx and tx gain to +3 db and repeating the test
> >resulted in a measured -30.5db signal, but these settings create
> >unwanted echo issues. Therefore adjusting channel gain is not an
> >option.)
> >
> >The end result is that retreiving any voicemail message left from
> >a distant location and retreived from a distant location can hardly
> >be heard. By adding the proposed voicemail=6.0 statement to the
> >appropriate channel, any calls connected to voicemail via that
> >channel would effectively increase transmission levels by 6db (or
> >whatever the setting happened to be). In this example case, the
> >setting would increase the vm volume by 12db (or about 24db measured
> >in the above).
> >
> >Anyone have any thoughts on this?
> >
> >Rich
> 
> Rich -
>    I'll say that this would be very useful.  Regardless of where the 
> loss is being inserted, it still exists.
> 
>    I like the idea of associating the voicemail db adjustment on a 
> per-channel basis.  I don't want to have to dink around with yelling 
> at the telco to fix something that "just works" otherwise.  Their 
> answer will be "Well, turn up the volume on your phone!" which is 
> exactly what your proposed patch will do.  A simple trial-and-error 
> process should be able to sort out the proper adjustment on any 
> typical system that doesn't have radical db changes across time.  I'm 
> heartily in favor of this idea; I'll even throw a donation towards 
> it, if you have a PayPal account.
> 
>    Another cool feature would be app_volume, which would turn up/turn 
> down tx/rx levels dynamically, but that's left for a different day, 
> and after we have an enhanced app_dial that lets single-digit dtmf 
> sequences jump to dialplan routines and then can reconnect bridged 
> calls.  See my various rantings about this in months (years!) past. 
> When I get some spare time (ha ha ha) I should really learn how to 
> code this stuff...
> 
> JT

The above "feature" request has been entered as bug #2023.

It also appears that VM has an issue (by itself) with recording/playing
volume. Transmitting a 1004hz tone at 0db through a ata186 (set for
-1db fxs loss), and then retreiving the same VM results in that tone
measured at ~ -10db. Doing the same from a pstn location (via TDM FXO)
suggests the same -10db loss (in addition to the pstn loss). Zapata.conf
rxgain and txgain set to 0. Using CVS-HEAD-07/12/04, but same result
with CVS-HEAD-07/1/04. Entered as bug #2022.

Add comments to either if you'd like.