Sign in

username:

password:



Not a member?

Search speechcoding



Search tips

Subscribe to speechcoding



speechcoding by Keywords

ACELP | ADPCM | AMBE | AMR | AMR-NB | CELP | Codebook | DTMF | G.723 | G.726 | G.729 | GSM | Interpolation | LPC | LSF | LSP | MELP | PCM | Perceptual | Pitch | PSOL | QCELP | Quantization | SMV | VAD | Vocoder

Ads

Discussion Groups

Discussion Groups | Speech Coding | Re: G.723.1 Optimization

Technical discussions related to Speech Coding (all itu and other vocoders, ACELP, CELP, AMR, etc)

  

Post a new Thread

G729 Annex D Porting - Gram Smi - Mar 13 15:14:00 2003



Hi All,

First of all, I'm new in speech coding. Some days ago, I got the G.729 Annex D source code from ITU's page.

Now, I'm involved in porting this code to our system, with a H8/2633 (16 bits) microcontroller, and my first attempt has been awful.

¿Is it possible to do? ¿Am I crazy?

¿Which are the lowest hardware requirements to do it? ¿Is it mandatory to use a DSP?

Help please!!!!

Thanks in advance

 




Tu cuenta de correo gratuita Mixmail
Ya.com ADSL: Router U.S. Robotics ¡Gratis!





(You need to be a member of speechcoding -- send a blank email to speechcoding-subscribe@yahoogroups.com )

Re: G729 Annex D Porting - Mark Lakata - Mar 13 18:33:00 2003

For G.729, you need roughly a 100MHz 32 bit general processor to do
encode+decode with optimized C. With a DSP, you can do it in about 20MHz. A
64bit processor with DSP extensions can do it in about 30-40MHz. I have not
looked at the Annex D... I don't know how that affects the baseline
performance. You also need to be able to do multiplies very fast, about 2 16x16
multiplies per cycle on average. It appears that the H8S/2633 doesn't even have
a multiplier. So, it is basically impossible. Mark Lakata, Staff Engineer 1225 Charleston Road voice 650-567-5170
MIPS Technologies Mountain View CA 94043 fax 650-567-5002

On Thu, 13 Mar 2003, Gram Smi wrote: > Hi All,
>
> First of all, I'm new in speech coding. Some days ago, I got the G.729 Annex D
source code from ITU's page.
>
> Now, I'm involved in porting this code to our system, with a H8/2633 (16 bits)
microcontroller, and my first attempt has
> been awful.
>
> ¿Is it possible to do? ¿Am I crazy?
>
> ¿Which are the lowest hardware requirements to do it? ¿Is it mandatory to use
a DSP?
>
> Help please!!!!
>
> Thanks in advance >
________________________________________________________________________________\
___________________________________________________
>
> Tu cuenta de correo gratuita Mixmail
> Ya.com ADSL: Router U.S. Robotics ¡Gratis!
>
> Yahoo! Groups Sponsor
>
> ADVERTISEMENT
> width
> width
> _____________________________________ >
> Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.






(You need to be a member of speechcoding -- send a blank email to speechcoding-subscribe@yahoogroups.com )

Re: G.723.1 Optimization - Hemant - Mar 14 7:11:00 2003

Hi List,
I have been involved with G.723.1 for quite some time. Right now, my
implementation which is compatible with ITU G.723.1 is working fine. I want to
further optimize it.
Can someone guide me about the steps needed to do that ?
rgds,
Hemant Kumar





(You need to be a member of speechcoding -- send a blank email to speechcoding-subscribe@yahoogroups.com )

Re: G729 Annex D Porting - chetan k - Mar 14 9:09:00 2003

Hi,
U R not crazy, It is possible provided the processor
is running at arround 100 Mhz. It is recomended to
Implement G.729 on a DSP since the code is MAC
intensive, and DSP is meant for this purpose.

This can be Implemented on a low cost low power DSP
such as a TiC54x, which is a 16bit 100 Mhz processor.
Arround 7 channels of G.729 can be run on this
processor, ie only about 14% of processor speed is
used by a single channel G.729.

U can also choose a RISC microprocessor like ARM with
or without DSP extension's where about 25-30% of the
processor speed is used.

Good luck, go ahead and Implement. =====

With Rgds
Chetan k.




(You need to be a member of speechcoding -- send a blank email to speechcoding-subscribe@yahoogroups.com )

Re: G.723.1 Optimization - Author Unknown - Mar 20 20:58:00 2003

Hi Hemant:

This is Hemant who was at AT&T Wireless, right? I have c code for 723.1 which
you could use to create an optimized assembly language version.

Or perhaps you were asking for algorithmic optimization? I'm not sure. In any
case, if you want the c code let me know and I will try to find it in my pile of
things from AT&T Wireless.

Still getting all A's?????? Chuck
In a message dated 3/14/2003 2:11:58 AM Eastern Standard Time,
writes: >
> Hi List,
> I have been involved with G.723.1 for quite some time. Right now, my
implementation which is compatible with ITU G.723.1 is working fine. I want to
further optimize it.
> Can someone guide me about the steps needed to do that ?
> rgds,
> Hemant Kumar > _____________________________________




(You need to be a member of speechcoding -- send a blank email to speechcoding-subscribe@yahoogroups.com )