iain duncan wrote:
>>I think that 'groove quantise' does have quite a specific meaning :
>>> To take the timing from one midi part, and apply it to another.
>>> So, you get the midi timing from a real drummer grooving away and
>>apply
>>> it to your beat. Then it pulls your beats to the nearest beat in the
>>> groove, rather than to the normal 16th or whatever.
>
>
>That was one sequencer's implementation of the term, but certainly not >a
>universally accepted strict definition.
I guess so. It's only Cubase that calls it that. Emagic logic, Fruity
Loops, Pro Tools, Groove Slicer, Cakewalk Sonar and Digital Performer
all call using the timing from one part on another a 'groove template',
which is much clearer.
If the 'groove' was pulling the notes towards a dotted or triplet feel,
I'd call it 'shuffle' or 'swing' to differentiate it.
'groove quantise' should perhaps be used when a sequencer or drum
machine is applying a groove more complex than a simple triplet feel,
but not using a template based on another recording.
>Iain
HI all,
There was a discussion on the LAU list about jamin using FFT
based filtering. I I missed much of the discussion but that
particular point just jumped out at me.
Has anyone thought of trying linear phase FIR filters instead of
FFT methods? Any filter that can be specified in the frequency
domain can be implemented in the time domain and vice-versa.
Often (but not always), the time domain version is significantly
cheaper in CPU cycles.
Erik
--
+-----------------------------------------------------------+
Erik de Castro Lopo nospam(a)mega-nerd.com (Yes it's valid)
+-----------------------------------------------------------+
"Neither noise nor information is predictable."
-- Ray Kurzweil
Hi all,
Does anyone know of a page somewhere that explains just what (on a
developer level) MIDI "groove" is? I want to implement it in a
sequencer, but all I can find is user documentation pages with useless
information.
Is it as simple as each note having a time offset (ie snare is early .5
ms, hi-hate late 1ms, etc.) or something more?
Thanks..
-DR-
Hi!
I want to programme my mic and my speakers on my FC2 box. I am running
KDE. Is there a library which I should look at or is the ALSA API the way
to go?
Thanks,
Catalin
--
<<<< ================================== >>>>
<< We are what we repeatedly do. >>
<< Excellence, therefore, is not an act >>
<< but a habit. >>
<<<< ================================== >>>>
Dave Robillard wrote:
>Hi all,
>
>Does anyone know of a page somewhere that explains just what (on a
>developer level) MIDI "groove" is? I want to implement it in a
>sequencer, but all I can find is user documentation pages with useless
>information.
>
>Is it as simple as each note having a time offset (ie snare is early .5
>ms, hi-hate late 1ms, etc.) or something more?
>
>Thanks..
I think that 'groove quantise' does have quite a specific meaning :
To take the timing from one midi part, and apply it to another.
So, you get the midi timing from a real drummer grooving away and apply
it to your beat. Then it pulls your beats to the nearest beat in the
groove, rather than to the normal 16th or whatever.
I seem to remember having some floppys of grooves back in the atari days
and doing this with midi shaker patterns and the like, but it's been a
long time....
Hans Fugal:
>
> ALSA. If you want to tie yourself to KDE and endless misery, try artsd,
> but I warned you.
>
Hmmm, I think the alsa api is a bit huge/complicated. I would never
reccomend doing alsa directly, and I think it was a very bad advice
actually. Check out portaudio, sndlib or jack instead, which provides
easier interface to the soundcard than alsa, and works on top of alsa (and
others).
--
Hello,
I trying my first steps with Supercollider. Build and install
went without any probles, and I learned just enough elisp to
get scel running in emacs.
Now the problem seems to be that sclang and scsynth do not
talk to each other. How is the sclang -> scsynth UDP link
configured ? There is no option in sclang to set the port
number it sends to so probably this is fixed. OTOH, scsynth
requires a port number for the -u option, but what should
this be ?
TIA,
--
Fons
The second stable release (0.9.0) of JAMin - the JACK Audio Mastering
interface is now available for download.
JAMin is a GPL-licensed, realtime mastering processor designed to
bring out the detail in recorded music and provide a final layer of
polish. Every effort has been made to ensure a clean, distortion-free
signal path. All processing elements use linear-phase filtering to
eliminate phase distortion.
JAMin runs on Linux using the JACK Audio Connection Kit, a low-latency
audio server. JACK connects multiple applications to a single sound
device, and also share audio among themselves. We rely on other JACK
applications (like ardour, ecasound, or rezound) for playback and
recording.
Homepage
http://jamin.sourceforge.net/
Download
http://prdownloads.sourceforge.net/jamin/jamin-0.9.0.tar.gz?download
New dependencies (since jamin-0.8.0)
swh-plugins >= 0.4.6 <http://www.plugin.org.uk>
liblo >= 0.5 <http://plugin.org.uk/liblo>
(optional, for OSC scene change control)
All other dependencies remain unchanged, the README file has a
complete list.
