Commit graph

1590 commits

Author SHA1 Message Date
derselbst
f2cf1f82c7 move fluid_rvoice_buffers_mix() to rvoice_mixer 2018-04-25 14:57:40 +02:00
derselbst
907ec27a9e optimize away memset/mix if voice is quiet 2018-04-25 14:56:20 +02:00
derselbst
291075c31f fix build for gcc 2018-04-25 12:33:47 +02:00
derselbst
044c5e3238 fuse preset iteration tests to preset+sample loading tests 2018-04-24 21:32:49 +02:00
Marcus Weseloh
df3bab0b40 Only import instruments once from Soundfont
This changes the way instruments and their related information is imported
from a Soundfont. Previously, each preset zone got a complete copy of the
instrument and related information, wasting a lot of precious memory.

After this change, an instrument and related information is only
imported once and then linked into all preset zones that use this
instrument.
2018-04-22 19:09:23 +02:00
Marcus Weseloh
c4003bef39 Store mutable information about inst zones in new fluid_voice_zone_t struct
This change separates the static instrument zone information read from
the Soundfont from the information that gets modified on import and
later via the legato handling. It opens opens up the possibility of
having unique instruments that only get imported once and then linked
into the individual preset zones.
2018-04-22 19:08:09 +02:00
Marcus Weseloh
a95a4864fd
Merge pull request #366 from FluidSynth/dynamic-sample-loading
Dynamic sample loading
2018-04-22 19:06:53 +02:00
Marcus Weseloh
3603a34a77 Also load 46 zero words area after sample for uncompressed samples 2018-04-22 12:20:37 +02:00
Marcus Weseloh
3ce3575269 Ensure that loops are sanitized and sample is optimized for dynamically loaded samples 2018-04-22 10:37:14 +02:00
Marcus Weseloh
4cfdacd905 Fix typo in dynamic-sample-loading settings documentation 2018-04-21 23:38:12 +02:00
Marcus Weseloh
7240461e5b Merge branch 'master' into dynamic-sample-loading 2018-04-21 23:34:20 +02:00
Marcus Weseloh
9dcb4148cb Only open the Soundfont for sample reading if necessary 2018-04-21 23:22:45 +02:00
derselbst
c433fc887c simplify dependency handling of unit tests 2018-04-21 23:22:45 +02:00
derselbst
02de830043 fix dangling pointer on inst_zone sample lookup
fixes #369
2018-04-21 23:17:19 +02:00
derselbst
323b9b8cbf avoid assigning incompatible pointers 2018-04-21 22:31:18 +02:00
derselbst
17dded3700 avoid NULL deref in fluid_rampreset_preset_delete() 2018-04-21 16:55:08 +02:00
derselbst
71592ec679 fix memory leaks during sfont loading 2018-04-21 16:48:56 +02:00
derselbst
08147cc3d6 avoid double free in fluid_sample_set_sound_data() 2018-04-21 16:33:27 +02:00
derselbst
a051d57ab3 avoid preprocessor stringification to expand to special '%' character
in fprintf() call
2018-04-21 10:37:04 +02:00
Marcus Weseloh
7987446fdf Fix tests without libsndfile. Not pretty, welcome a better solution! 2018-04-18 09:14:55 +02:00
Marcus Weseloh
073da2141c Implement dynamic sample loading
This change adds a new feature that enables loading and unloading of
sample data on demand. As soon as a preset is selected for a channel, all
of it's samples are loaded into memory. When a preset is unselected, all
of it's samples are unloaded from memory (unless they are being used by
another selected preset).

This feature is disabled by default and can be switched on with the
"synth.dynamic-sample-loading" setting.
2018-04-18 09:14:55 +02:00
Marcus Weseloh
a985c68a13 Enable decompressed Ogg Vorbis samples to be stored in sample cache
This change moves the Ogg Vorbis decompression to fluid_sffile, so that
this is the only place where we have to deal with compressed audio.

It also changes the way the samples are loaded for SF3 files: previously,
the compressed data was copied into memory, then the individual samples
were decompressed (resulting in both compressed and decompressed data to
stay in memory). Also, decompressed data wasn't cached, so previous loads
of the same file ran the decompressed again for each sample.

