[Fwd: LF: PSK31 tests]

[Andre Kesteloot]

Peter Martinez wrote:

> Re Charles' comments on the PSK31 tests.
>
> The test done by Andy and I did show that BPSK was indeed better
> than QPSK at the lower SNR, but we did not measure the SNR
> itself, so I couldn't say whether the cross-over occurs at 4dB or at
> some other level.
>
> The point I was trying to make was that, unlike on HF where the
> noise is far from random and QPSK is always best, the fact that
> BPSK was better on this tests implies that the noise was random.
>
> The constraint length is 5 in the error-correcting code used. I
> wouldn't expect a longer code to give much better performance in
> random noise. Longer codes are better for QSB and I don't think we
> are going to get much of that on this band. The real reason for not
> using a longer code is that real-time QSOs become unpleasant
> because of the time-delay. Although BPSK seemed better on this
> test, it would be interesting to repeat the comparison in the summer
> when the noise is more speckly than smooth: the error-correction
> might start to show some benefit.