Re: Questions about how to observe BER using CPFSK modulation

[Christophe Seguinot]

Hi

• -0.3 correspond to 0.5=10^-0.3 error rate which is the maximum error rate obtained when demodulation is completely wrong.
  
• You must use a delay in your input bytes (between throttle and BER block) so that receives bytes are synchronized with emitted bytes.
  
• Since no synchronization is used here, even if bits are decoded without error, there is absolutely no warranty that unpacking one byte (give 8 bits), repacking 8 bits latter in the freceiver give the same byte (because whitout synchronization, the 8 repacked bits are no the same as the 8 unpacked bits) .

I would suggest:

• add a delay on emitted bits (not bytes)
  
• compare emitted and received bits (bit_stream QT time sink) to find the delay between them
• compute BER from these 2 bit sstreams
  

On 10/09/2020 13:02, Hiroki Sagara wrote:

Hello everyone!
I wanted to observe the error rate when using CPFSK modulation, so I created a flow graph as shown in the figure.
As a result, the output of the "BER block" did not change from -0.3 even if the noise increased or decreased.
(Voice communication using CVSD vocoder can be done without any problem, so I think the error rate is not so high ...)

How can I observe it correctly?

Thank you.