On Sunday 14 Aug 2005 12:54, Johannes Stezenbach wrote: > On Sat, Aug 13, 2005 Jeremy Hall wrote: > > Is it possible that FE_READ_BER and FE_READ_UNCORRECTABLE might be > > swapped for stv0299? > > > > I'm looking at a transport and the uncorrectable blocks are much higher > > than the BER blocks. Is this expected behavior? > > By definition BER is an error rate (i.e. number of wrong bits > per 10^x received bits, with x typically = 12..18), which > UNC is an absolute count of errored TS packets since > the last call to FE_READ_UNCORRECTABLE. > > However, the stv0299 cannot measure both, so what Andrew > implemented is rather crude: switch the error mode and > block 100msec and return an absolute count in both cases. > > Which means that the number of wrong bits per 100msec > before error correction should be higher than the > number of errored packets per 100msec after error correction. > > Maybe Andrew can comment on this, I would guess that switching > the error mode doesn't reset the error counter, and a dummy > read needs to be added. Yeah, it should really block for a calculated time depending on the error accumulation settings. I don't have the time to do this properly right now. I can revert the change if you want.
