Bad (popping clicking interrupts) audio when running ASL 1.01 USBRADIO on Intel ATOM boards

I tried D510MO & D945GCLF2D boards with same result. HPET is enabled.
DAHDI_TEST Best: 99.999% – Worst: 99.192% – Average: 99.930983%

Pops noticable when transmitting tones, beeps or whistle and causes bad transcoding to DMR, voice transmissions are OK.

Upd: radio tune txvoice gives clean signal with no pops

Some folks will vehemently disagree with me on this, but that’s why I don’t like USBradio. SimpleUSB performs much better I my humble opinion.

On the other hand, you are talking DMR here, which is really a pile of poop… just saying.

UPD: 90% of pops affects rx input, PL tone from uplink sounds as if it was DPL.

In some cases DSP functionality of USBradio is too much useful. Previous time i made repeater, it was D945GCLF2D board + Xipar + GM300 RX + Nucleus transmitter, it was pretty nice to generate PL by DSP. And by the way i had no pops on Xipar.
In my current project Debian distro is prefferd, just to be able to easily install lots of additional ham radio packages not supported by centos repos.

I hate DMR’s poor half rate audio too, but unfortunately cheap chineese radios and flexibility made it popular with lots of nodes and activities as a result, so i would like to be able to dial in there from my app_rpt nodes. Actually P25 is supported too but unfortunately much less nodes and activity there.

Transcodiing from IAXrpt to DMR is awesome, transcoding from analog node initially affected by poor voice spectrum (would be nice to be able to expand it with some kind of neural algorithm before transcoding), finally degrades by pops which are, for some reason much more noticeble to AMBE vocoder than to ear.

You’ve got dahdi timing issues. If dahdi_test generates momentary results below 99.9%, you will have tone generation issues, CW ID issues, timing problems, etc. This is the first issue to fix. Once your dahdi_test numbers are perfect, then move on to looking for other problems.

73, David KB4FXC

1 Like