2
=======================================
4
=======================================
5
* added: missing headerfile in dssihost.cpp, required for building on Ubuntu Maverick (rj)
7
* Fixed: Mixers not restoring their size properly upon song load. (T356)
8
- Rearranged MusE::loadProjectFile1() slightly so mixer resizing done at end. Tested OK.
9
* Changed: Disabled some optimized code always enabled in Dsp::cpy(), for now, in case of 64-bit issues. (T356)
11
* Created rc2 release for 1.1 (rj)
12
* Fixed: Delete track(s) not disconnecting track routes, causing routes to hang around and store in song file. (T356)
13
- Fixed major regressions (my fault) in Song::insertTrack2() and ::removeTrack2().
14
* Fixed: Dummy audio device: Crashes when Jack midi device created in device manager, or loaded from song. (T356)
15
- This allows to create Jack midi devices even when dummy audio driver is used.
16
- Tested OK loading complex Jack midi based songs with dummy audio driver.
17
* Fixed: AlsaTimer cannot start, with dummy audio device. (T356)
18
- Reverted: Back to default SND_TIMER_GLOBAL_SYSTEM in AlsaTimer, for now until a better fix.
19
- Also replaced pop-up warning in dummy loop with terminal output instead.
20
* Changed: MidiDevice <-> Track routes replaced with MidiPort <-> Track routes. (T356)
21
- This allows to change a port's device (even to <none> !) while keeping all the track channel routes.
22
* Changed: Single midi <-> track route with channel mask, instead of one route-per-channel. (T356)
23
- Potentially saving very many routes from memory and in song file.
25
* Fixed: regression with oR routing for Aux strips (rj)
27
* Fixed: shortcut for insert moved to Ctrl+Shift+X (rj)
28
* Added: Select all toggle for midi channels iR menu (rj)
29
* Added: Auto-route-connect old songs with obsolete midi input port/channel masks (before iR button added). (T356)
30
- Added MidiTrack::setInPortAndChannelMask() and called it from MidiTrack::read().
32
* Changed: Updated German localizations muse_de.qm and muse_de.ts from <babarosa [at] gmx.at> (T356)
33
* Added: Instrument definitions and drum maps for Kurzweil SP2X, from <babarosa [at] gmx.at> (T356)
35
* Created rc1 release for 1.1 (rj)
36
* Added: Instrument definitions for AKAI SG01 and Waldorf-microQ from <babarosa [at] gmx.at> (rj)
38
* Added: Possibility to turn on/off move-armed-track with track selection in settings (rj)
40
* Renamed: Soft synth configuration changed to Synth configuration, as per patch from Geoff King (rj)
41
* Fixed: Shortcuts with modifier keys wasn't working for global shortcuts, should be now (rj)
42
* Added: Shortcuts for changing len of events, default Ctrl+Shift+Left/Right (rj)
44
* Fixed/Changed: Grid reacts to midi resolution change (rj)
46
* Added: Move track selection when part selection is moved, is this a good approach? (rj)
47
* Added: Backspace unarms all armed tracks (rj)
49
* Feature: Audio and midi routing popup menus now stay open, for making rapid connections. (T356)
50
* Fixed: Delete some objects hanging around upon close like song, audio, midiSeq, and prefetch. (T356)
51
TODO: Find a way to delete the 'muse' object without causing problems like seg faults.
53
* Improved: Note lanes in pianoroll and marking of C-notes (rj)
55
* Fixed: illegal shortcut no longer reported on startup (rj)
57
* Fixed: Handle errors during timer initialization more gracefully (rj)
59
* Added: Volume control for audio metronome (rj)
61
* Fixed: When wave files were edited allocation was on stack which caused crashes for larger selections, now done on heap (rj)
62
* Added: Print svn version in -v and About box. (rj)
63
* Fixed: Bug with RemoveShortEvents plugin script, would not remove anything due to index error (rj)
65
* Added: Insert measure, inserts empty space and moves everything after playpointer to the right and
66
Part-insert, pastes parts and moves everything after the paste point to the right
67
it's currently not possible to change the shortcuts for these, wip (rj)
69
* Added: Button in transport to enable/disable jack transport. (rj)
70
* Fixed: Dropdown Snap and Type in arranger toolbar no longer steal focus, annoyed the helloutame (rj)
71
* Fixed: Making edits to larger chunks in wave files no longer cause a crash, heap is used for allocation (rj)
72
* Improved: Part selection when navigating with keyboard, previously there was sometimes a need to use the mouse
73
to get a part selected, now the first one should be selected (rj)
74
* Added: some automation (very)wip stuff, not enabled, just to get it off my harddrive (rj)
76
* Fixed: Crashes loading a song while another is loaded. Or worse, it loads but saving corrupts the file! (T356)
77
- Song::clear(): Clear all midi port devices. Delete Jack midi devices, and remove all ALSA midi device routes.
