Wim Taymans
8513150bdf
resample: use calloc to make sure fields are 0
2020-05-08 16:13:34 +02:00
Wim Taymans
79fd4c941f
debug: Don't debug EBUSY params
...
When a device returned EBUSY when enumerating params, don't try to
enumerate (and fail) them again.
2020-05-07 17:15:46 +02:00
Wim Taymans
713f242853
audioadapter: handle -ENOENT from enum_param
2020-05-03 11:45:06 +02:00
Wim Taymans
aafd1e7298
improve debug
...
Improve log so that debug level 3 gives a reasonably readable overview
of what is going on.
2020-04-22 12:47:18 +02:00
Wim Taymans
e73431d541
resample: use right define to compile sse functions
...
See #220
2020-04-14 20:29:27 +02:00
Wim Taymans
c446dfb1d6
resample: improve debug
2020-04-12 20:46:59 +02:00
Wim Taymans
2bb2e50eef
audioadapter: stop when the follower needs data
2020-04-12 20:44:56 +02:00
Wim Taymans
c1f3593ef0
audioadapter: small cleanup
2020-04-12 20:44:19 +02:00
Wim Taymans
b18dacde9a
spa: improve draining
...
Make a new DRAINED status.
Place the DRAINED status on an input IO when a stream is out of
buffers and draining.
All nodes that don't have HAVE_DATA on the input io need to copy
it to the output io and return the status. This makes sure the
DRAINED is forwarded and nodes return DRAINED from _process()
DRAINED on the resampler flushes out the last queued samples and then
forwards the DRAINED in the next iteration.
Emit a new drained signal from the context when a node returns
DRAINED. Use this to trigger the drained signal in the stream.
2020-04-07 17:58:43 +02:00
Wim Taymans
fdb5911e86
resample: use a 0 sized buffer to drain
2020-04-06 18:01:38 +02:00
Wim Taymans
7febf7ea25
improve debug
2020-04-06 18:01:19 +02:00
Wim Taymans
95906ca52e
audioconvert: fix compilation
2020-04-04 20:03:11 +02:00
Wim Taymans
6937ec5e63
audioconvert: compile c version with right flags
2020-04-03 18:12:04 +02:00
Wim Taymans
28b0da626e
Add missing file
2020-04-03 18:05:29 +02:00
Wim Taymans
0d1cef6b3a
audioconvert: move some things around
...
To make it easier to add other implementations later.
Improve selection of resampler function
2020-04-03 18:03:42 +02:00
Wim Taymans
bf3ebb67aa
neon: add alignment hints
2020-04-03 14:26:02 +02:00
Wim Taymans
167460a9bc
fmt-ops: add arm32 optimizations
2020-04-03 14:09:46 +02:00
Wim Taymans
a81e269a7e
neon: fix neon detection and compilation
2020-04-03 13:30:08 +02:00
Wim Taymans
df5721ff47
neon: fix taps arguments
2020-04-03 10:31:55 +02:00
Wim Taymans
a87337cae8
neon: fix fallback
2020-04-03 10:22:58 +02:00
Wim Taymans
22e590c7cc
stream: somewhat implement drain better
...
Abuse the xrun callback in the adapter to emit the drained signal until
almost all data left the resampler. This needs more work with a proper
signal and a buffer flag to signal the drain.
2020-04-02 17:46:50 +02:00
Wim Taymans
1b3aaba206
fmt-ops: set the right flags on optimized functions
...
Or else we could select an avx2 version on a non-capable CPU.
Fixes #227
2020-04-02 17:34:47 +02:00
Wim Taymans
b87c0a9270
clean up options
...
make sure we can use -v for verbose mode
2020-04-02 14:34:02 +02:00
Wim Taymans
1fd5be7101
neon: fix resample interpolation
2020-04-01 13:07:52 +02:00
Wim Taymans
9b24a84ce6
fmt-ops: fix copy/paste error
2020-03-30 11:02:13 -04:00
Wim Taymans
527dc89365
fmt-ops: add neon optimizations for format conversion
2020-03-30 10:59:21 -04:00
Wim Taymans
67a5dd35be
resample: add neon optimizations
2020-03-27 08:27:11 -04:00
Wim Taymans
397787ec3e
spa-resample: getopt_long returns an int
2020-03-26 18:10:11 +01:00
Wim Taymans
a880012e72
fmtconvert: order formats better
...
