[Discuss-gnuradio] False Tags Correlation Estimator

Hello everyone,

I'm seeing double tags being placed occasionally by the correlation estimator block around a true correlation peak. This is causing me issues in my receiver. I have come up with a hypothesis that could use confirmation by more experienced developers.

I have attached a picture with my question on it to make it most clear, it really needed a picture. Please take a look at that before going on. If the scheduler happened to hand the correlation estimator a chunk of data such that the peak occurs at the end of the chunk (noutput_item = 4095 -> peaks occurs on item index 4095), then will the implementation maintain a record of this and not tag the first item in the next chunk, since it too is above the threshold?

The above scenario is handled when the peak is not placed at the end of the chunk, but it is unclear to me that it is handled for the above mentioned scenario.

Thanks,

Rich