[asterisk-users] FXO Cards - T38

[Rob Hillis]

T.38 is for all intents and purposes a codec.  It's purpose is to
re-encode a fax transmission as a data stream to be re-assembled at the
other end as if it were a fax call.  Seems to me to be pretty close to
the definition of a codec to me.

Your original comment was that you cannot use T.38 and G.729 in Asterisk
at the same time.  On a technical level, this is /not/ true, especially
if the T.38 implementation does not rely on SpanDSP. (whether or not
such an implementation exists is another question)  Breaking license
conditions is a separate issue altogether.

You also appear to have answered another one of your questions on this
forum to someone else ("why on earth would you want to remove SpanDSP as
a dependency?") by telling us that you can't run G.729 at the same time
as T.38.

I'm also curious as to why you assert that using G.729 in Asterisk
(/not/ ABE) at the same time as a T.38 implementation that relies on
SpanDSP since these are two completely separate plugins that are
installed and acquired separately.  That's almost like asserting that
you can't run any commercial X application if you've installed my XYZ
web browser on the same machine.  Just because they use a common
software base (X in this instance) /doesn't/ mean that you're violating
the GPL by running non commercial software on the same machine.

A more meaningful interpretation of the GPL would be that you either can
or can't run a T.38 implementation with Asterisk /full stop/.  Either
the license is compatible, or it isn't.  Trying to force any other
interpretation on people will end up with you being dismissed as an
extremist.


Steve Underwood wrote:
> T.38 is not a codec. A codec has one input and one output. T.38 is an 
> interactive protocol. This, however, has nothing to do with what I said. 
> If you use G.729 in the same asterisk as my spandsp library, you are 
> breaking my licence conditions.
>
> Steve
>
>
> Rob Hillis wrote:
>   
>> T.38 is a codec in exactly the same way that GSM or G.729 is a codec, 
>> so yes it /can/ be used at the same time as any other codec - just 
>> that only /one/ codec will be used at a time.  What often happens is 
>> that the call will initially be established with a codec such as G.729 
>> or G.711a, but once fax tones are detected the call will change codecs 
>> to T.38.
>>
>> According to the release notes for 1.6.0-b4...
>>
>>  - 11873, Added core API changes to handle T.38 origination and termination
>>    (The version of app_fax in Asterisk-addons now supports this.)
>>   
>>
>> This should be all that is necessary to run a T.38 gateway.
>>
>>
>> Steve Underwood wrote:
>>     
>>> Rob Hillis wrote:
>>>   
>>>       
>>>> Not unless you're running CallWeaver or Asterisk 1.6.0-beta4.  Asterisk 
>>>> has had passthrough support for T.38 for a while (somewhere in 1.4 it 
>>>> became available IIRC) but is currently completely incapable of 
>>>> terminating or encoding a fax call to T.38.
>>>>   
>>>>     
>>>>         
>>> I thought * was still not capable for T.38 gateway operation. Doesn't 
>>> beta 4 just added T.38 termination? And, I believe it misses out some 
>>> key elements of doing that properly. Note that T.38 termination is an 
>>> addon, so it can't be used with, say, G.729.
>>>   
>>>       
>>>> The only real option available at the moment is to keep one PSTN line on 
>>>> an ATA with an FXO port and T.38 support available and direct calls from 
>>>> the fax machines through to it.  However, I should point out that while 
>>>> I believe this should be possible, I haven't actually tried it myself.
>>>>
>>>>   
>>>>     
>>>>         
>>> Steve
>>>     
>>>       
>
>
> _______________________________________________
> -- Bandwidth and Colocation Provided by http://www.api-digital.com --
>
> asterisk-users mailing list
> To UNSUBSCRIBE or update options visit:
>    http://lists.digium.com/mailman/listinfo/asterisk-users
>   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20080224/eebdc4c9/attachment.htm