Re: [Openbts-discuss] can I adjust the sample rate to 400khz * 4

[John Wu <jw...@gm...>]

thanks

On Thu, Aug 5, 2010 at 11:48 AM, Joshua Lackey <jl...@th...> wrote:

> The training sequence is actually in the middle of the burst.  So, yes,
> you can miss the first tail bits entirely and still be able to center
> your burst.
>
>
> Quoting John Wu (jw...@gm...):
> > so even if the burst lost some start tail bit, it can also be sync by
> > interpolation?
> >
> > On Thu, Aug 5, 2010 at 6:50 AM, Joshua Lackey <jl...@th...> wrote:
> > > Quoting Sylvain Munaut (24...@gm...):
> [...]
> > > > How do you deal with phase alignment ?
> > >
> > > The burst is first correlated with the training sequence.  Then the
> peak
> > > of the correlation is calculated using interpolation.  Because of
> > > interpolation, the index of the peak will be at a fractional index and
> > > not necessarily at an integer index.  (I.e., beacause of interpolation
> > > the resolution is higher than even 4 samples per symbol.) The burst is
> > > then centered using that peak.
>