Re: RF signals are pure noise in Python implementation

[Verónica Toro Betancur]

Hi Marcus,

Thanks for the explanation.

In fact, the block I'm having trouble with is self.iio_pluto_source_0_0, which is defined as

pluto_source_impl::pluto_source_impl(fmcomms2_source::sptr block) :
        hier_block2("pluto_source",
                 io_signature::make(0, 0, 0),
                 io_signature::make(1, 1, sizeof(gr_complex))),
        fmcomms2_source_f32c(true, false, block)

As far as I understand, the variable type of the output here is gr_complex, which in Python corresponds to numpy.complex64. So, in that sense, I think my input variable type is correct, otherwise, I would probably get an error. But, I'd like to know if the output variable in pluto_source_impl is fixpoint and, in that case, how do I define my input variable in Python to match the fixpoint type?

Thanks in advance.

- Verónica

On Thu, Nov 25, 2021 at 7:49 PM Marcus Müller <marcus.mueller@ettus.com> wrote:

Hi Verónica,

in your long/short_sync_block's __init__, you set the in_sig to [np.complex64], which is
of a complex number composed of two 32 bit floats.

You can change that to other types!

But: your wifi_phy block outputs something specific, it needs to match that.

Best regards,

Marcus

On 25.11.21 17:47, Verónica Toro Betancur wrote:
> Hi Martin,
>
> Yes, that could definitely be the case. I don't have my radios right now with me, but
> I'll try it tomorrow.  And sorry for the silly question, but how should I define it in
> Python to be fixpoint?
>
> On Thu, Nov 25, 2021 at 6:25 PM Martin Braun <martin.braun@ettus.com> wrote:
>
>     Verónica,
>
>     have you maybe mismatched data types? Like, the real signals are fixpoint, but your
>     Python is doing floating point?
>
>     --M
>
>     On Thu, Nov 25, 2021 at 2:59 PM Verónica Toro Betancur <vetorobe@gmail.com> wrote:
>
>         Hi,
>
>         I am trying to detect and decode WiFi and ZigBee signals in GNURadio. For the
>         detection, I have implemented my own blocks in Python. It all works well with
>         simulated signals but the problem comes when I use radios to acquire real
>         signals. I'm using Pluto SDR and it works perfectly when I use it in workflow
>         examples but not in my own implementation. I mean, I plot the data that comes
>         directly from the radio and it looks good in the given examples but, in mine, it
>         looks like noise.
>
>         I am using the exact same parameters in both cases. The only difference I see is
>         that the blocks in the example are all in C++ while mine are in Python. Could
>         this be the problem? If so, is there a way to solve it other than writing the
>         blocks in C++?
>
>         Thanks in advance.
>
>
>         Best regards,
>         Verónica
>