Linux-audio-dev December 2004

linux-audio-dev@lists.linuxaudio.org

74 participants
64 discussions

[linux-audio-dev] ANNOUNCE: liblrdf-0.4.0 release

by Taybin Rutkin

Changes: SWIG support. Better generic id creation. Moved to sourceforge. Available at http://lrdf.sf.net -- Taybin Rutkin

19 years, 11 months

[linux-audio-dev] dssi plugins / UIs - compiling without ui

by Julien Claassen

Hi! I've got a question: Could you give an option to the configure scripts (for hexter and the like) to NOT compile the GUIs? I've got a problem with the guis, for I don't have a lot of the graphic devel stuff installed. This would be very kind! As to text-based UIs, I've looked into the simple_sampler UI, but don't understand it yet (very much)... :-) I've to continue looking into it, I'm not a very good programmer. But I'll do it! Kindest regards Julien -------- Music was my first love and it will be my last (John Miles) ======== FIND MY WEB-PROJECT AT: ======== http://ltsb.sourceforge.net - the Linux TextBased Studio guide

19 years, 11 months

[linux-audio-dev] Ardour 0.9beta22 released

by Paul Davis

I am happy to announce the release of Ardour 0.9beta22. With over a hundred notable new items, and only a handful of bug fixes between this and 0.99, this is a significant release of Ardour. Ardour's bug tracking database continues to have 260+ live bug reports and feature requests, and there are likely twice that number still present in the program, but we identified those that will be fixed for the 1.0 release in order to try to gain control of the release process. 0.99, which just might make it out by Christmas (December 25th, but don't hold your breath), will be the endpoint of code development for 1.0 per se - the difference between 0.99 and 1.0rc will be related to release engineering (making sure that the install of Ardour does all that it should). Fetch it here: http://ardour.org/releases/ardour-0.9beta22.tar.bz2 RELEASE NOTES FOR ARDOUR 0.9beta22 * Added support for midi parameter feedback. When enabled, controls that are bound to incoming midi events will send out that event when modified. This lets you control generic midi control surface with motor faders and/or led encoders (like the incredibly affordable Behringer BCF2000, which was used for testing), to match the internal state. Note the extra/changed options in the option editor's midi tab. Also note that automation playback does not yet send midi feedback. Reminder: to bind a fader or bar-control to midi, do a Control-Middle-Click on it, then send some midi control. * Added generic midi bindability for mute, solo, and rec-enable controls. to do it, you can do Control-RIGHT-click on the mute and solo buttons (the normal binding click is taken on these buttons for other functions). For the Record-enable button, it is the normal ctrl-middle-click. You can also pick it from the context menu of the mute button. A menu will be added to other buttons later. * fixed some midi prompter dialog issues * fixed some midi binding state saving issues * change entire buffer management scheme to be (more) RT-safe * add -V/--novst flag to avoid using VST even if it was compiled in * templates are mpeinstalled * major changes in thread handling to fix some serious errors in the way ardour handles threads when exiting. * make export work when sync'ed with JACK * add "timecode-source-is-synced" option to ardour.rc(.in) * new playlist selector (uses a tree structure in a scrolled window, instead of menus) * fix bug with MMC rec-enable handling (only odd-numbered tracks could be enabled) * make region export progress bar move in the right direction * more MTC slaving fixes and improvements (not done yet) * remove wrong-thread-calling of Session::clear_event_type(), which should help a number of loop-related crashes * remove erroneous use of "abort" in function call * new playlist selector * Automake-1.7 or higher is required for libmidi++ now. Should be pretty standard on distributions by now. * add "nudge by capture offset" operation for regions * enlarge playhead/edit cursor arrows a bit * avoid duplicate keyboard target registration * pay attention to "virtual" window enter/leave events, but continue to ignore "inferior" ones (fixes many keyboard focus issues) * cancelling tempo/meter create dialog doesn't insert a default tempo/meter marker * tempo/meter create dialogs have minimal WM decor, and use standard ArdourDialog API * 2d panner fixed in many ways (still not to my taste and not really correct) * context-click on panner shows context menu (Bypass is the only entry so far) * general panner state load/restore fixes * master outs added to connections menu * added optimizations for peak metering (use of fabsf, flag constants as floats, avoid implicit double/float casting) * handle disk over- and under-run errors in the GUI, and be more comprehensible to regular users. * remove zoom changes from the undo/redo history * fix panning bypass for 2d panner * fix stereo panner "mute" menu item so that it doesn't toggle panner mute state every time it pops up * add "OK" button and remove window decoration from "can't connect to JACK" dialog * sessions still load when audio files are missing or corrupt * complete redesign of slave-following design * stop disk-(over|under)run handler from executing in the wrong thread * fix appalling C/C++ bug preventing proper use of MIDI port tags * make solo/mute button mouse bindings more consistent with rest of interface (Ctrl-middle click is now used for