[Date Prev][Date Next][Thread Prev][Thread Next] - [Date Index][Thread Index][Author Index]

Contest on algo



Dear friends,
just a short message.
I ran test on Thomas 's soft using calibrated noisy files generated with
the soft available on amsat-na download.
For the first test the parameters are shown here :
ENTRER DEMI AMPLITUDE (0 - 127) 28
AMPLITUDE=-13.2dB
ENTRER NIVEAU DE BRUIT  (0- 100) (128 si niveau FP=0) 20
DENSITE SPECTRALE=-49.1dB/Hz
S/N DANS B.P. 100 Hz= 15.9dB **** S/N DANS B.P. DE 0 A 4000Hz= -0.1dB

The file was comprising 4 frames of ascii text in order to perform
visual test on BER since at that low level of Eb/N0, you cannot expect
to have a correct CRC.

PSKDEM succesfully demodulated the four frames albeit with errors (the
measured BER was in the range of 7e-3/1e-2)
Thomas 's algo totally missed 3 frames out of 4.

I ran another test with higher Eb/N0 with the parameters shown :
ENTRER DEMI AMPLITUDE (0 - 127) 28
AMPLITUDE=-13.2dB
ENTRER NIVEAU DE BRUIT  (0- 100) (128 si niveau FP=0) 14
DENSITE SPECTRALE=-52.2dB/Hz
S/N DANS B.P. 100 Hz= 19.0dB **** S/N DANS B.P. DE 0 A 4000Hz=  3.0dB

PSKDEM succesfully decoded the 4 frames with an OK crc
Thomas 's algo missed 1 frame out of 4. (loss of char sync)

So , altough it would be hazardous to extrapolate the results, it can be
said (tentatively) that his algo is 3.1dB worse than Amsat-F its one.

I used the unlimited version of pskdec, should anybody want to perform
the tests, that version will be soon released after decison of amsat-f's
board.

Thomas 's soft is a beautiful piece of soft anyway and runs faster than
pskdem. (but you get for what paid for)

Tests were not performed on Nico's demod.

Best 73'.

-- 
--------------
Ghislain RUY
F1HDD
AMSAT FRANCE
--------------
----
Via the amsat-bb mailing list at AMSAT.ORG courtesy of AMSAT-NA.
To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org



AMSAT Top AMSAT Home