dfd81b055b
begin work on class-based plugins
...
ladspa/crap_eq is working, seemingly.
2015-06-07 13:18:59 -07:00
79ffa24c08
consistency
2015-06-06 23:03:04 -07:00
e7ca27ca55
obsolete file
2015-06-06 22:59:50 -07:00
bbbfa7690c
move -ladspa/-vst suffixes to directory prefixes
2015-06-06 20:19:58 -07:00
9aa9fb8e3a
bump
2015-06-06 19:41:23 -07:00
5b82fcfac3
it works!
2015-06-06 19:31:38 -07:00
7bcf6c40d2
fix alignment
2015-06-06 16:50:19 -07:00
e2b74d736c
rework crap_tube to use templates
...
i'm so sorry
2015-06-06 16:45:09 -07:00
ad4551af16
use templates for process functions
...
there's some duplication across the biquad plugins
but this can be resolved later
2015-06-06 11:27:35 -07:00
39e385f1f9
beginning conversion to sepples
2015-06-06 10:42:14 -07:00
42c23a95d5
Update README.md
2015-06-04 09:58:15 -07:00
ec0fcf908c
moog filter
2015-06-03 09:35:22 -07:00
834e424087
fix sepples compilation
2015-05-28 21:14:13 -07:00
e72a42b6e0
unused code
2015-05-28 19:47:55 -07:00
99c4a214ce
working svf matricies
2015-05-28 19:28:55 -07:00
79fe905ec5
incomplete svf implementation
2015-05-28 16:45:14 -07:00
9f58751a04
shuffle around definitions
2015-05-28 12:21:38 -07:00
e2f8a22f95
don't need this
2015-05-17 16:07:57 -07:00
ffa6c66a10
take three
2015-05-08 14:57:19 -07:00
db5933978a
take two
2015-05-06 13:32:45 -07:00
9054360ded
first attempt at T420 speaker compensation
2015-05-06 11:10:20 -07:00
1fbefa9eab
happy belated new year
2015-04-11 09:24:29 -07:00
f1a8bcd876
use single-precision floats for ARM
...
NEON doesn't support doubles ;_;
anyway, this turns the two-times slowdown
caused by previous commits into a two-times speedup.
2015-04-07 13:55:12 -07:00
95c3e58ff9
split decimate, interpolate
2015-04-07 11:33:11 -07:00
33c0ef8f14
vectorize tube
2015-04-07 11:25:26 -07:00
8e7dde59f6
bump eq; factor out process funcs
2015-04-06 17:35:57 -07:00
4022d11349
use vectors instead of intrinsics
2015-04-06 12:10:37 -07:00
75fa193a90
optimize biquads
...
using process.h and running in-place
is 19% faster on this machine.
from there, using intrinsics yields another 94%,
for a total speedup of 130%.
2015-04-05 17:52:39 -07:00
db0cd0a7dd
move process() to its own file
2015-04-05 17:52:39 -07:00
4a4e4f20d0
personal EQ for Sennheiser HD 598
2015-04-05 17:52:01 -07:00
315265e2e8
Merge branch 'master' of github.com:notwa/crap
2015-04-04 10:20:16 -07:00
80e60fab1c
inline everything; fix types
2015-04-04 06:48:27 -07:00
a394384555
optimize by splitting into blocks
...
+67% faster on this machine
2015-04-04 06:31:57 -07:00
e6ab4e4473
handle LADSPA's port range crap
2015-04-04 02:44:33 -07:00
e943a99966
tweaks; use gnu11 now
2015-04-04 02:06:41 -07:00
3072985b0a
fix segfault: allocate all ports
2015-04-04 02:06:13 -07:00
1cd27b1c68
fix potential segfault
2014-06-19 08:27:25 -07:00
889ece8080
new polyphase IIR oversampling filter
2014-06-18 05:54:19 -07:00
46a4d07e58
move tube oversampling to its own file
2014-06-18 04:39:24 -07:00
eb2eec3709
refine make output (like linux and automake)
2014-03-13 01:24:32 -07:00
5bf857dee4
eq_const: new settings
2014-03-05 00:56:22 -08:00
e8cf7505e7
rudimentary ARM support
2014-02-24 16:18:58 -08:00
195e33d2a5
update readme
2014-02-24 13:32:48 -08:00
85774c0341
tube: another better oversampler (6x now)
2014-02-24 13:22:42 -08:00
c0ef645404
use exported objcopy if available
2014-02-24 11:56:09 -08:00
7545689477
tube: fix oversampling
...
the original cascade was missing a first-order filter entirely!
N is incremented so the cascade is only second-order filters
which allows minimal changes to the code.
additionally the output sample is from the last
zero-padding for a slightly better phase response.
2014-02-11 03:00:41 -08:00
1b1e4353fb
tube: better oversampling filter to reduce aliasing
2014-02-08 16:53:19 -08:00
1550230dc5
tube: add parameters with smoothing
2014-02-07 22:57:42 -08:00
d20bea5fa5
rewritten makefile, separate SSE2 test, readme blurb
2014-02-07 17:54:05 -08:00
21059bbf7e
make benchmark target work again, whitespace
2014-02-06 02:49:04 -08:00
