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

Contents of /linuxsampler/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1534 - (show annotations) (download)
Sun Dec 2 19:07:17 2007 UTC (16 years, 3 months ago) by persson
File size: 19007 byte(s)
* made signal handling a bit more robust

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

  ViewVC Help
Powered by ViewVC