morse-wip/ChangeLog at master · ag1le/morse-wip · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
(C) 2013,2014  AG1LE Mauri Niininen
2014-JUL-19
	git tag -a v0.1.04 -m "Release v0.1.04"
	Removed all static variables to enable running multiple instances - used in FLDIGI experimental release.

2014-JUN-22
	Removed most static variables to reduce size of .bss segment in order to be able to run in Raspberry Pi.

2014-JUN-14
	Added TCL/sqlite3 based testing framework to capture and process CER vs SNR test data.
	See test/cer_vs_snr_test.tc  and test/sqlite.db

2014-JUN-13
	Added GIT version tagging feature   bmorse -ver  returns the tag+version number
	Use git tag -a v0.1.02 -m "Release v0.1.02"  to update tag  -
	see http://stackoverflow.com/questions/1704907/how-can-i-get-my-c-code-to-automatically-print-out-its-git-version-hash


2014-MAY-25
	Fixed bugs and syncronized with FLDIGI 3.21.77 Bayesian v3  version

2014-JAN-22
	Fixed bug in savep.cxx - check if psum is zero before trying to divide by it.
	I think is is the root cause of bug mentioned in Sep-26-2013 - Pmax growing to huge values


2014-JAN-02
	New release bmorse-0.1.01
	Added FFT filtering code from FLDIGI cw.cxx and fftflt.cxx
	Converted all files from C to C++.
	Found a workaround for P(max) bug - see Sept 26. Adjusting dur[] initialization value from 1000.0 to 9000.0 in proces.cxx worked in most test files so far.


2013-DEC-29
	Added more options to enable debugging and testing bmorse decoder while using WAV audio files as source.
	Bit filter (-bfv) option does envelope smoothing - 10 msec rise time assumption based on experiences with FLDIGI.
	Plot option (-plt) enables piping such as  "bmorse -plt <sndfile> | xplot" to visualize the CW envelope.
	AGC  option (-agc) has fast rise and slow decay in case original signal has very large dynamic range.
	AMP option  (-amp) allows to multiply envelope signal with a constant.
	DUR option  (-dur) allows manually to set Bayesian decoder sample duration (used in path.c).
	FFT option  (-fft) enables either FFT based signal recovery or just low pass filtering.
		- FFT has also peak_detection() to find if there are multiple signals in the audio frequency range.
		- Parameter "delta" sets peak detection threshold
		- Width (-wid) and length (-len) parameters can be used to set FFT size and chunking


2013-DEC-28
	Renamed this program as "bmorse" - it stands for Bayesian Morse code decoder.
	Built new functionality and options for testing both sound files and text files.


2013-SEP-26
	enabled decoding from FLDIGI data feed. clamping x value to 1.0 max as FLDIGI sends values over 20.0 during startup before AGC kicks in.
	BUG: P(dah) abruptly goes from near 1.0 down and the bounces back => P(max) jumps to huge values
	too long dahs?

2013-SEP-25
	BUG: missing word space /pause between words QUICK BROWN  when < 30 db SNR ?
	FOUND: enabled  noise.c processing in morse.c
	changed to:
			noise_(&x, &rn, &zout);
			retstat = proces_(&zout, &rn, &xhat, &px, &elmhat, &spdhat, &imax, &pmax);
	much better decoding with low SNR test signals.

2013-SEP-25
	run "./morse t test/test20db.in | less"
	BUG: QUICD  and FOB  when high 20 dB SNR?
	D should be K  and  B should be X
	for some reason last 'dah' following word space gets decoded as 'dit'.
	FOUND:  Initl.c - line 123    1, 1, 0, 0, 0, 0,   // mauri 2013-09-25 bugfix
	had 0 instead of 1 in state k=4


2013-SEP-02	Morse decoding works on C++ version. Added decoding struct TREE in transl.c
			and logic to translate incoming morse symbols. This is marked as version v01.


2013-SEP-01  	Initial version. Original Fortran sources compiled with
			gfortran -g  *.f
		produces a.out  executable program.

		Based on initial testing the element state estimation works, but translating letters
		has still problems. Produces a sequence of letter states but translation to actual
		characters produces incorrect letters.

 		Compilation of C-sources produced by f2c is done with following commands:

			gcc -c *.c
			gcc  *.o -lf2c -lm

		produces a.out executable program. The output of both versions of a.out is in
			output_c.txt
			output_f.txt

		These have minor differences - source still unknown.