Hi everyone,
I am working on G726. I have the source code in C (from Sun Microsystems).
I am trying to test the algorithm using the ITU test vectors.
I have converted the NRM.M and OVR.M hex files to binary files using the
cvt_h_b.c (code from ITU site).
My code takes in linear PCM and outputs linear PCM so I have converted the
mu-law and a-law binary files
into linear PCM files.
I have two things to say :
1) When I send in NRM_PCM test vector to my code, and look at the ADPCM
output values
The first 95 samples are bit exact but then after that there are errors (not
all of them but
about 7300 samples). And the PCM output does also match even for the first
95 samples.
(I have converted the RN16FM.O file also to PCM linear to compare).
2) When I send in OVR_PCM test vector as input to my code and look at the
ADPCM values,
all of them are bit exact. But if I look at the output PCM values of my code
and the test vector
given (RV16FM.O) they do not match up.
Could someone tell me if I am doing any thing wrong in testing or if there
is any obvious
mistake that I am not able to see. Please do help and I will get back to
you.
Thank you,
Aditi.
