Hi Bill,
Thank you for the explanation.
The only conceptual think I am not clear for my pjsua application now is
if I can block in my audio device until my hardware has or requires audio data.
Best Regards
Dimitar
From: Bill Gardner
Sent: Friday, March 27, 2015 15:56
To: pjsip at lists.pjsip.org
Subject: Re: TDM sound device to the pjmedia, how?
Hi Dimitar,
You want to write a new audiodev following pjmedia/include/pjmedia-audiodev/audiodev_imp.h,
then set up a macro to enable/disable this device, such as PJMEDIA_AUDIO_DEV_HAS_TDM,
and then add the hooks in pjmedia\src\pjmedia-audiodev\audiodev.c so pjmedia will use this device,
and then enable the device in config_site.h and disable the others.
You can then enable/disable the hardware echo cancellation by using the pjmedia capability
PJMEDIA_AUD_DEV_CAP_EC, then pjmedia can use the built-in AEC.
Regards,
Bill
On 3/27/2015 8:30 AM, dpn at switchvoice.com wrote:
Hi All,
I?ve ported pjproject-1.16 on an embedded platform based on Mediatek MT 7620 CPU running OpenWRT.
(Version 1.16 is rather old but I don?t need video and it seems it covers my needs. )
On this platform there is SLIC chip supporting FXS interface to an analog phone.
The SLIC has TDM interface to the CPU and I have completed the TDM driver (running in the Linux kernel space)
which works fine with Asterisk.
Now I am trying to test pjsua (or some other basic pjsip user agent? ) so I would like to integrate my driver to pjmedia.
I?ve checked the very clean pjsip documentation but still I am not fully clear how to approach the project.
I think using PortAudio or ALSA will overcomplicate my experiment.
My TDM driver expose ioctl interface which I can use from my pjsua application from the user space.
I think the best would be to derive the application clock from my tdm driver.
As far as I understand I should use ?Sound Device Port Framework? so eventually the application hooks
two callbacks rec_cb and play_cb which are called when my TDM driver has frame of data or needs frame of data.
I am looking at /pjproject-1.16/pjmedia/src/pjmedia-audiodev/null_dev.c which looks as the simplest starting point I could hack.
However sound devices in pjmedia provide lot of stuff which I don?t think I need.
For example acoustic echo cancelation. I don?t need that.
I rather use line echo canceler in my TDM driver which is more important I think for my softphone example.
From the other hand the ?Media Ports Framework? looks much cleaner than ?Sound Device Port Framework?.
For example /pjproject-1.16/pjmedia/src/pjmedia/null_port.c looks much simpler than null_dev.c.
But as far as understand using this framework the application calls the functions get_frame and put_frame so what I could do
is to block the execution of those functions (using linux system call select() for example)
and return when I get/send the frame data from/to my driver. This way I will get the timings provided from my driver to the pjsua.
However I am not sure how this will fit with the SIP part of my test softphone.
Anyone willing to clarify if I am on or fully out of track?
Thanks
Dimitar
_______________________________________________
Visit our blog: http://blog.pjsip.org
pjsip mailing list
pjsip at lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
--------------------------------------------------------------------------------
_______________________________________________
Visit our blog: http://blog.pjsip.org
pjsip mailing list
pjsip at lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.pjsip.org/pipermail/pjsip_lists.pjsip.org/attachments/20150327/4ac2e0db/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/png
Size: 1046 bytes
Desc: not available
URL: <http://lists.pjsip.org/pipermail/pjsip_lists.pjsip.org/attachments/20150327/4ac2e0db/attachment.png>
