Re: real to iq conversion



On Jun 18, 10:27=A0am, "alexander.tudor" <alexander.tu...@xxxxxxxxx>
wrote:
Greetings:

I have a radio that produces 14 bits at 56MS/s. The radio has no flow
control and I have to read samples into the PC's memory on one thread,
an=
d
write on a different one to disk.

Given today's PCs, 20GMIPS+, two 14 bit multiplications, and two NCOs
would not seem to be a problem. In addition, I would like to phase and
amplitude correct, which I presume would be a lot less intense than
the
real to iq operation above.

Is there a repository of open source for these types of computation?

Thank you,

Alex

Don't let those GMIPS numbers fool you, they apply to data that is
already sitting in fast memory. Forget the processing, the hard part
of your problem is IO. How is the data acquired? GIGE? PCI? What is
your disk configuration? What O/S?

John


Thanks to all who responded.

I use Linux. There is no flow control when reading from the ADC but I can
buffer to disk and feed the reals as the iq conversion is taking place. The
question is really, given X real samples how long does it take to convert
them to X IQ pairs.

Alex
.



Relevant Pages

  • Re: real to iq conversion
    ... I have a radio that produces 14 bits at 56MS/s. ... operations and you said your data stream has no flow control. ...
    (comp.dsp)
  • Re: real to iq conversion
    ... I have a radio that produces 14 bits at 56MS/s. ... control and I have to read samples into the PC's memory on one thread, ... already sitting in fast memory. ... your disk configuration? ...
    (comp.dsp)
  • Re: Memory allocated during run time
    ... Since the memory needed for a is unknown until run time, ... disappear when control flow exits the enclosing block in which it is ... you call a function control ...
    (comp.lang.c)
  • Re: real to iq conversion
    ... I have a radio that produces 14 bits at 56MS/s. ... The radio has no flow ... control and I have to read samples into the PC's memory on one thread, ...
    (comp.dsp)
  • Re: R> DROP in GForth
    ... It provides powerful flow control for slightly less overhead ... So maybe you could reverse the flow of arguments and first tell us ... flows remaining until you reach a structured exit. ... and wherever it starts getting snarled you can do a LOB like ...
    (comp.lang.forth)