Commit Graph

85 Commits

Author SHA1 Message Date
Jordan Sherer c8476b3308 Moved ctwk computation to syncjs8d 2020-05-02 15:24:20 -04:00
Jordan Sherer 2d6ac83606 Bumped max candidates to 300 (configurable) and cleaned up some sync code 2020-05-02 14:48:46 -04:00
Jordan Sherer a276a643bb Reorder nfqso reording after dedupe 2020-05-01 17:23:21 -04:00
Jordan Sherer 123cf9c5a0 Sync on ABC, AB, BC 2020-05-01 17:19:07 -04:00
Jordan Sherer c7354f1fc5 Ensure pass 3 does subtraction if we have a fourth pass 2020-05-01 10:31:01 -04:00
Jordan Sherer ff21c67e09 Fixed a problem in the synchronization code that caused issues with failed decodes on signals that should be able to be decoded 2020-05-01 10:16:12 -04:00
Jordan Sherer 59ce3b1118 Improving decoder performance 2020-04-30 22:48:20 -04:00
Jordan Sherer 1a10ab5d13 Working through decoder fixes, getting closer 2020-04-30 21:14:29 -04:00
Jordan Sherer 60f9a03898 Working through decoder fixes 2020-04-30 17:17:56 -04:00
Jordan Sherer 0e50639160 Merged https://bitbucket.org/hobbes1069/js8call/branch/gcc-10 gcc fortran fixes 2020-04-15 16:11:57 -04:00
Jordan Sherer 3875de4950 Added some decoder logging and js8 tone generation tweaks to ensure the right costas is getting generated for the right submode 2020-01-04 15:55:20 -05:00
Jordan Sherer ecf77ba04e Fixed legacy real array index warning. Fixed NCOSTAS warning 2020-01-02 22:39:17 -05:00
Jordan Sherer bf49b9b167 Added debug statements for costas 2019-12-13 19:59:28 -05:00
Jordan Sherer ee9ed66f82 Updated ft8 and js8 params 2019-12-13 11:09:31 -05:00
Jordan Sherer cc63277adb Decoder does not need to emit debug by default 2019-12-13 09:24:07 -05:00
Jordan Sherer c6407b2479 Fixed symlink for ft8 params 2019-12-13 08:48:03 -05:00
Jordan Sherer 5b9c71f380 Proper SNR calculations for each mode speed under additive gaussian
white noise
2019-12-12 20:56:16 -05:00
Jordan Sherer 77405cc968 Removing leftover ft8 files. Cleaning up sync code. 2019-12-12 20:35:16 -05:00
Jordan Sherer 7dc0298b18 I believe SNR should be the min between xbase and xnoi...let's try 2019-12-11 16:56:43 -05:00
Jordan Sherer 8e0e94e8bb Moved JS8E to 3.125 baud instead of 3 baud. 2019-12-10 22:24:50 -05:00
Jordan Sherer 5215fd324d Added I decoder and reorganized the E selection code 2019-12-09 14:00:23 -05:00
Jordan Sherer f706685cc9 Fixed signal subtraction for E speed signals 2019-12-08 14:14:58 -05:00
Jordan Sherer 3be107b348 Updated decoder with better debug handling. Fixed issue with E mode crashing on some decodes 2019-12-07 15:53:41 -05:00
Jordan Sherer eaff44aa38 Revert "Fixed array alignment in decoder"
This reverts commit cfeed7fc87.
2019-11-29 20:46:18 -05:00
Jordan Sherer cfeed7fc87 Fixed array alignment in decoder 2019-11-28 21:56:38 -05:00
Jordan Sherer f1ffc1de9d Fixed the decoder deadlock caused by waiting for the lockfile after a decode, which resulted in a deadlock on back-to-back decodes when a CPU was slow to decode TURBO within its 6 second cycle. Most of the time this was highly unlikely, but every now and then the timing was just right that the lock was created and removed before the decoder could observe that it was created. Sad panda. 2019-11-20 00:46:48 -05:00
Jordan Sherer 1a03619a2f Move decoder process handling to separate thread to avoid exhaustuion 2019-11-19 10:39:27 -05:00
Jordan Sherer 2a320ddcdd Added debugging fromthe decoder 2019-11-19 10:13:10 -05:00
Jordan Sherer 16da1235a0 Clean up decoder fortran 2019-11-17 02:14:20 -05:00
Jordan Sherer d0d7aaa286 Tweaking decoder, ensuring negative values are omitted 2019-11-12 09:34:54 -05:00
Jordan Sherer 084dd77b42 Added better baseline valid frame thresholds to prevent fewer false decodes 2019-11-07 22:18:05 -05:00
Jordan Sherer c001303bba Fixed windows crash for E decoder 2019-11-07 14:44:36 -05:00
Jordan Sherer cc6e28d5a7 Enable E decoder for windows testing 2019-11-07 14:26:08 -05:00
Jordan Sherer 0d785888d8 Broken E decoder, so we'll disable it for now. Also...don't try to decode if the decoder is curently busy 2019-11-07 11:07:01 -05:00
Jordan Sherer 0569bb0167 Optimized E settings 2019-11-06 20:33:37 -05:00
Jordan Sherer f7d253e7e3 Fixed lower limit for snr 2019-11-06 19:39:46 -05:00
Jordan Sherer 31137082f9 Tweak decoding for E 2019-11-06 16:32:13 -05:00
Jordan Sherer a43c913a19 Clean up js8d to jd8e to make it clear the difference 2019-11-06 16:00:16 -05:00
Jordan Sherer e5f07e3a0d Fix waterfall display by extracting global symspec vars and resetting the ja to the correct value after detector reset 2019-11-05 21:27:41 -05:00
Jordan Sherer 720bd6b22a Fixed issue with duplicate decodes based on the ring buffer after tx cycles completed when it lined up with a decode cycle. 2019-11-05 15:43:34 -05:00
Jordan Sherer bd357ed4c4 Working multidecoder using relative positioning in the frame buffer 2019-11-04 14:38:00 -05:00
Jordan Sherer 862e827ddf Refactored decode functions to be more independent and easier to grok 2019-11-03 15:29:59 -05:00
Jordan Sherer a24e4429ea Working decoder refactor using decode cycles based on buffer size instead of symbol half step nonsense 2019-11-01 15:07:24 -04:00
Jordan Sherer dc6d161c6f Set defaults back to standard decode procedure 2019-10-30 11:49:10 -04:00
Jordan Sherer 0ca7b0e5ed Updated ultra to 4, reorgnazed fixStop, and testing new decode patterns 2019-10-29 09:55:32 -04:00
Jordan Sherer 952d9d7080 Simplifying to help clear up some confusion 2019-10-26 23:01:05 -04:00
Jordan Sherer 56d457a056 Initial experimentation 2019-10-25 14:15:45 -04:00
Jordan Sherer 3841e1901f Ripped out echograph as it was causing strange linking issues with the MacOS compiler 2019-10-22 22:08:20 -04:00
Jordan Sherer 066cdb8423 Dedupe in the decoder b4 move fqso to top of list 2019-10-22 09:37:31 -04:00
Jordan Sherer 5313706818 Refactor decoder output to include which submode is decoded so we could eventually run three 2019-10-21 14:01:56 -04:00