/[svn]/linuxsampler/trunk/ChangeLog
ViewVC logotype

Contents of /linuxsampler/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1644 - (show annotations) (download)
Sat Jan 19 16:55:03 2008 UTC (11 years, 10 months ago) by persson
File size: 20626 byte(s)
* fixed memory leaks that occurred when liblinuxsampler was unloaded
* fixed a memory leak that could happen when a channel was deleted
  while notes were playing
* fixed memory management bug in ASIO driver
* optimized the SynchronizedConfig class so it doesn't wait
  unnecessarily long after an update

1 Version CVS HEAD (?)
2
3 * packaging changes:
4 - autoconf bugfix: the PKG_CONFIG variable wasn't initialized properly,
5 causing e.g. the libgig test to fail when
6 "./configure --disable-jack-driver" was used
7 (patch by Alexis Ballier)
8
9 * general changes:
10 - bugfix: on some POSIX systems instrument editor plugins refused to
11 load as we used a non-portable Linux specific struct field
12 (fixes bug #70, patch by Ronald Baljeu)
13 - fixed endless loop which occured when loading an instrument editor
14 plugin DLL on Windows systems failed
15 - fixed memory leaks that occurred when liblinuxsampler was unloaded
16 - optimized the SynchronizedConfig class so it doesn't wait
17 unnecessarily long after an update
18
19 * audio driver:
20 - removed the nonsense audio channel constraint (which was hard coded to
21 max. 100 audio channels) for most audio drivers
22 - JACK audio driver did not offer a device parameter "SAMPLERATE" as
23 opposed to the LSCP specs
24 - bugfix: the SAMPLERATE parameter of some drivers (e.g. JACK)
25 reflected the wrong value
26 - fixed a memory management bug in ASIO driver
27
28 * instruments database:
29 - avoid time consuming samples scanning when adding instruments
30 to the instruments database
31
32 * Gigasampler format engine:
33 - fixed a memory leak that could happen when a channel was deleted
34 while notes were playing
35
36 Version 0.5.1 (6 December 2007)
37
38 * packaging changes:
39 - added autoconf checks for pthread library
40 - added autoconf check for pthread bug found on certain NPTL-enabled
41 glibc versions (see Gentoo bug report #194076)
42 - added autoconf checks for MS Windows
43
44 * general changes:
45 - bugfix: the thread used by an editor plugin didn't die when the
46 editor closed
47 - bugfix: Ringbuffer.h: fill_write_space_with_null() did not zero
48 out all the space. operator--() did not apply size_mask after
49 decrementing the read_ptr. DEFAULT_WRAP_ELEMENTS set to 0 in
50 order to avoid problems with the _NonVolatileReader functions.
51 - bugfix: Stream.h: added a missing divide by BytesPerSample in
52 GetWriteSpace(). Since this function is currently only used in
53 the stream's qsort() compare function, it didn't trigger any
54 bugs.