After this change, the vorbis decompression is changed so that it reads the
compressed data directly from the Soundfont file. And the resulting WAV
data is stored in the sample cache.
2018-04-18 09:14:55 +02:00
Marcus Weseloh
f52bbf53a4 Add VintageDreamsWaves-v2 in SF3 format and some tests for sf3 loading 2018-04-18 09:14:55 +02:00
Marcus Weseloh
3d05360f33 Split preset parsing from Soundfont file open 2018-04-18 09:14:55 +02:00
derselbst
8051b43c02 fix result checking for fwrite()
Potentially fixes infinite loop in file renderer, when fwrite() returns zero. Addressing #367.
2018-04-17 16:43:52 +02:00
Tom M
d3e6781abd
Merge pull request #364 from FluidSynth/197
Unify calling conventions for rvoice update functions
2018-04-11 19:42:19 +02:00
derselbst
ff19788fda avoid memory leak in new_fluid_voice() 2018-04-11 19:33:40 +02:00
derselbst
78de299da4 make param array of fluid_rvoice_function_t const 2018-04-11 19:22:55 +02:00
derselbst
88606f1538 test_sfont_loading: update unit test purpose 2018-04-11 11:12:01 +02:00
derselbst
36881f4ba8 add test for sample rate changes 2018-04-11 11:12:01 +02:00
derselbst
cb35fac1c6 make VintageDreams sf2 an explicit requirement for unit tests 2018-04-11 11:03:12 +02:00
derselbst
713a4b680d remove redundant TEST_ASSERT_[N]EQ macros 2018-04-11 10:23:17 +02:00
Tom M
f4f439180d
Merge pull request #365 from FluidSynth/preset-stack-removal
Preset stack removal
2018-04-11 10:14:32 +02:00
derselbst
b76aeead01 remove obsolete alloc preset comments 2018-04-11 10:05:24 +02:00
derselbst
0af950a1e2 avoid swapping rvoices when can_access_rvoice is false 2018-04-11 09:56:42 +02:00
derselbst
036595f7db avoid duplicate call of fluid_rvoice_set_output_rate() 2018-04-11 09:56:03 +02:00
Marcus Weseloh
839f62f89d Remove unused variable 2018-04-10 22:09:33 +02:00
Marcus Weseloh
fdd6804ae8 Prevent uninitialized use of prev_preset and assume VintageDreams soundfont for test 2018-04-10 22:08:09 +02:00
derselbst
0ccb0460fb update rvoice function related docs 2018-04-09 16:57:34 +02:00
derselbst
5034e0ca8b rename fluid_rvoice_eventhandler_push*() 2018-04-09 16:40:25 +02:00
derselbst
b4456d1b4c rename EVENT_PARAMS to MAX_EVENT_PARAMS 2018-04-09 16:32:32 +02:00
Marcus Weseloh
efa97718fa Increase soundfont references only when assigning preset to channel
Previously, the refcount was increased when retrieving a preset from
a soundfont. That doesn't really make sense anymore.
2018-04-08 23:03:14 +02:00
Marcus Weseloh
c6bdc4bc12 Add test for sfont preset iteration 2018-04-08 23:03:14 +02:00
Marcus Weseloh
5a8880560d Remove now invalid comments about having to free presets 2018-04-08 23:03:14 +02:00
Marcus Weseloh
1a1bf7d0ea Remove now obsolete preset stack code and struct members 2018-04-08 23:03:14 +02:00
Marcus Weseloh
66610abcb9 Refactor fluid_preset_t handling in defsfont and ramsfont
Removes the need for a pre-allocated stack of fluid_preset_t's. Upon
adding a loader specific preset, a fluid_preset_t is automatically
created. defsfont and ramsfont now keep track of the fluid_preset_t's,
not the loader specific ones.

Also changes the sfont::iteration_next signature to directly return
the next preset. This is possible as presets are now only created once
and returned as a pointer.
2018-04-08 22:59:55 +02:00
derselbst
0bfa332fd9 remove redundant call to fluid_synth_set_sample_rate()
on synth creation
2018-04-08 18:49:18 +02:00
derselbst
ddd14cb00b duplicate rvoice_eventhandler in fluid_voice_t
avoiding several indirection to access it
2018-04-08 18:35:54 +02:00
derselbst
7c2e76f2f3 avoid NULL deref during voice creation
by executing rvoice event function directly instead of enqueueing events
2018-04-08 18:12:53 +02:00
derselbst
9c44aaba62 remove fluid_rvoice_eventhandler_push5() 2018-04-08 17:35:31 +02:00