Closed GoogleCodeExporter closed 9 years ago
In theory, it would be possible, but g7221 is not optimized :-( So I think the
CPU is
not enough strong.
Some tests are necessary.
Do you know SIP providers that use g722.1 ?
Original comment by samuelv0...@gmail.com
on 3 Oct 2009 at 9:53
G7221 is a fairly lightweight algorithm, so it may in fact fit within the
available
cpu cycles even without optimization.
I don't think you'll find a SIP provider who explicitly says they allow G.722.x
calls. But you should be able to test by calling directly to an IP address?
Original comment by hartnett...@gmail.com
on 5 Oct 2009 at 1:20
As g722 codec, g722.1 codecs work wi 16 or 32 kHz (issue 289), so we need to
resample
signal because Siphon works to 8kHZ.
And the resample step use much CPU.
Antoher things, for the moment it's impossible do disable g722 and not g722.1
Original comment by samuelv0...@gmail.com
on 11 Oct 2009 at 9:19
resampling the 8khz I/O to 16 or 32 will mean audio bandwidth is still below
4KHz so
there is no value in having resampled g.722.1
Original comment by hartnett...@gmail.com
on 13 Oct 2009 at 3:31
Issue 289 has been merged into this issue.
Original comment by samuelv0...@gmail.com
on 7 Nov 2009 at 11:51
I released 2.2 rc2 on http://code.google.com/p/siphon/downloads/list
I tried to include g722 and resample but I didn't test.
If you can try and tell me if it works. If yes, I will enable g722.1 too.
Original comment by samuelv0...@gmail.com
on 7 Nov 2009 at 11:52
Partially working! I can hear G.722 audio coming from my hardphone, but no G.722
audio is heard back on Siphon earpiece.
Also, G.722 is 4KHz filtered, I assume due to resampler. G.722 is not useful
unless
it is wideband (16KHz native sampling)
Original comment by hartnett...@gmail.com
on 9 Nov 2009 at 5:58
Fixed.
The RC 4 will include the fix.
Original comment by samuelv0...@gmail.com
on 16 Nov 2009 at 10:27
I just release RC 4 : http://code.google.com/p/siphon/downloads/list
Original comment by samuelv0...@gmail.com
on 1 Dec 2009 at 10:50
Original issue reported on code.google.com by
hartnett...@gmail.com
on 1 Oct 2009 at 5:56