[asterisk-dev] Bugid 5162: Packetization

Dan Austin Dan_Austin at Phoenix.com
Thu Aug 10 20:02:02 MST 2006


I've been trying to keep anthm's original patch up to date with svn, and
avoiding trying to address the modifications needed to make it IAX
friendly.

Not because I didn't want to, but because I suspected the result would
be
garbage.  After exchanging some ideas with mantis user SB, I'm not so
sure.
His ideas closely matched my own, and even our first passes at the code
looked
similar.

Attached is a patch that demonstates the basic direction.  This is NOT
1.4
Material, it does require trivial changes to a couple header files.
Still
I'd like to find out if this is the wrong direction while it is early.

High level details-
	1.  Extend AST_FORMAT_LIST to include these values-
		Default payload in bytes, default payload in ms
		minimum payload in ms, maximum payload in ms and
		payload incriment.
	2.  Move the function ast_rtp_lookup_smoother from rtp.c
		to frame.c and rename it to ast_lookup_smoother
	3.  Modify ast_lookup_smoother to use AST_FORMAT_LIST instead
		of RTP_TABLE 

The attached patch is a hybrid between where I think it should go and
anthm's original.  In fact other than loading the values from the config
files and a change to chan_sip to confirm the values are properly
stored,
the code does not use the vaules.

If this approach is acceptable, I'll continue to work with SB to finish
the RTP implimentation and the see what it will take to make IAX happy.

So comments are more that welcome.  I am pretty sure my C is lacking,
but
syntax and style can be fixed if the design is valid.

Thanks,
Dan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: frame-packetization.diff
Type: application/octet-stream
Size: 20641 bytes
Desc: frame-packetization.diff
Url : http://lists.digium.com/pipermail/asterisk-dev/attachments/20060810/94efc248/frame-packetization.obj


More information about the asterisk-dev mailing list