Hi Alex,
hope your client is not the Texan company RevCord ;o)
If you are comfortable with that you can send the troublesome code
off-list and I'll check in-depth.
>From what i can now guess, you don't need an extra merge_port as you
can directly use the decoded RTP buffer buffer in your pseudo-code
like this...
while(RTP decode into PCM frames and such into &frame) {
pjmedia_frame frame;
frame.buf = your RTP buffer
frame.size= the size of the current frame, rtp payload size
frame.type = PJMEDIA_FRAME_TYPE_AUDIO;
pjmedia_put_frame(rev, &frame);
}
Cheers,
Alain
On 19-Dec-11 11:30, Alex Balashov wrote:
> Alain,
>
> On 12/19/2011 05:20 AM, Alain Totouom wrote:
>
>> the splitter port is the one connected to the bridge. That port is
>> in sync with the bridge. You'll use it's slot ID to do the mixing.
>> The reverse port is the port you'll async and exclusively read/write
>> from/to.
>> The bridge is just aware of the splitter (slot)...
>>
>> pjmedia_port* sc, *rev;
>> int slot;
>> pjmedia_splitcomb_create(...,&sc);
>> pjmedia_splitcomb_create_rev_channel(pool,sc,..,&rev);
>> pjsua_conf_add_port(.., sc,&slot);
>
> Understood. In that case, I understood your original post correctly
> and did all this.
>
>> /* async write */
>> pjmedia_frame rdf;
>> rdf.buf = ..;
>> rdf.size= ..;
>> rdf.type= ..;
>> pjmedia_port_put_frame (rev,&rdf);
>
> I think here is where I am running into the problem. I don't
> produce the frames endogenously within this async write function;
> instead, they come from the outside, read from a PCAP packet
> capture, and RTP-decoded into PCM frames in the usual manner. So,
> I'm doing something like this with the PCM frames in that capture
> processing loop:
>
> while(RTP decode into PCM frames and such into &frame) {
> pjmedia_put_frame(&merge_port, &frame);
> }
>
> And here's how I have implemented the 'merge_port':
>
> pjmedia_port merge_port;
> pj_str_t merge_port_name = pj_str("merge");
> ...
> memset(&merge_port, 0, sizeof(pjmedia_port));
> ...
> pj_strdup2(pool, &merge_port.info.name, "merge_player");
>
> merge_port.get_frame = &merge_get_frame;
> merge_port.put_frame = &merge_put_frame;
>
> pjmedia_port_info_init(&merge_port.info, &merge_port_name,
> 0x1, 8000, 1, 16, (8000 * PTIME) / 1000);
>
> And the callbacks for 'merge_port':
>
> static pj_status_t merge_get_frame(pjmedia_port *port,
> pjmedia_frame *f)
> {
> return PJ_EINVALIDOP;
> }
>
> static pj_status_t merge_put_frame(pjmedia_port *port,
> const pjmedia_frame *f)
> {
> if(f->size == 0)
> return PJ_SUCCESS;
>
> return pjmedia_port_put_frame(conf_port, f);
> }
>
>
> And here is what I end up with:
>
> recombine: ../src/pjmedia/splitcomb.c:742: rport_put_frame:
> Assertion `frame->size == this_port->info.bytes_per_frame' failed.
>
> A little backtracing reveals:
>
> #6 0x0000000000404d90 in merge_put_frame (port=0x241, f=0x0)
> at media_proc.c:653
>
> So, for some reason it's getting a NULL frame. Any idea what I'm
> doing wrong?
>
--
""
(o)(o)
_____o00o__(__)__o00o_____
3072D/146D10DE 2011-09-29 Alain Totouom <totouom at gmx.de>
PGP Fingerprint 39A4F092 FFA7C746 CC305CB0 69091911 146D10DE
