Oliver Wilcock wrote:
> Hi Steve (and the list),
>
> A dialtone signal is the best the Meridian can do with the hardware I
> have.  I discovered the dialtone workaround in Cisco documentation.
>   
Actually, a Merdian can be configured to do a wide range of things. 
Figuring out just how can be the problem. :-)
> Any chance you can answer some of my questions?  I've reproduced a few
> below.  But first, you gave me new questions.  (Ahhh! Too many questions!)
>
> If dsp.c is being replaced, what, when, how, who?  Where can I read about it?
>   
Its crap, and and its being replaced chunk by chunk by functions in spandsp.
> When you say "dsp.c should report DSP_TONE_STATE_DIALTONE" what do you
> mean?   Only if call progress is enabled, I presume,  Report to whom?  And
> when you say "should" is there an architecture document I can read.  My
> first suspicion is that the idea may be to offload such features to
>   
If you look through dsp.c you will see that DSP_TONE_STATE_DIALTONE is 
generated when dial tone is detected. If these events are being 
processed, and this one occurs after disconnect, I think things will get 
confused.
> spandsp, but I didn't see evidence in dsp.c that this was occurring.  Does
>   
Evidence would be things like it already using the DTMF detector in 
spandsp, and some of the other tone processing. :-)
> it occur in the channel drivers instead?  Perhaps in chan_zap.c, for
> example.  Or is dsp.c intended to be the interface to spandsp?  O wait,
>   
The channel drivers will use spandsp directly when things are finished.
> goertzel_ comes from spandsp and I see your name in spandsp.h.  I guess it
> is too late to add a prefix to spandsp exports such as spandsp_goertzel_.
>   
Why would you want to do that?
> Back on topic, I think the features I'm interested in and willing to
> contribute are useful.  They are certainly useful to me, and if you search
> for asterisk, hangup and disconnect supervision you will find plenty of
> people having trouble many of whom were able to resolve or reduce their
> problems, and some who still have difficulty.
>   
If you need something quick, I think you should just try to fudge it for 
now. There is a better call progress tone detector in spandsp than the 
one in dsp.c, but it still isn't great. The great one is in progress. 
:-) I think I'm getting a little free time to work on it again, now, so 
it should be done soon.

Regards,
Steve

_______________________________________________
Callweaver-dev mailing list
[email protected]
http://lists.callweaver.org/mailman/listinfo/callweaver-dev

Reply via email to