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

hi friends, I have just started working with AMR speech codec. I would like post few general queries which are written below.... 1)What is the application of different bit rates(12.2,10.2,7.95,7.40,6.70,5.90,5.15,4.75 kbits/sec in practical field? 2)After using the auto-correlation why we are using lagging window?? 3)By using Levinson- Durbin algorithm we are transforming the coefficients into LP filter coefficients. Then These LP filter coefficients are transformed into LSP coefficients. But why do we need to transform back the LSP filter coefficients into LP filter coefficients? 4) What are the major functions in the standard AMR encoder which effect the MIPS rating of the encoder?? Anyone experienced in this field please let me know the answers of the above queries. Thanks in advance bye Subrata Chatterjee

______________________________
New Code Sharing Section now Live on DSPRelated.com. Learn about the Reward Program for Contributors here.



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

hi Subrata! i can try to answer the last 2 queries.. query 3. LPC coefficients are the coeffts of the SYNTHESIS filter. Since they are not suitable for direct quantisation,they are converted to LSP coeffts to guarantee the stability of reconstruction filter at the receiver end. now, once u get the LSP coeffts at the decoder, u need to convert them back to LPC coefficients again because that's how u will have a synthesis filter to reconstruct the original speech. query 4: the best way to find out which functions take the max time and contribute heavily in terms of MIPS is to PROFILE the code... do this for all the test vectors. this feature is available in VC and also in any other processor's simulator/cross compiler.(the only problem is that it is a very time consuming business... u will find out soon!) Sameer. --- subrata_etc <> wrote: > hi friends, > I have just started working with AMR speech codec. > I > would like post few general queries which are > written > below.... > 1)What is the application of different bit > rates(12.2,10.2,7.95,7.40,6.70,5.90,5.15,4.75 > kbits/sec in practical field? > > 2)After using the auto-correlation why we are using > lagging window?? > > 3)By using Levinson- Durbin algorithm we are > transforming the coefficients into LP filter > coefficients. Then These LP filter coefficients are > transformed into LSP coefficients. But why do we > need > to transform back the LSP filter coefficients into > LP > filter coefficients? > > 4) What are the major functions in the standard AMR > encoder which effect the MIPS rating of the > encoder?? > Anyone experienced in this field please let me know > the answers of the above queries. > Thanks in advance > > bye > Subrata Chatterjee > ------------------------ Yahoo! Groups Sponsor > > _____________________________________ > __________________________________________________

______________________________
Start your Android Ice Cream Sandwich development on TI's AM35x Sitara ARM Cortex-A8 processor today.



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

hello as I know.. 1) Diffrence rate is to adapt to nework status, not for the application 2) LPC analysis does not provide accurate estimation of spectral envelope for female speakers. To overcome this problem auto-correlations are multiplied by lag window. For more detail see, Chapter 12 (2.2.2) Speech Coding and Synthesis Edited by W.B kleijn and K.K Paliwal Elsevier This is one of the bibles of the speech coding 3) Dynamic range of LPC so wide that It is difficult to quantize efficiently That is reason why LPC is converted to LSP see, Digital Speech Coding for Low Bit Rate Communication Systems A.M.Kondoz JOHN WILEY & SONS *********************************************************** À̵¿¿ø() »ï¼ºÀüÀÚ Á¤º¸Åë½ÅÃÑ°ý / Åë½Å¿¬±¸¼Ò IMT2000´Ü¸»¿¬±¸ÆÀ Core±â¼ú¿¬±¸±×·ì / DSP sub-part ¼±ÀÓ ¿¬±¸¿ø Phone : 031-279-5451 H.P : 016-9530-5451 Dong-Won Lee Core Tech Research Team/ Telecommunication R&D center Information and Communication Business SAMSUNG ECELTRONICS CO., LTD. ***************************************************************

______________________________
New Code Sharing Section now Live on DSPRelated.com. Learn about the Reward Program for Contributors here.



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