55 - Resampler.h, Synthesizer.h: cubic interpolation now works in
56 24bit mode too. Faster method to read 24bit words on little
57 endian machines (x86): replaced 3 byte reads + shifts with a 1
58 unaligned 32bit read and shift
59 - experimental support for MS Windows (MIDI input via MME, AUDIO
60 output via ASIO)
61 - made handling of SIGINT signal (Ctrl-C) a bit more robust
62 - support for monitoring the total number of active disk streams
63 (new LSCP commands: GET TOTAL_STREAM_COUNT,
64 SUBSCRIBE TOTAL_STREAM_COUNT, UNSUBSCRIBE TOTAL_STREAM_COUNT)
65
66 * AUDIO driver:
67 - added Windows ASIO low latency audio driver
68
69 * MIDI driver:
70 - added MME Windows MIDI driver
71
72 * LSCP server:
73 - added support for Windows style path / filenames, however with
74 forward slash path separators instead of backslash
75 (i.e. "C:/foo/bar.gig")
76 - allow naughty liblscp to send non-string device parameters within
77 apostrophes as well
78 - added new LSCP commands: "GET FILE INSTRUMENTS <file>",
79 "LIST FILE INSTRUMENTS <file>" and
80 "GET FILE INSTRUMENT INFO <file> <index>" for retrieving informations
81 about an arbitrary instrument file on the system where the sampler is
82 running on
83
84 Version 0.5.0 (15 October 2007)
85
86 * packaging changes:
87 - config.h is not going to be installed along with liblinuxsampler's
88 API header files anymore
89 - only the API relevant header (and source) files will be exported to
90 the Doxygen API documentation (explicitly listed in Doxyfile.in)
91 - added completely new XCode project files for Mac OSX which is now
92 capable to execute our autoconf environment, thus no need anymore
93 to maintain the compile time configuration file (osx/version.h) for
94 OSX manually (patch by Toshi Nagata)
95 - fixed buggy boolean --enable-foo / --disable-foo configure script
96 parameters
97 - global.h now only covers global definitions that are needed for the
98 C++ API header files, all implementation internal global definitions
99 are now in global_private.h
100 - atomic.h is not exposed to the C++ API anymore
101 - no need to include config.h anymore for using LS's API header files
102 - fixed warnings in API doc generation
103
104 * general changes:
105 - replaced the old, confusing MIDI program change mechanism by a
106 flexible MIDI instrument mapper which allows to map arbitrary
107 (MIDI bank, MIDI program) pairs with arbitrary
108 (engine type, instrument file, file index) triplets which will be
109 loaded on the respective channel when such MIDI program change
110 messages arrive, beside that, each entry allows to define a life-time
111 strategy for the instrument, so the user can define whether the
112 instrument should i.e. be permanently loaded (to be able to switch
113 quickly among sounds) or i.e. loaded just on demand when the
114 respective program change arrives (to safe RAM space), as well as a
115 global volume factor for each entry, so the user can adjust the volume
116 dynamic of his mapped instrument collection without having to modify
117 the instrument files, also one can manage arbitrary amount of such
118 MIDI instrument maps and assign each sampler channel individually
119 a certain map, so that i.e. separation of normal instruments and
120 drumkits is possible
121 - new notification events for tracking changes to audio/MIDI devices,
122 MIDI instrument maps, MIDI instruments, FX sends, global volume.
123 - sampler was limited to load max. 200 instruments in the background
124 due to a constant size RingBuffer FIFO which is now replaced by a
125 dynamic (unlimited) size std::list FIFO
126 - added FX sends, these allow to route audio signals to arbitrary audio
127 output channels for being processed by external effect processors
128 (i.e. jack-rack), the send levels are controllable via arbitrary MIDI
129 controllers
130 - global (sampler wide) volume can now be controlled at runtime
131 - Implemented new, improved notification system
132 - fixed compilation errors regarding OSX
133 (patch by Toshi Nagata)
134 - implemented instruments database
135 - added support for escape sequences to the instruments database
136 - added highly experimental support for on-the-fly instrument editing
137 within the sampler's process (by using instrument editor plugins),
138 you'll notice the new "Registered instrument editors:" message on
139 startup, added a new LSCP command:
140 "EDIT CHANNEL INSTRUMENT <sampler-channel>"
141 to spawn a matching instrument editor for the instrument on the
142 given sampler channel, the plugin path can be overridden at compile
143 time with ./configure --enable-plugin-dir=/some/dir
144 - added experimental code for synchronizing instrument editors hosted