Usage instructions
http://jamin.sourceforge.net/Using_JAMin.html
Changes since jamin-0.8.0
* Limiter improvements:
-- uses new fast lookahead limiter (LADSPA:1913) from the
swh-plugins >= 0.4.6. The old limiter had some undesirable
sonic artifacts.
-- changed order of final gain and limiter stages.
* OSC control for scene changes:
-- accepts "osc.udp://localhost:4444/jamin/scene" messages
-- new `jamin-scene' command sends them
-- new jamin-cont plugin (LADSPA:1912) sends scene change
messages to the JAMin process via OSC.
(This plugin works with ardour, ecamegapedal and
applyplugin, but segfaults in jack-rack when removed.)
* Increase number of scenes from 6 to 20
* GUI improvements
-- extensive context-sensitive help
-- color editor for highlights and text
-- options pulldown menu for spectrum, crossfade, and EQ
* Better bypass control
-- specific bypass controls for EQ and limiter
-- separate Active/Solo/Bypass for each crossover band
* Global state settings saved in XML file.
* Expanded Russian translation
* Many bug fixes.
--
joq
http://sourceforge.net/projects/liblo/
Theres a mailing list, liblo-devel.
For those who dont know, liblo is a GPL'd implementation of the OSC
protocol, which is a Remote Procedure Call language, designed for audio
software.
http://www.cnmat.berkeley.edu/OpenSoundControl/
I feel that liblo is now mature enough to stand on its own, so I'm moved
it to sourceforge. I welcome help and advice from anyone who feels like
giving it, though I do want to stop the API getting too sprawly.
The current state is that I've just added UNIX domain and TCP transport
support, but they are a bit rough around the edges.
The current contents of the TODO list are:
* Bundle support [needs NTP, argh]
* Add source address to received messages
* More/better regression tests
* More/better example code
* Normalise URI handling [needs discussion]
* GTK/Qt/etc. tie code [maybe belongs in another library]
* OSC pattern matching [still not sure about this one]
* Rendevous/OpenWhatever (eg. howl) service discovery [low priority]
* Add a JACK transport layer [maybe, low priority]
* Make TCP work to spec (currently the TCP server only reads the 1st packet
from every conneciton) [low priority]
Cheers,
Steve
(not the "feature complete" release I was planning, but the list
of changes is enough to warrant a new beta number)
Changes since 0.9beta18.4:
* optimization flags patch from rob holland
* drag-n-drop onto the editor track area where there is no track
creates an appropriate track (or multiple, if appropriate)
* URL-decode drag-n-dropped filenames
* numlock can no longer mess up region dragging
- if numlock generates whatever "Meta" is supposed to be
then Ardour changes its notion of what Meta is.
- if you see a warning about this in the terminal at
startup, you should fix your keyboard map using xmodmap
or equivalent
* remove debugging output from process() tree
* removed VST audiomaster callback debugging output
* UI for VST plugins now offers presets
* VST plugins now save and restore their state
- "chunk"-using plugins are not yet supported
* initial tempo and meter markers cannot be moved, even after
editing
* when cleaning up, prompt user about
each unused playlist (options are
delete, keep, cancel cleanup)
* never decrease the number of panning buffers
* Can't create a new session on top of an old one anymore. It didn't
overwrite the old session, but you would get 2 master tracks and
whatever was in the template, if any.
* freezing tracks
- implementation of freeze completely redone
- unfreeze now works across invocations of ardour, so you
can freeze, exit, restart and unfreeze.
- freeze hides any insert editors, makes rec-enable
and redirect display widgets insensitive and
locks the new region.
- freeze can be used over and over again
* reworked relationship between bounce + freeze
* removed "Loop Region", renamed "Loop Selected Region" as "Loop
Region" because of click-selects selection model
* finally made solo button in mixer strip have the same font size
as other buttons
* removed edit history button and associated code from editor track
* removed "automation style" buttons since they did nothing
* more automation line + point patches from gerard
* fix a couple of valgrind complaints
* region trimming now obeys ignore-snap modifier
* search unused playlists in Session::playlist_by_name()
* added 2-phase unserialization mechanism (Session::StateReady signal)
* empty playlists assigned to correct diskstream at startup
* removed some debug printfs and some dead code
* make Shift_R work like Shift_L
* fix crashing bug with zero-length fade ins
* internationalize all editor operation strings
* transport control is insensitive when there is no session
* fix problems with "region->fill track" where region is at or
beyond current session end
* new mixer strips use the current default width setting
* always save to snapshot
* snapshotting changes the current snapshot to whatever was saved
* delete click_io object at session close, thus allowing new
sessions to load correctly
* fix problem with code ordering in normalize-region operation
(undo after normalize-then-cut will now do the right thing)
* force selection.clear() in file selection browsers when widget
selections are cleared
* committed patches so that ardour now compiles with gcc-3.4. There were
several patches from multiple people. So please let me know if I missed
one.
* patch from nick_m that fixes a problem with the editor-mixer
button.
