mossmann/clock-recovery

clock recovery experiments

clock-recovery's Issues

wcpr crashes on Arch Linux

Traceback (most recent call last):
  File "./wpcr.py", line 94, in <module>
    symbols=wpcr(samples)
  File "./wpcr.py", line 43, in wpcr
    d = numpy.diff(b)**2
  File "/usr/lib/python2.7/site-packages/numpy/lib/function_base.py", line 1926, in diff
    return a[slice1]-a[slice2]
TypeError: numpy boolean subtract, the `-` operator, is deprecated, use the bitwise_xor, the `^` operator, or the logical_xor function instead.

Looks like there are issues within the following construct because d has not the right types for the subtraction in numpy.diff() (boolean instead of numerical).

b = a > midpoint(a)
d = numpy.diff(b)**2

Recovery without a preamble

First, thanks for this tool and the corresponding talk!
I guess I’m exactly the person you’re talking about, too little radio experience and completely thrown back by the scariness of the Gnuradio clock recovery blocks, which I guess wouldn’t work in my case anyways.

I’ve got this frequency modulated signal:

The base/modulation frequency seems to be 9.8kHz and the symbol rate 480Hz, thus the signal is clearly visible already in the FFT plot.
Your tool is usually doing a good job at recovering the symbol rate, but it’s also outputting gibberish data for the flatline part before the actual data starts.

Any advise on how to handle this?
I guess I could do some numpy magic to detect when the actual data starts, but I wonder if there’s a smarter way.

Thank you!