145 in the sampler's process to safely edit instruments while playing
146 without a crash (hopefully) by either suspending single regions
147 wherever possible or if unavoidable whole engine(s)
148 - fixed several issues in fundamental "Thread" class: set scheduling
149 policy and priority on thread level, set a minimum stack size for
150 thread (TODO: a reasonable value yet to be tested), bugfix: non-RT
151 threads simply inherited properties of starting thread instead of
152 setting their own policy and priority
153 - minor fix in our automatic stack trace mechanism on crashes, the
154 main process did not wait for the stack trace process to finish
155 its output
156 - fixed some minor memory leaks
157 - reenabled assembly features support, at the moment only for
158 enabling a fast denormal FPU mode (x86 platforms supporting SSE2)
159 - minor assembly fix in x86 features detection (don't use the PIC
160 register, to avoid relocations in the text segment at runtime)
161 - POSIX callback functions of Thread.h are hidden
162
163 * MIDI driver:
164 - dispatch bank select (MSB and LSB) messages
165
166 * audio driver:
167 - the ALSA audio output driver parameters now reflect the correct
168 parameter value ranges for the respective selected sound card
169 (patch by Till Wimmer, a bit fixed and extended)
170
171 * Gigasampler format engine:
172 - if a filter is used and EG2 finishes before EG1, let the voice
173 die when EG2 has finished (this fixes a problem with clicks and
174 voice starvation for some gigs)
175 - playback is no longer disabled during instrument loading
176 - all notes playing on a channel that changes its instrument keep
177 playing with the old instrument until they get a note off
178 command
179 - EG fix: a release value of zero could cause noises or crash
180 - handle MIDI coarse tuning messages (MIDI RPN #0 MSB #2 LSB)
181 - EG fine tuning: when attack is zero the EG starts at a level
182 above max sustain level, which means that there is a decay phase
183 even if sustain is 100%
184 - more EG fixes: the level could sometimes go below zero and cause
185 noises or crashes
186 - minor fix of EGDecay (patch by Toshi Nagata)
187 - fixed compiler error when --enable-override-filter-type was
188 supplied to the configure script (fixes #46)
189 - disk thread: queue sizes are now proportional to CONFIG_MAX_STREAMS
190 instead of a fix value
191 - behavior fix: on MIDI CC# 65 (portamento on / off), 126 (mono mode),
192 127 (solo mode) only kill voices if the respective mode really
193 changed
194
195 * LSCP server:
196 - fixed compile time error for old Bison versions
197 (i.e. v1.28 found on MacOS 10.4, patch by Toshi Nagata)
198 - parser now supports extended ASCII character set
199 (up to ASCII code 255, i.e. includes now umlauts and accents)
200 - filename arguments in LSCP commands now allow to use escape
201 sequences, that is directly literal as one of: \', \", \\, \n, \r,
202 \f, \t, \v, or as octal ASCII code value like \132, or as hex ASCII
203 code value like \xf2) (fixes bug #24)
204 - the following LSCP commands now also support escape sequences for at
205 least one of their text-based parameters (i.e. name, description):
206 "ADD MIDI_INSTRUMENT_MAP", "MAP MIDI_INSTRUMENT",
207 "SET MIDI_INSTRUMENT_MAP NAME", "SET FX_SEND NAME", "CREATE FX_SEND",
208 "SET DB_INSTRUMENT_DIRECTORY NAME",
209 "SET DB_INSTRUMENT_DIRECTORY DESCRIPTION", "SET DB_INSTRUMENT NAME",
210 "SET DB_INSTRUMENT DESCRIPTION", "FIND DB_INSTRUMENTS",
211 "FIND DB_INSTRUMENT_DIRECTORIES"
212 - returns verbose syntax errors (line and column where syntax error
213 occured, the unexpected character and the actually expected, possible
214 character(s), the latter only if less than 5 possibilities)
215 - made sure that LSCP syntax is not affected by gigedit locale
216 settings
217 - bugfix regarding strings parameter lists: all comma separated lists
218 of strings were treated as being one string containing commas
219 (fixes #57)
220
221 * Bug fixes:
222 - fixed crash occurring on certain LSCP scripts (Bug 39)
223 - another thread safety fix for lscp "load engine" and "set
224 channel audio output device"
225 - fixed a crash which occurs when reassigning the same engine
226 on a sampler channel with connected MIDI device
227 - fixed a crash which occurs when changing the number of ports of a MIDI
228 device connected to a sampler channel to number less then or equal
229 to the index of the port to which the sampler channel is connected.
230 - The previous bindings were not been disconnected when altering
231 the ALSA_SEQ_BINDINGS parameter. Introduced a NONE keyword for
232 unsubscribing from all bindings (e.g. ALSA_SEQ_BINDINGS=NONE).
233 - The active stream/voice count statistic was incorrect.