78
- Route::read() Ignore bogus midi routes in med file if the device they refer to is not in use by the song (port is -1).
79
- MidiJackDevice::writeRouting(), MidiAlsaDevice::writeRouting(): Ignore and do not write midi routes if the device
80
they refer to is not in use by the song (port is -1). This prevents bogus midi routes writing to med file.
81
- Fifo::~Fifo(): Free the buffers! Fifo::getWriteBuffer(), Fifo::put(): Verify allocation. Fifo::get(): Verify buffers.
83
* Fixed: Jack midi output: Sent pitch bend and program values were incorrect, if coming from midi keyboard. (T356)
84
- Reported by Pieter while using Hurdy Gurdy vst under fst. Tests OK now, here.
85
* Fixed: Drum editor: Controller graph not in current tool mode (pencil, eraser etc.), when first opened. (T356)
86
- Call setTool() in DrumEdit::addCtrl().
87
* Fixing: Drum editor: Velocity controller graph showing incorrect for selected drum, or showing for all drums, when first opened. (T356)
88
- WIP. Still not quite correct, but at least now it doesn't show all drum velocities when first opened.
89
(This is interesting! I will try to allow 'no drum' list selection, to show all drum velocities at once, just like piano roll.)
91
* Fixed: More fixes to marker list, selected item was still not right. (T356)
93
* Fixed: Time signature editing problems in master track list and gui. (T356)
94
- Fixed SigEdit::outOfRange(). Added Sig::isValid(), and check it in LMaster::returnPressed() to fix crash with 0 n or z.
95
* Fixed: Marker/list problems. Double entries. Pianoroll/arranger timescale 'shift-left/right-click-to-add/del-marker' fixed.
96
Marker 'lock' (smpte/ticks) works now. Up/down buttons fixed. Marker list selection 'stays put' now
97
when adding or deleting markers. (T356)
98
* Fixed: Graphics corruption in arranger and track list when vertically scrolling. (T356)
99
- Changed rectangle to full w/h in 'shift down' section of View::setYPos() and TList::setYPos().
100
Should not have to do this, but it cured my problems. No noticable change in speed.
101
- Arranger corruption occured on more than one machine with xorg nv, ati drivers. More severe arranger AND track list
102
corruption occured with proprietary drivers. Corruption is also observed in several other windows (ex. ladspa browser)
103
but there's not much I can do, the corruption is found in ordinary usage of QListView for example.
104
* Changed: Increased arranger vertical scroll step, was too slow to scroll. (T356)
105
* Possible fix: Auto-scroll in pianoroll and arranger take waaay too long to stop scrolling. (T356)
106
- Increased timer timeout from 40 to 80 ms in Canvas::scrollTimerDone(). Helped for me, on a 1.6Ghz P4...
107
* Changed: Ladspa plugin guis now have scroll bars. Helps when there are many controls. (T356)
108
- Still toying with the code a bit, please bear with me while I try to get it right.
109
* Started: Addition of MusE-native ladspa guis for dssi synths. (T356)
110
- No visible changes yet, just under the hood. Added class PluginIBase which is inherited by
111
DssiSynthIF and PluginI. Changed class PluginGui and DssiSynthIF to fit.
113
* Fixed: Xml file trouble: Tag attributes not storing xml-friendly entities, trouble reloading song. (T356)
114
- Fixed all usages of Xml::nput, ::put, ::tag, and ::etag. Discovered by Geoff B.
116
* Fixed: Audio Input tracks: Loud noises or DC output with unconnected input routes. (T356)
117
- AudioInput::getData(): Do not get buffers of unconnected client ports. Causes repeating leftover data.
118
* Fixed: Audio Input tracks: Stereo processing not correct if both input routes connected to same port. (T356)
119
- AudioInput::getData(): Just copy the buffers always, don't set buffer pointer directly.
120
* Fixed: Effect rack plugins: Not saving correct number of plugin channels. Upon reload, channels are wrong. (T356)
121
- PluginI::writeConfiguration() and PluginI::readConfiguration(): Write and read the channels, not instances.
122
- Optimize TODO: Is xml plugin 'channel' really necessary? Well not now, but later if we optimize the effect rack
123
by not creating redundant plugin instances if the next plugin in the rack doesn't need them.
