Mmh, maybe it's related to the u8 output. Inside FormatConverter.cpp, can you try with +127 on line 68? Maybe that's off by one.

from odr-dabmod.

No, it's larger than before ...

from odr-dabmod.

BTW: s8 and s16 show both the same spike, but into the other direction

s8

s16be

Please mind the y axis. I don't know if this is an issue with welle-io or with odr-dabmod

P.S. The doc-file says

Also supported is s16, with system endianness (little endian on x86_64 and ARM)

but it looks like s16be, not s16le which looks this way in welle-io:

from odr-dabmod.

The plots for s8 and s16be look good, the central carrier is not supposed to have any energy.

For u8, it would be great it you could measure the average value of I and Q over a complete transmission frame. For example with numpy. You can easily load complex float data like that:

import numpy
frame = numpy.fromfile("your_file.iq.cf32", numpy.complex64)

Maybe it's not as easy with interleaved u8 I/Q, let me know if you want to do this and if you're stuck somewhere.

And then use numpy.average on the frame.

from odr-dabmod.

Do you fear it's related to my PC? If not I would like you (or someone else) to measure that.

from odr-dabmod.

No I don't think so. I would just prefer not doing it myself :-)

from odr-dabmod.

I don't think we're going to investigate this further, and since I believe it is more a consequence of the u8 representation rather than a real issue I'll close this ticket.

from odr-dabmod.