Commit graph

37 commits

Author SHA1 Message Date
ha7ilm
c7d9aa05e7 Fixed a lot of things related to fractional decimator 2017-02-28 00:27:41 +01:00
ha7ilm
875434bf96 added comment 2017-02-18 16:37:49 +01:00
ha7ilm
5f772ec3e3 made that spectrum even better 2017-02-18 16:36:36 +01:00
ha7ilm
1fab87af1d haha, fixed that artifact too. Strange flickering line in the spectrum at -100 dB was caused by reading some samples out of region 2017-02-18 16:26:51 +01:00
ha7ilm
0306b5b5bf fixed fractional_decimator_ff for N=6 interpolator, still interesting effect on the spectrum 2017-02-18 16:24:19 +01:00
ha7ilm
5f8f8695d3 fixed fractional_decimator_ff 2017-02-18 16:17:53 +01:00
ha7ilm
139d3d731b single samples bad, possibly at the beginning 2017-02-18 16:01:14 +01:00
ha7ilm
f4da40ffa7 fractional_decimator_ff gets new algorithm based on Lagrange interpolator formula 2017-02-18 15:49:04 +01:00
ha7ilm
3246e20864 Merged origin/master into nmux 2017-01-19 17:00:16 +01:00
ha7ilm
6386e3f4d1 merged master into nmux 2017-01-15 21:12:34 +01:00
Tatu Peltola
38d567d96e Fix logaveragepower_cf for FFT sizes below 16384 2016-10-22 22:16:47 +03:00
Tatu Peltola
8cbf028732 Merge branch 'master' of https://github.com/tejeez/csdr 2016-10-22 21:32:43 +03:00
ha7ilm
ef39d8dc27 Fixed 24 bit conversions 2016-06-23 10:50:08 +02:00
ha7ilm
de110ff719 Fixed 24 bit conversions 2016-06-21 00:23:43 +02:00
ha7ilm
c4b3527490 Added convert_s24_f and convert_f_s24 2016-06-21 00:14:03 +02:00
Tatu Peltola
d2a0099227 Precalculate FFT window
Before this calculating FFT window was taking much more CPU time than calculating the FFT itself.
2016-06-07 21:30:00 +03:00
Tatu Peltola
f9d6d22fe2 Implement averaged FFT 2016-06-07 21:30:00 +03:00
ha7ilm
42ce9d4669 Fixed handling a null input source for some functions, in order to work with the new squelch. 2016-03-20 23:30:21 +01:00
ha7ilm
4230198d91 Added a squelch 2016-03-20 16:41:37 +01:00
ha7ilm
277075b397 Renamed convert_f_i16 -> convert_f_s16 and convert_i16_f -> convert_s16_f. (csdr also triggers for the original functions, but they are deprecated. 2016-02-14 11:19:36 +01:00
ha7ilm
03a3a9ab50 Merged dev into master. 2016-02-14 11:08:53 +01:00
ha7ilm
4380bec6fa Tested all shifts, and they seem to work! 2015-11-30 00:20:22 +01:00
ha7ilm
95ebc0e790 Fixed shift_addfast_cc/NEON, now it works! Added shift_unroll_cc. 2015-11-29 22:46:06 +00:00
ha7ilm
965ea631fc Now at least we get a continuous output signal from shift_addfast_cc/NEON. And we get a good result if the shift_rate is 0! (Okay, we get a bad result for anything else...) 2015-11-29 20:47:00 +00:00
ha7ilm
d28edcca3d Fixed a jump. 2015-11-29 19:13:24 +00:00
ha7ilm
d24164db03 Added NEON implementation of shift_addfast_cc. 2015-11-29 19:05:28 +00:00
ha7ilm
4697928c38 Added shift_addfast_cc and tests. 2015-11-29 12:54:09 +00:00
Michael Ossmann
a3ad1e62d9 converters for signed chars, compatible with hackrf_transfer 2015-11-22 12:53:10 -07:00
ha7ilm
0e8ae4bdff Added fixed_amplitude_cc; fixed add_dcoffset_cc. 2015-11-01 11:30:03 +00:00
ha7ilm
fb5899f087 Added FM modulator. 2015-10-31 15:42:17 +00:00
ha7ilm
0800e4f431 Added AM modulator and test. 2015-10-31 15:22:33 +00:00
ha7ilm
c89f6e026b Of course can't null by OR instructions 2015-09-30 15:04:46 +00:00
ha7ilm
9cec3782e2 Added NEON optimization for DDC.
Buffer size can now automatically adjust to sampling rate changes between csdr processes.
2015-09-30 13:52:43 +00:00
ha7ilm
c50ffaac1b Added new features, e.g. the IMA ADPCM codec and sdr.js 2015-08-16 23:40:42 +02:00
ha7ilm
9e5ce0cc3b fix parameter naming mistake: rolloff -> transition_bw 2014-12-13 23:46:41 +01:00
ha7ilm
a671a43898 add a predefined filter for 44100sps to deemphasis_nfm_ff 2014-12-12 10:22:57 +01:00
ha7ilm
07c6a7df3a initial commit 2014-11-28 16:44:41 +01:00