124
What I have changed is always make the number of plugin channels equal to the number of track channels.
125
PluginI::setChannels() and PluginI::initPluginInstance() then sort out the details of how many instances to create.
126
* Fixed: Effect rack plugins: Some plugins crash MusE, for example plugins with no audio outputs. (T356)
127
- PluginI::setChannels() and PluginI::initPluginInstance(): Be more aware of different audio ins and outs.
128
* Added: First draft help file README.effects-rack titled "Understanding the Effects Rack". (T356)
129
- Details how MusE uses plugins, especially now with these changes.
131
* Changed: Building: Separate --enable-dssi --enable-osc configure options. (T356)
132
--enable-dssi (Enable dssi synths AND special handling of dssi-vst ladspa effect plugins. No guis if by itself.) and
133
--enable-osc (Enable dssi guis. Useless by itself at the moment).
134
Use both for best features.
135
* Feature: If dssi and osc support are enabled, audio track effects rack dssi-vst plugins can now show their NATIVE guis. (T356)
136
* Fixed: Storing dssi synth states. Remembers current program, all configure strings, and control settings - for current program ONLY. (T356)
137
- With dssi-vst synths, some better are than others at remembering program.
138
- Todo: Ignore unwanted dssi-vst config strings/values like "guiVisible" "<name-of-fifo-file>" (gives non-fatal error on reload).
139
* Fixed: Dssi native synths (hexter, LTS etc.) should actually work now. Their gui controls operate the sounds now. (T356)
140
- Todo: If using large audio buffers (Jack), my 'one-at-a-time-per-audio-process' osc control fifo streams kinda suck,
141
with a delay in their processing. Want to try MULTIPLE 'run's or 'run-multiple's PER audio process, for SAMPLE accuracy and
142
better quicker message processing. Just need to add a timestamp to my osc fifo items, then all forms of change - whether
143
control, midi event, or program - have a timestamp with which we can SPLIT UP the runs into 'runs between changes'.
144
* Tip: Currently a patch to dssi-vst is required for their guis to update properly. (For now, ask about this on our lists.)
146
* Fixed: Shortcut for moving between tracks in arranger. Ctrl+Up/Down (rj)
148
* Patch: Fix for libdir bug causing synths to not be available on some 64bit systems. By Orcan Ogetbil (rj)
149
* Fixed: Drawing of grid when snaping to less than measure (rj)
151
* Fixed: LADSPA rack effects: Do not display LADSPA output controls (such as latency). (T356)
152
* Fixed: Bug when cancelling 'save as' operation, says 'file exists'. Caused by my earlier compressed save fixes. (T356)
154
* Fixed: DSSI requested midi mapped controls were not correct. (T356)
155
* Changed: Working on controllers, and saving their state. Synths like Organ should remember their state now. (T356)
156
- Working on DSSI controls... Working on OSC stuff...
158
* Fixed: Jack midi, and DSSI: Midi controller knobs, sliders, boxes etc. not updating with current values. (T356)
159
- Copied code in MidiPort::sendEvent() to JackMidi::processMidi() and DssiSynthIF::getData().
160
* Fixed: DSSI: Crashing with unkown controllers sent to synths. (T356)
161
- Ignore unkown controllers in DssiSynthIF::processEvent().
163
* Added: Rec enabled track moved with selection when only one track is rec enabled (rj)
164
* Changed: Made canvas show grid the default (rj)
165
* Added: Jack midi ports can now be renamed in the ports list dialog, by clicking on the name. (T356)
166
- Also cleaned up ports list behaviour and added tooltips.
168
* Major reworks: Jack midi, routing system, multichannel synth ins/outs, midi strips and trackinfo pane. (T356)
169
- WORK IN PROGRESS. Should be usable for current projects.
170
- ADVISORY: If opening current projects you are advised to "save as" or back up your med files (waves should be OK)
171
until all is deemed to be working OK. It is possible more changes to .med file structure are needed later.
172
- Some .med project file structures have CHANGED (like routing). I have tested opening large existing projects,
173
then re-saving them. OK so far.
174
However, I have not added importing code for the old midi trackinfo panel 'input port' and 'input channel' boxes (gone now).
175
If you want to wait, I am planning to do that, but later... Otherwise you must re-enter them with the new midi input routing.
176
- ** Multichannel synth inputs and outputs: Fairly satisfied with popup menus, hopefully no major reworks...
177
Routing system fully supports multichannel synth mono/stero paths but is NOT COMPLETE yet.
178
When changing a connected track from stereo to mono, there is no coding yet to ensure the existing stereo routes are
179
changed from stereo to mono (just like the audio input and output tracks have always done with Jack routes).
