I'll take a look at it. Our issue boils down to the fact all the volumes are the same, announcements and user voice, and we'd like to have the announcements about half the volume when the repeater is in use. I'm looking at holdofftelem=1 as well as http://docs.allstarlink.org/drupal/node/126 to see if we can achieve a workable compromise. Thanks, Bob kk6ecm Sent from iPad On Jan 5, 2016, at 11:11 AM, Ken <ke2n@cs.com> wrote:
The audio was setup using the prescribed method, without the "MK-1 calibrated ear." If we raised the rx level, and reduced the tx level, then we'd have difficulty hearing the station id. I suspect the level of our custom recordings is too high, an I was trying to avoid the brute force method of redoing each one.
Thanks, Bob kk6ecm
Record level is tricky in Asterisk. When seen on any editing program, "full scale" is -6 dB f.s. or 0.5 on the voltage scale. I suspect the reason for this is that uLaw only produces about 14 bits when converted to linear and that is what asterisk expects for full scale. (I guess losing those two bits is equivalent to dividing by four or 6 dB down). Have a look at some of the Allison GSM files by importing them into Audacity and you will see the max level is 0.5. Even this is too loud for most people (as newbie postings have indicated many times). I think a reasonable level is 3 dB down from full scale - in other words, minus 9 dB relative to full scale (!) or about 0.3 on the voltage scale.
Ref: http://images.ohnosec.org/app-rpt-audacity.pdf
Regards Ken KE2N