Place float and higher resolution formats first so that they are
selected first.
2020-03-24 11:47:41 +01:00
Wim Taymans
b8a1ea1d3a
fmt-ops: fix 32 bit compilation
...
_mm256_extract_epi64 is only for 64 bits, add workaround for 32
bits.
Fixes #220
2020-03-23 14:18:10 +01:00
Wim Taymans
acccccd2c0
Use SPA_IO_BUFFERS_INIT when we can
...
We should not have to initialize the state to NEED_DATA, anything
that is not HAVE_DATA is good. Also we need to set the buffer to
INVALID because else it might be recycled.
2020-03-19 13:09:21 +01:00
Wim Taymans
07f935a5ed
add some more debug
2020-03-19 13:00:35 +01:00
Wim Taymans
8a2af908a7
small optimizations
...
Add some SPA_LIKELY
Enable FASTPATH by default
2020-03-18 15:43:29 +01:00
Wim Taymans
57f84ae5ae
fmt-ops: use gather to read samples
2020-03-18 11:41:14 +01:00
Wim Taymans
55633ebf9a
fmt-ops: move AVX
2020-03-18 10:06:54 +01:00
Wim Taymans
949dba7bfc
fmt-ops: flesh out avx optimizations
2020-03-17 17:27:47 +01:00
Wim Taymans
4dcd0dacc9
fmt-ops: add more benchmark cases
2020-03-16 16:45:20 +01:00
Wim Taymans
37d3864949
merger: add some SPA_LIKELY
2020-03-16 16:12:34 +01:00
Wim Taymans
3a911dfe3b
fmt-ops: add avx2 optimized version
...
Only one optimized version but the sse2 version are compiled with
the avx2 flags so that they get optimized better.
2020-03-16 16:11:29 +01:00
Wim Taymans
6eca935e61
also test optimized versions
2020-03-16 13:08:33 +01:00
Wim Taymans
fd3dd14ebe
fmt-ops: add optimized f32 to s16 stereo conversion
2020-03-16 13:04:21 +01:00
Wim Taymans
0548fbf690
fmt-ops: fix bug in s32 to f32 conversion
2020-03-16 12:57:21 +01:00
Wim Taymans
cb7bfdf98a
sprinkly SPA_LIKELY/UNLIKELY around
2020-03-16 12:52:28 +01:00
Wim Taymans
c4b2be2aad
use true and false for boolean properties
...
So that it's easier to convert them to booleans in introspection
code later.
2020-03-03 18:17:46 +01:00
Wim Taymans
ee617682c3
audioconvert: implement Portconfig query
2020-02-21 14:59:38 +01:00
Wim Taymans
d8bac82e72
improve logging
...
Remove some harmless warnings from the log
Remove some excessive info logging
2020-02-21 11:09:34 +01:00
Wim Taymans
a528189d26
slave -> follower
...
We use master/follower terminology everywhere.
2020-02-21 10:47:32 +01:00
George Kiagiadakis
0bff0e345f
audioconvert: disable monitor port logic if both sides are in dsp mode
...
There is a use case where you may want to use audioconvert with
both input & output sides configured in dsp mode, for altering
the stream volume. When it is configured like that, the
IS_MONITOR_PORT macro wrongly thinks that all output ports except
the first one are monitor ports, originating in the merger instead
of the splitter.
This change fixes that and it also completely disables exposing
monitor ports, even if they are enabled, when both sides are in
dsp mode (fmt_input_port_info() also uses the same macro to check
if the monitor ports are to be exposed)
2020-02-20 12:34:42 +00:00
Wim Taymans
75e14ecadc
resample: tweak resample qualities a little
2020-02-14 20:23:15 +01:00
Wim Taymans
621f6ca10b
resample: reset props first, then override
2020-02-14 20:22:55 +01:00