180
Also coding must be added to avoid simultaneous mono and stereo routes - make them mutually exclusive - although
181
there's nothing technically wrong with it, it's just not desirable.
182
- ** Jack midi: You now create your own Jack midi ports. I have not added any means to delete them yet, but I believe
183
if you 'unselect' them in the midi ports list so that they are not used, then the saved .med file will ignore them.
184
- Multi-instances of MusE should work OK now - all ports should hopefully have unique names.
185
- Number of midi ports increased from 32 to 128. It's a bit much for the ports list, was planning "create your own
186
midi port as needed" (like RG). Obstacles seem to have been cleared now, but it would be a big job (lots of places to change).
187
- Along the way various tweaks and fixes when I spotted them.
189
* Added: Separate Jack midi client ports per MusE port. (T356)
190
- For now, you will have to manually add further connections to those ports if desired, each session.
191
- Tested OK rec/play so far.
192
* Fixed: Some inconsistencies in midi controller number decoding. (T356)
193
- Everything now references the values listed in midictrl.h
194
* Fixed: Some Jack code was not calling free after calling certain functions. (T356)
195
* Changed: Midi priority now deaults to audio priority +1. (T356)
196
- Thanks to RJ for the convincing argument.
198
* Added: Jack midi input should be working now. Controllers, too. (T356)
199
- Tested with large 1024 frame buffer. Events are recorded with good sub-tick (frame) resolution,
200
as viewed in pianoroll with 64T snap grid.
202
* Fixed: Jack midi output should now be frame-accurate, support variable length events (like sysex),
203
and most controllers should work. (T356)
204
- TODO: No record yet. Just playback.
205
- TODO: I have code in place for multiple device listing (like ALSA) with auto-connect,
206
instead of single 'jackmidi' device, but it doesn't seem to want to actually connect.
207
Simply enable the #define JACK_MIDI_SHOW_MULTIPLE_DEVICES in driver/jackmidi.h
208
and it will magically switch. Shows port names as their alias #1 (with a rem'able line to
209
revert to non-alias names - to be made configurable later). I swear it was working,
210
then intermittent, then not at all. Hmm...
212
* Fixed: MusE hanging on close, with Jack driver. (T356)
213
- Unregister Jack midi ports on destruction of JackAudioDevice.
215
* Fixed: Top level windows: Do not open if there are no parts. Case: bogus -1 top-level part index stored in med file. (T356)
216
- Changed MusE::readToplevels() and stuff in listedit.cpp
217
- Was causing crash of list editor upon loading a song saved with the list editor open.
218
Somehow a bogus -1 top-level part index can make it in to the .med file. Checking this...
219
* Changed: Preparations for audio processors: Moved stuff out of class Event and into class WavePart. (T356)
220
- WavePart needs a converter map to lookup events in the part's (shared) event list, in case clones exist !
221
Ughh, it's going to be tough to ensure the map is always correct because an event cannot tell what part owns it...
223
* Improved: Grid handling in arranger improved, it should now be actually usable (rj)
225
* Fixed: Jack midi output: Stuck or missing notes. (T356)
226
- Must only be one jack_midi_event_reserve() per event in handle_jack_midi_out_events().
228
* Changed: Midi sync window: Clock is now seperated from other real time commands (play/stop etc). (T356)
230
* Fixed: Midi sync in: Should be very solid and 'in time' now, play/stop/continue. (T356)
231
- Re-coded to begin incrementing immediately upon first-clock detection. Forbid MusE to send transport commands
232
while in external sync mode - our sync router handles that. Was causing problems.
233
- Tested OK: Ensoniq ESQ-1 KB (sends clocks ony between start and stop) and Roland TR-505 drums (clocks always running).
234
The measured general timings were quite different, but (thankfully) clocks and events were not.
235
- MusE responds better to the TR-505 than the ESQ-1 does with a direct midi cable connection! (ESQ gains ticks on continue).
236
* Added: Midi sync window: 'Rewind on start' column. (Also a 'Send first clock delay' box - not implemented yet). (T356)
238
* Feature: Dummy audio driver: Added global settings for sample rate and period size. (T356)
239
- The period size affects midi resolution, and the default was just too long. So this was added.
241
* Changed: Don't process audio prefetch, or getData() in Audio::copyData(), if track is 'Off'. (T356)
242
* Added: Arranger track list: Quick 'right-click' or 'ctrl-click' or 'ctrl-mouse-wheel' toggling of Track On/Off. (T356)
243
- Note this includes midi tracks now! Remains to be seen whether allowing midi off is useful and will work,
244
or should be filtered out. **TODO: Turn off remaining controls in midi strips, turn off actual midi playback and record.
