Forums

g.726 test vectors problem

Started by mste...@yahoo.com December 9, 2005
Hello,
I'm trying to run the test vectors for the G.726 (reset and homing) over the original code. They are in hexa format, so I have first converted them to binary using the cvt_h_b.c.
Then I feed g726demo with nrm.a(the binary version) and the corresponding parameters (A-law, load, 256 samples, first block =1, number of blocks = 64 and reset = 1, as depicted inthe makefile.cl) and I expect to obtain an output file that is identical to the binary version RN16FA.I file. But this is not true... the two files differ.
Any hint is more than welcome! :)
Best Regards,


Hey,

1 - the reference test vectors given along with G726 also in hex
format, Did you convert all your reference test vectors to binaries?

2- load, lolo, adlo are there command line options
lolo- log input and output
load- log input, adpcm output
adlo,- adpcm input, log output

log means pcm output

rn40fa.i is adpcm output, so the command line option shd have been
load. did you given that?

>
> 1. g.726 test vectors problem
> From: mstelian@mste... > ________________________________________________________________________
> ________________________________________________________________________
>
> Message: 1
> Date: Fri, 09 Dec 2005 02:56:19 -0500
> From: mstelian@mste...
> Subject: g.726 test vectors problem
>
> Hello,
> I'm trying to run the test vectors for the G.726 (reset and homing) over the
> original code. They are in hexa format, so I have first converted them to
> binary using the cvt_h_b.c.
> Then I feed g726demo with nrm.a(the binary version) and the corresponding
> parameters (A-law, load, 256 samples, first block =1, number of blocks = 64
> and reset = 1, as depicted inthe makefile.cl) and I expect to obtain an
> output file that is identical to the binary version RN16FA.I file. But this
> is not true... the two files differ.
> Any hint is more than welcome! :)
> Best Regards, >
Satheesh Ram,
Digibee Microsystems Plc (www.dgbmicro.com)



hi,
for running the homing test
have intailised the codec paramets to a predetermined value by first
executing the
"pcm_init.a" file for the coder.

this file intialises the codec parametrs and structure to a predetermined
state and then
u need to run the test case nrm.a

regards
vikas On 12/10/05, Satheesh Ram <satheesh.ram@sath...> wrote:
>
> Hey,
>
> 1 - the reference test vectors given along with G726 also in hex
> format, Did you convert all your reference test vectors to binaries?
>
> 2- load, lolo, adlo are there command line options
> lolo- log input and output
> load- log input, adpcm output
> adlo,- adpcm input, log output
>
> log means pcm output
>
> rn40fa.i is adpcm output, so the command line option shd have been
> load. did you given that? >
>
> >
> > 1. g.726 test vectors problem
> > From: mstelian@mste...
> >
> >
> > ________________________________________________________________________
> > ________________________________________________________________________
> >
> > Message: 1
> > Date: Fri, 09 Dec 2005 02:56:19 -0500
> > From: mstelian@mste...
> > Subject: g.726 test vectors problem
> >
> > Hello,
> > I'm trying to run the test vectors for the G.726 (reset and homing) over
> the
> > original code. They are in hexa format, so I have first converted them
> to
> > binary using the cvt_h_b.c.
> > Then I feed g726demo with nrm.a(the binary version) and the
> corresponding
> > parameters (A-law, load, 256 samples, first block =1, number of blocks =
> 64
> > and reset = 1, as depicted inthe makefile.cl) and I expect to obtain an
> > output file that is identical to the binary version RN16FA.I file. But
> this
> > is not true... the two files differ.
> > Any hint is more than welcome! :)
> > Best Regards,
> >
> >
> >
> Satheesh Ram,
> Digibee Microsystems Plc (www.dgbmicro.com)