234 - notification events were not been sent for some sampler
235 channel changes
236 - added default min and max values to restrict the number of allowed
237 audio output channels and MIDI input ports
238 - the connection to the PCM interface is now closed when destroying
239 an audio output device
240 - files with slash in their path or filename could not be loaded
241
242 * test cases:
243 - updated and fixed (haven't been touched in a while)
244
245 Version 0.4.0 (24 November 2006)
246
247 * packaging changes:
248 - changed deprecated copyright attribute to license; added ldconfig
249 to post-(un)install steps; added devel package for liblinuxsampler;
250 to linuxsampler.spec (RPM)
251 - install necessary development header files for allowing 3rd party
252 applications to link against liblinuxsampler
253 - liblinuxsampler's API documentation can be generated with 'make docs'
254 (Doxygen required)
255 - added benchmark to automatically detect the best triangle LFO
256 implementation (currently either an integer math solution or a
257 di-harmonic approximation), automatic detection can be overriden
258 with --enable-signed-triang-algo=x and --enable-unsigned-triang-algo=x
259 configure script argument though (mandatory for cross-compilation)
260 - do not automatically pick optimized gcc flags if the user already
261 provided some on his own (as CXXFLAGS)
262 - added compile time option to disable processing of All-Notes-Off MIDI
263 messages
264 - added compile time options to allow disabling the various audio and
265 MIDI drivers
266 - fixed automatic GCC CPU switch detection on PPC
267 (patch by Ebrahim Mayat)
268
269 * Gigasampler format engine:
270 - extensive synthesis optimization
271 (reimplementation of EGs and LFO(s), removed synthesis parameter
272 prerendering and the synthesis parameter matrix in general, splitting
273 each audio fragment into subfragments now where each subfragment uses
274 constant pitch and filter coefficients. The volume coefficient is
275 linearly interpolated inside a subfragment, unless
276 --disable-interpolate-volume is set.)
277 - fine tuning of the EG modulation parameters
278 - improved filter cutoff calculation by adding support for the
279 following gig parameters: Cutoff freq (used when no cutoff
280 controller is defined), Control invert, Minimum cutoff, Velocity
281 curve and Velocity range. The keyboard tracking now scales
282 cutoff frequency instead of resonance.
283 - added support for gig parameter Resonance.
284 - fixed bug in sysex handling (patch by Juan Linietsky)
285 - added global volume attenuation of -9 dB (0.35f) to prevent clipping
286 which can be overridden with --enable-global-attenuation
287 - EG fixes: made the length of "attack hold" stage more
288 accurate. Release stage can now start before attack stage
289 ends. Cancel release didn't work when sustain was zero. Attack
290 time now has a minimal value to prevent clicks.
291 - fixed pitch changes larger than one octave
292 - fixed EG3 (pitch envelope) synthesis which was neutral all the time
293 - implemented portamento mode and solo mode (a.k.a 'mono mode'):
294 all modes can be altered via standard GM messages, that is CC5 for
295 altering portamento time, CC65 for enabling / disabling portamento
296 mode, CC126 for enabling solo mode and CC127 for disabling solo mode
297 - fine tuning of the curves for volume (CC7), pan (CC10 and gig
298 parameter) and crossfade
299 - added support for the "attenuation controller threshold" gig
300 parameter
301 - added smoothing of volume changes caused by control change
302 messages
303 - sample loop parameters are now taken from the DimensionRegion
304 instead of the wave chunk
305 - fixed keyswitching for v3 gigs with a number of keyswitch splits
306 not equal to a power of two
307 - reimplementation of the filter algorithm. The new filter is more
308 accurate and supports all gig filter types, including bandreject
309 and lowpass turbo.
