|
Hi all, I was looking for a c implementation on C5409 for A-law compression (both encoder and decoder). It would be really helpful if anybody could provide me the same or links from where i could get the c code... Thanks and Regards kp |
|
|
A law c code
Started by ●April 10, 2003
Reply by ●April 11, 20032003-04-11
|
Hi, If your DSP has got a bootloader in the on-chip ROM, then there a built-in 256-word A-law expansion table that you can use. This table is located at address 0xFD00. Make sure that you map the on-chip ROM into data space (DROM=1). Remember that A-law can only handle 12-bit linear. You need to right/left shift the linear sample in order to work with 16-bit or other width linear. The C code for A-law expansion (decoding) would look like this: // built-in expansion table in on-chip ROM int *alaw_exp_table = (int*)0xFD00; int *ulaw_exp_table = (int*)0xFC00; // decode a-law into 12-bit linear int alaw_expand(int sample) { // make sure sample is between 0 and 255 return alaw_exp_table[sample & 0xFF]; } For A-law compression (encoding), using lookup table would be inefficient. You can use the following C code to compress 12-bit linear into 8-bit A-law. #define MASK 0x0080 #define BIAS 0x01F0 #define INVT 0x00D5; #define ABS(x) (((x) > 0) ? (x) : -(x)) // encode 12-bit linear into a-law int alaw_compress(int sample) { int result; int exp; result = sample >> 6; exp = (result) ? _lnorm(result) : 0x1F; result &= MASK; result += BIAS; result -= (exp * 16); result += (ABS(sample) >> (32 - exp)); return result^INVT; } The function _lnorm() is an intrinsic function equivalent to EXP assembly instruction. Hope it helps. --- In , Benoy Antony <rec_kp@y...> wrote: > Hi all, > I was looking for a c implementation on C5409 for A-law compression > (both encoder and decoder). > > It would be really helpful if anybody could provide me the same or links > from where i could get the c code... > > Thanks and Regards > kp |
