Thanks for Quick response.

I tried studying TSM code layer1 and Osmocom layer1 for purpose of implementing them in cellphone it self but i as far as i know that init of some variables required for amr are completed but couldnt be able to figure out what exactly part remaining .

also for decoding AMR convolution i made some changes to structures given by sylvain. I didnt know what i was doing but gave me correct output for encoding and decoding. but still i doubt its correctness. It would be great if some one could comment on these files.

may be need to pay attention soon On Thu, Feb 28, 2013 at 2:37 PM, Harald Welte laforge@gnumonks.org wrote:

Dear Akib,

I think it is best if you start to read the actual relevant GSM specifications if you want to implement part of it. The key word to search for is probably the 'active code mode set'.

I will try to check specs again. i am really doubting about hysteresis

part. cause i dont know whether its needed or not while storing raw frames. if any one know please let me know shall save my time.

At any given point in time, there can only be a set of four active codecs in an AMR connection. The codec currently used within that set is selected with CMI/CMR bits on the L1 of the air interface.

thanks for pointing out.

The selection of which four codecs are part of the currently active set is explicitly configured by signalling messages on RSL between BTS and BSC, as well as on Layer3 (RR) between BSC and MS.

On Thu, Feb 28, 2013 at 01:15:37PM +0300, Akib Sayyed wrote:

...

Dear Sylvain

While Decoding AMR do need to just decode and pack it into raw amr frames or we need to think about hysterasis and all parts also

also there are total 8 codec modes

but in band data have details of on 4 codec modes

which one to choose out of 8 codec modes from those 8 modes.

-- Akib Sayyed Matrix-Shell akibsayyed@gmail.com akibsayyed@matrixshell.com Mob:- +91-966-514-2243

--

• Harald Welte laforge@gnumonks.org

http://laforge.gnumonks.org/

============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)

Okies

I shall follow your instructions

On Thu, Feb 28, 2013 at 6:43 PM, Harald Welte laforge@gnumonks.org wrote:

On Thu, Feb 28, 2013 at 05:54:27PM +0300, Akib Sayyed wrote:

...

I tried studying TSM code layer1 and Osmocom layer1 for purpose of implementing them in cellphone it self but i as far as i know that init

of

...

some variables required for amr are completed but couldnt be able to figure out what exactly part remaining .

I remain with what I stated before:

you first need to understand the detailed operation of AMR in the context of GSM by studying the specs. I'm mostly talking about the encoding of the various bits, signalling of codec mode changes, etc. - not the mathematical background of the actual voice codec. It might also help to study related documents like the IF2 and RTP formats for AMR.

Only with that backgorund knowledge _and_ knowledge of the TSM30 L1 code as well as traces from actual AMR calls on A-bis and possibly Um, I think you will be in a position to fully understand the porblem and implement what is needed.

Asking for somebody to explain every little detail for you is not the way to solve any problem. If Sylvian, Dieter, Andreas or I sit down, read the code + dosc, and give minute instructions what needs to be done, this consumes more of our time than if we actuall implement it ourselves.

If you have _specific_ technical questions, please come back to the list. If you have a specific question, please include a precise reference to the chapter/section of the specification (+ exact version of the document), the exact function or data structure you refer to.

--

• Harald Welte laforge@gnumonks.org

http://laforge.gnumonks.org/

============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)