[Asterisk-Users] Digium cards, so disappointing !

Tim Panton tim at mexuar.com
Sat Apr 15 01:07:58 MST 2006


On 15 Apr 2006, at 06:53, George Pajari wrote:

> Kevin:
>
> You wrote:
>> FAX transmission is massively more complex than modem  
>> transmission. At
>> higher speeds, it involves 3 or 4 different 'carrier' frequencies and
>> signaling rate shifts, and these are done with very critical timing
>> requirements.
>>
>
> I'm sure you didn't quite mean to write what you have said above.  
> Fax transmission builds upon exactly the same ITU-T standards as  
> data transmission. For example, 33.6 kbps fax transmission (so  
> called "Super G3") uses the same V.34 standard as 33.6 data modems.  
> At slower speeds, fax modems use a half-duplex standard that is  
> less complex than that used by full duplex modems running at the  
> same speed.
>
> If there is a problem with Digium hardware handling fax signals it  
> cannot be laid at the purported "massively more complex" signalling  
> of fax transmissions.

I think it is more an issue of being massively more complex to  
support than 'pure' data
e.g. a few data channels on a T1 . I don't think the comparison was  
with analog modems.

The problem (as I see it) is that there are some tricky compromises  
to be made
when engineering a system like Asterisk. The Asterisk developers have  
concentrated
on their main aim - Voice over IP - and engineering decisions are  
made based on that.

Humans can't hear small phase shifts, but they can hear latency, so  
asterisk is optimized
to try and reduce latency at the cost of some (possible) phase shifts.
Fax machines (and V34 modems) can't cope with phase shifts, but can  
(probably) cope
with a bit more latency - if it is constant.

In my view, it would be possible to re-engineer the channel and  
jitterbuffer code to support
fax and V34 by tuning it to do (just) that. The amount of work  
involved would be huge
and almost certainly not worth it.


Tim Panton
tim at mexuar.com






More information about the asterisk-users mailing list