7397c3cd9b
Fix build warnings about ternary operator precedence
...
Be more explicit with parenthesis to avoid any possibly confusion
about operator precendence.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
2016-03-21 12:42:29 -07:00
f801ae3f28
update copyright dates 2016
2016-02-19 12:29:58 -08:00
639b695f31
eliminate all instances of implicit double-promotion
2015-08-23 10:16:15 -07:00
ff0346f5a8
Really use 64 data types for demodulation
...
[ kamal: cleanup ]
2014-11-16 12:46:14 -08:00
cc6b5c9123
Change code to use 64-bit variables for decoding
2014-11-16 12:46:13 -08:00
40f1d50cbd
fsk: calc separate divergences for mark and space bits
...
Calculate and consider separate signal-strength divergence values for
mark and space bits, so that imbalanced (but consistent) mark/space
strengths do not drive down confidence.
Note: This change makes all existing test cases yield very slightly
higher confidence values.
2014-11-16 11:00:36 -08:00
c2fa2dae93
fsk: fix FSK_DEBUG character decode
2012-11-03 13:26:10 -07:00
577b84d7f2
(portability) use stdlib.h not malloc.h
2012-10-04 18:10:33 -07:00
1ed3b8465a
fsk: confidence algo #6 : "SNR * (1-divergence)"
...
Disabled all the hardcoded noise limiters.
2012-08-24 22:38:41 -07:00
dd6040336d
minimodem: stats report ampl={amplitude}
2012-08-25 09:17:05 -07:00
501ee25c7d
fsk: expect_bits_string param, no start/stop bits required
...
fsk_find_frame() takes expect_bits_string (e.g. "10dddddddd1"), as a
parameter supplied by the caller allowing for encodings which don't use
any start/stop bits (e.g. just "dddddddd").
2012-08-24 13:20:20 -07:00
e9263c9c9d
fsk: fix noise <= EPSILON handling
...
Handle noise <= EPSILON better; fixes (im)perfect test cases on i386.
2012-08-19 15:58:08 -07:00
f81f069d95
minimodem: --limit (2.3) sets fsk_confidence_search_limit
...
Significant performance enhancements from fsk_confidence_search_limit:
runs far fewer FFT's when --limit (default 2.3) is less than INFINITY.
2012-08-18 21:24:01 -07:00
e1b055cf4f
fsk: AVOID_TRANSIENTS noise limiter re-enabled
...
Algo #5 needs this, just as #3 did.
2012-08-18 19:15:33 -07:00
076343a9c3
fsk: MIN_MAGNITUDE noise limiter
2012-08-18 19:15:07 -07:00
725355aa43
fsk: performance: do not bzero fftin
2012-08-18 16:52:07 -07:00
cf9ba20ee0
fsk: confidence algo #5 (SNR)
...
Computes SNR confidence value, where "signal" is the magnitude of the FSK
driven frequency, and "noise" is the magnitude of the non-driven frequency.
(Also includes unused algo #4 code, which may not compile anymore.)
2012-08-14 22:31:30 -07:00
bca34371e1
fsk: do not consider the stop bit twice
2011-07-08 13:00:50 -07:00
25e6e41939
fsk: no FSK_MIN_STRENGTH limitation
2011-07-08 09:26:02 -07:00
845c3e1be6
fsk: CONFIDENCE_ALGO 3 (worst divergence)
...
CONFIDENCE_ALGO 3 (worst divergence):
- single worst diverging bit in the frame
- amplitude normalized
- drop AVOID_TRANSIENT
2011-07-08 08:03:31 -07:00
cb3e81e576
fsk debug
2011-07-08 07:31:54 -07:00
fd0546d80d
fsk expect_bits
2011-07-07 13:31:10 -07:00
2d1310febf
reduce FSK_MIN_STRENGTH to 0.005
2011-06-23 23:05:35 -07:00
702c0be223
minimodem licenced as GPL3+
2011-06-22 17:10:50 -07:00
3a22116159
fsk: new confidence algo
2011-06-22 10:06:19 -07:00
2b014b50fb
fsk_find_frame: search across an entire bit
2011-06-21 22:30:37 -07:00
243af5da56
fsk: floating point band_width
2011-06-21 20:27:51 -07:00
9002cf2694
fsk: confidence considers start/stop bits
2011-06-21 12:13:47 -07:00
775ef649e1
fsk_bits_analyze / fsk_frame_analyze rework
...
- confidence calc only uses data bits metrics (for good or ill)
2011-06-12 12:42:08 -07:00
e7b16118e4
fsk: respect n_data_bits
2011-06-10 16:19:11 -07:00
439b606a97
minimodem 0.2.0
2011-06-08 14:00:22 -07:00
54f8efde20
minimodem: use fsk.h API
2011-06-08 13:21:12 -07:00
35121ab63c
fsk: move advance to top of loop
2011-06-06 21:36:23 -07:00
0509066f34
fsk cleanup
2011-06-06 21:33:50 -07:00
336a057233
fsk: carrier autodetect for 300 baud
2011-06-05 12:10:59 -07:00
2ea1085b45
fsk cleanup
2011-06-05 08:49:06 -07:00
cd6ab7f77d
fsk: noconfidence_bits increase to 20
...
... and probably should be set even higher, to allow for even longer
idle-bit (stop-bit) periods.
2011-06-05 08:44:47 -07:00
3522182394
fsk new confidence equation
2011-06-04 16:31:26 -07:00
1e96d42d16
fsk_bit_analyze refactor
2011-06-04 15:55:22 -07:00
0e888d2bf2
fsk cleanup
2011-06-04 14:41:41 -07:00
08654e8466
fsk: AVOID_TRANSIENTS allows for miniscule FSK_MIN_CONFIDENCE
2011-06-04 14:15:54 -07:00
71baa6f916
fsk cleanup
2011-06-04 11:56:26 -07:00
b73f7fb5b8
fsk: cleanup
2011-06-04 09:57:33 -07:00
80c8a62073
fsk: multibit decoder
2011-06-03 23:26:00 -07:00
Colin Ian King