JT65 Software Comparsions
These test are based upon the lowest SNR before no decode with various types of conditions.
The three software in question is MultiPSK, WSJT, JT65-HF for the JT65a mode.

First test is a direct path at the lowest SNR before no decode..


WSJT
PatSim was at -25 noise level
-24 DB is what the software reported

MultiPSK
PathSim was at -23 db noise level
Software reports -24 DB

JT65-HF
Pathsim -25 db SNR
Software reports -23DB

Mid Lattitude Quiet .5 hz spread .5 mSec delay

MultiPSK
-21 db SNR noise level
software reported -23DB

WSJT
-27 db SNR noise level
software reported -25 db

JT65-HF
-27 db SNR
software reported -27db

High-Latitude Quiet
1mSec delay .5hz spread
MultiPSK
-19db SNR
software reports -23db

WSJT
-26 db SNR
software reports -28db

JT65-HF
-26 db SNR
software reports -28db

Mid-Latitude Disturbed
1hz spread 2mSec delay

JT65-HF
-25 DB SNR
software reports -24 db

WSJT
-25 db SNR
software reports -24 db

MultiPSK
-20 db snr
software reports -23db

Mid-Latitude Disturbed NVIS

MultiPSK
-21 DB SNR
software reports -24 db

JT65-HF
-25 db SNR
software reports -24db

WSJT
-25 db SNR
software report -24db

I am not real sure why adding some instability to a path actually helped out the sensitivity till the instability got severe.

As you noticed from the data above, MultiPSK seemed to score the lowest in every case.

As you also can see from the data above the software is not real accurate on the correct SNR reading , MultiPSK seemed to be the worse for this as well.