[Asterisk-Users] Open G.729A codec

[Steve Underwood]

Hi Dan,

Dan wrote:

>----- Original Message ----- 
>From: "Steve Underwood" <steveu at coppice.org>
>To: <asterisk-users at lists.digium.com>
>Sent: Wednesday, August 13, 2003 9:49 AM
>Subject: Re: [Asterisk-Users] Open G.729A codec
>
>
>  
>
>>Steve Underwood wrote:
>>After writing this I got curious about how fast/slow the ITU reference
>>code really is. I built and ran the G.729A reference code on a 2.4GHz
>>Xeon machine, running RedHat 9. Its actually a dual Xeon, but the test
>>is only able to use 1 CPU.
>>
>>G.729A is the fixed point reduced complexity version of G.729. Reduced
>>complexity means it needs about half the MIPs of the more complex
>>version. I compressed and decompressed a 3.5 minute file of 16 bit
>>linear speech. It took 25 seconds to compress and about 5 seconds to
>>decompress. That ratio seems about right for a codec of this type. So,
>>using this code you can only do 7 bidirectional channels, using 100% of
>>a 2.4GHz Xeon. Not exactly great, huh?
>>
>>Perhaps I should try the floating point version. That might perform
>>somewhat better on an x86 machine, as any scaling and saturation steps
>>need not be performed.
>>
>>For comparison, can anyone tell me how fast the VoiceAge codec runs? If
>>is isn't a *lot* faster than that I would be rather surprised. It should
>>be algorithmically more efficient, and I assume as a commercial product
>>it should be using MMX, SSE and/or SSE2.
>>
>>    
>>
>
>There is any test made in the same conditions using Asterisk's GSM codec?
>I am interested in the scaling possibilities when using hardware IP phones
>with G.711 codec (like Cisco's 79x0, ATA, Budgetone, SNoM, etc.) and remote
>IAX connections plus all the local voice prompts and voicemail using GSM
>codec.
>
The GSM 06.10 codec is *much* less complex than G.729. Mark said he had 
over 200 channels of GSM running on a server a long time ago, but I 
don't know what that machine was. 06.10 isn't that great a codec, 
though. I don't think it is used very much on the GSM networks these 
days. Most of the time they use the enhanced full rate (EFR) or half 
rate codecs.

I just tried the ITU reference floating point code for G.729. Its 
considerably faster. It encoded a 3.5 minute speech file in less than 
5s, and decompressed it in about 1s.

I have no idea whether you are permitted to base an implementation on 
anything in the reference code. The code says its copyright by "AT&T, 
France Telecom, NTT, University of  Sherbrooke, Conexant, Ericsson. All 
rights reserved." but that doesn't make a whole lot of sense. Once you 
have the code, it is mearly impossible to make a clean room 
implementation, and its hard to get a clean room implementation bit 
accurate unless you play with the reference code. I couldn't find a 
clear statement about what you are permitted to do. I assume as long as 
you cough up the patent licence fees they wouldn't care too much, but 
who knows.

Regards,
Steve