310 - real support for 24 bit samples - samples are not truncated to
311 16 bits anymore
312 - support for aftertouch (channel pressure, not polyphonic
313 aftertouch)
314 - LFO1 behaviour fixed (dampening from max volume instead of
315 amplifying from 0)
316
317 * LSCP server:
318 - fixed application exit on broken pipe error (fixes bug #20)
319 - fixed the notification messages delay due to lack of
320 network activity (fixes bug #26)
321 - fixed parser bug which occured on space(s) within device parameters
322 (e.g. "SET AUDIO_OUTPUT_CHANNEL_PARAMETER 0 0 SOMEPARAM='foo bar'")
323
324 * audio driver:
325 - added aRts audio output driver (by no means RT safe)
326
327 * MIDI driver:
328 - fixed legacy sysex code which caused dispatching of MIDI SysEx
329 messages several times instead of once
330 - API extension for MIDI drivers which already supply exact time stamps
331 for events (i.e. for offline rendering based MIDI drivers)
332 - added checks for bad MIDI data
333
334 * linuxsampler application:
335 - show available sampler engine types on startup
336
337 * general changes:
338 - support for muting sampler channels and solo mode of the same, two new
339 LSCP commands ("SET CHANNEL MUTE" and "SET CHANNEL SOLO") and two new
340 fields ("MUTE" and "SOLO") for command "GET CHANNEL INFO" were
341 introduced for this, the behavior is the same like on a mixer console
342 (patch by Grigor Iliev, a bit adjusted). Also added configure option
343 --enable-process-muted-channels which can be used to enable the
344 processing of muted channels.
345 - support for sostenuto pedal
346 - support for monitoring the total number of active voices
347 (new LSCP commands: GET TOTAL_VOICE_COUNT, GET TOTAL_VOICE_COUNT_MAX,
348 SUBSCRIBE TOTAL_VOICE_COUNT, UNSUBSCRIBE TOTAL_VOICE_COUNT)
349 - fixed some memory management errors
350 - fixed some concurrency problems that could lead to crashes when
351 LSCP commands were executed
352 - fixed crash when instrument loading failed previously on the same
353 sampler channel (fixes bug #36)
354
355 Version 0.3.3 (15 July 2005)
356
357 * packaging changes:
358 - fixed compilation with gcc 4.0
359
360 * Gigasampler format engine:
361 - LFO preprocessing was performed even though the respective LFO was
362 disabled by the instrument patch (minor efficiency fix)
363 - if period time of chosen audio device is too small (< MIN_RELEASE_TIME)
364 for volume ramp downs in the same fragment (needed for current voice
365 stealing implementation) then simply inform the user with a warning
366 about possible click sounds and reduce the volume ramp down
367 appropriately instead of cancelling the audio device connection
368
369 Version 0.3.2 (24 June 2005)
370
371 * packaging changes:
372 - updated autotools build files to compile on OS X
373 (conditional compilation of CoreMIDI and MidiShare drivers)
374 - hand-crafted assembly optimization code can be disabled with
375 './configure --disable-asm' (definitely not recommended)
376 - fixed 'make dist' rule to include all necessary files
377 - require automake (>= 1.5) for 'make -f Makefile.cvs'
378 (due to 'dist-bzip2' automake option)
379
380 * Gigasampler format engine:
381 - support for the gig parameters for "release velocity response" curves
382 - fine tuning of the EGADSR envelope
383 - volume of release triggered samples now depends on note-on velocity,
384 note length and gig parameter "release trigger decay" instead of
385 note-off velocity.
386 - revised voice stealing
387 (fixes crash and endless loop caused by voice stealing)
388 - don't reset scale tuning on instrument or audio output device change
389 - handle key group conflicts right at the beginning of each fragment
390 instead of when the respective voice is actually launched
391 (fixes undefined behavior if stolen voices belonged to a key group -
392 this case was followed by a "killed voice survived" error message)
393 - fixed minor issue with null/silence samples
394 (those stole voices even though they don't need a voice at all which
395 resulted in "voice stealing didn't work out" messages)
396 - don't reset volume, pan, pitch and MIDI controller values on
397 instrument or audio output device change
398
399 * LSCP server:
400 - fixed some crashes (patch by Grigor Iliev, fixes #19)
401 - fixed LSCP event "CHANNEL_INFO" notification
402 (e.g. did not notify on volume changes or MIDI program change events)
403
404 * linuxsampler application:
405 - added command line parameters --lscp-addr and --lscp-port to override
406 default IP address and port of LSCP server
407
408 Version 0.3.1 (24 May 2005)
409
410 * initial release

  ViewVC Help
Powered by ViewVC