245
* Feature: Initial Jack midi support. Imported from Larry Valkama's GIT repo. (T356)
247
* Fixed: Import midi 'replace' broken last fixes. (T356)
248
* Fixed: External midi sync: Wait until first clock after start to start transport, and >= second clock to increment ticks. (T356)
250
* Added: Midi sync: Added 'Jack timebase master' checkbox, removed requirement of MusE being tempo map master. (T356)
251
* Added: While on external sync and Jack timebase master, reports correct MusE BBT info to Jack. (Tempo not correct yet). (T356)
252
* Added: Midi import option: Whether to split imported tracks into multiple parts. (T356)
254
* Fixed: External midi sync in: MusE transport not rewinding upon reception of start. (T356)
255
* Added: Midi sync: MusE now transmits and receives some MMC commands, and displays some MTC and SMTPE info. (T356)
256
- No transmit MTC or receive MTC syncing yet, just transport control stuff.
257
- Tested OK with Rosegarden and Ardour, but they do not seem to have an option to sync to the input midi clock,
258
only MTC or internally. Must test when MTC output added to MusE.
259
* Improved: Midi sync editor window: Shows MMC and MTC activity and works with them now. (T356)
261
* Fixed: First tests: External midi sync in works now! Should be rock solid no matter what is thrown at it. (T356)
262
- All changes are labeled "p3.3.25"
263
- Switched to a 'direct drive' method instead of 'calculation of tempo' method.
264
Drift between source of master and slave is non-existant since the incoming midi clocks
265
drive the midi engine directly. This is at the expense of exact 64th note resolution.
266
At MusE's fixed 24 midi clocks per quarternote, triplet 32nd notes are resolved (32nd note divided by three).
267
MusE's default 'division' (ticks per quarternote) is 384. 384/24 gives a division of 16 MusE 'ticks' per midi clock.
268
- Normally when using the tempo map (external sync is off), device play event times are in frames,
269
because the tempo map provides for that kind of sub-tick frame resolution.
270
But now, during external sync, device play event times are in ticks, not frames.
271
Hence the compromise between exactness of note times, and master/slave drift.
272
- Tested heavily with a Roland TR-505 which has a variable 'analog' tempo knob.
273
Combined that with a complete midi song, and other test 'tick' tracks driving external synth KB,
274
was rock solid going 'absolutely nuts' with the tempo knob up to 260BPM, surprisingly good resolution.
275
- TODO: Check safety of 'midiExtSyncTicks = 0' inside audio thread while midi thread is counting it up.
276
Also fix seek operations and check for any more place needing these 'tick'frame' alterations.
277
* Changed: Some more changes for compressed files operations. (T356)
278
- More to be done, but tests OK so far.
280
* Fixed: Saving/loading compressed .gz/.bz2 MusE .med files, + loading compressed .mid/.kar files (save broken, off for now). (T356)
281
- Changed filedialog.cpp:getSaveFileName() and MusE::loadProjectFile1.
282
- Added seperate save file dialog filters in globals.cpp.
283
- Disabled exporting of compressed midi/karaoke files (.mid/.kar) for now because
284
a compressed file is opened as a pipe, and pipes can't seek, resulting in a
285
corrupted midi file in MidiFile::writeTrack().
287
* Added: Piano roll and drum editor edit menus: "Select prev/next part". With default hotkeys Alt+Left/Right. (T356)
288
- This is an easy way to switch parts when multiple part editing in one window (via Ctrl-E or Ctrl-D).
289
And it was absolutely essential, for multiple dedicated controller track editing since there are no notes,
290
and no (other) way to select another part to edit.
292
* Feature: MusE now has two mixers, with selectable track type display. (T356)
293
- Changes to .med file were needed. Tested OK so far with old and new song files.
294
* Fixed: Builtin freeverb and TAP reverb ladspa GUIs, or custom user GUIs not showing controls. (T356)
295
- Broken since removal of musewidgetsplugin library months ago.
296
Added PluginWidgetFactory class to handle local creation of controls.
297
(Custom user GUIs must be done by hand since the removal of musewidgetsplugin library).
298
* Fixed: Appearance settings red colour adjustment larger than others. (T356)
299
* Changed: Default controller graph foreground colour from blue to medium orange. (T356)
300
- Avoid clash with blue colour used by left and right markers.
2
302
* Release: 1.0.1 (rj)
added
removed
ChangeLog