MIDI learn here as well) * add solo button context menu * add new MessageWithConfirmation object to libgtkmmext * fix crash in handling DiskOverrun or DiskUnderrun messages * correct region fade in/fade out curves after split * fix complete loss of all optimization flags * added patch to allow snap-to-SMPTE-frames * add redo-ability of automation track "clear" operation * add optimization to AutomationList::_eval, and use same design in Curve::_eval * fix failure to use a search path when looking for templates * fix inserting zero-input plugins (divide-by-zero error) * improve sync-with-JACK after MTC slaving breakage * session events are no longer queued during session loading, which should reduce or eliminate JACK kicking ardour during startup. instead, these events are now processed by the GUI thread as the session is loaded. this stops the process() callback from potentially having a long list of things to do when its first called, thus causing JACK zombification. * recover-from-crash functionality a) all audio files get a header written into them at creation b) when session rec-enable is turned on, we store information about the files we'll be using for recording c) if ardour (or the computer) crashes during capture, ardour discovers this information, and uses the actual size of the audio files to modify their headers. new regions are then created and added that correspond to this data. d) there is GUI dialog to check on the user's wishes as to whether or not to use this "pending capture" information * clear-all-markers and clear-all-ranges no longer crash with no session * "duplicate client/no jack server" dialog grabs focus for its "OK" button * change all of libardour + gtk_ardour to use <stdint.h> types (uint32_t etc.) and <inttypes.h> for printf-style tags. a few "int" declarations remain. * redesign + reimplement plugin input/output count handling. try it and see. i think its way better. your mileage may vary. A note on the new scheme: a) sends are ignored entirely in figuring out plugin configuration. they send a signal *out* of the route, and have no effect on the flow within the route, so this seemed right to me. b) the first insert (plugin or port insert) is configured the match the input count of the route c) then we work *backwards* through the inserts to make them match the configuration of the previous insert. the output configuration of the last insert is left "free". d) we recompute in/out counts whenever we - add an insert - remove an insert - reorder inserts - change the input configuration of the route e) if the computation "fails" (i.e. no way can be found to make them work), the order is left as is, and an error is delivered via the GUI. * fix ancient bug in computing lengths of overlaps under 1 of the 4 possible types of overlaps. this fixes problems with chunks (and possible range ops) that extended to the end of the session (and beyond) * automation tracks now have some modest logic for handling height changes. alas, i can't hide the "expansion arrow" when i want to. * there is now a context menu for the region list display, complete with "Hide", which will move the region into a special subtree titled "hidden". I am still working on what a "Remove" option should actually do. * selecting a region in the region list selects all regionviews that are considered "region-list-equivalent" * mouse mode buttons never end up with 2 active at the same time * fix for new insert in/out counting scheme as it affects port inserts * libardour: 0.843.2 * fix deep order-of-operations problem that prevented Session->{Open,Recent,New} working with new plugin in/out code. * delete failed redirect, not just plugin, when adding a new plugin fails * fix for crash at session close caused by list iterator botchups (thanks to nick murtagh for discovering this one) * (in previous un-announced commit) patch from nick mainsbridge for crashing bug induced by tempo change * JACK sync is mostly working again * new, not necessarily improved Route::_reset_plugin_counts() (still a work in progress, debug output generated) * fix for crashing bug when deleting a plugin-based redirect * every use of the %f specification is now bracketed by enforced use of the POSIX locale (for libardour; GUI is still 100% locale-sensitive) * main transport clocks shown meter+tempo in BBT mode * CD markers+CD TOC/CUE patch from Nick Mainsbridge * all connections to Session signals made by editor are dropped when session is closed (avoids title bar being reset to show session name) * add new "mtc_qtr" signal to libmidi++, emitted every time an MTC quarter frame message is received. * redesign MTC slaving to sample apparent master speed from the process callback, rather than set it from the MIDI signal handlers. * attempt to use jack_client_open(), if its available. * alter --jack-client-name argument to be just --name, in keeping with some other JACK clients. * Ardour now handles JACK buffer size changes - note: VST plugins are likely not correctly handled at this time * move JACK related material to its own top-level menu bar item * better handling of JACK multiple disconnect/reconnect actions * add control for JACK buffer setting to that menu * make reconnect/disconnect to/from JACK sensitive at the right times. * finally (?) fix mouse mode buttons to work correctly at startup AND change default bindings of "r" and "g" * remove "disk rate" metering stuff from GUI * modularized mixer strip components internally * renamed the Route Parameters window to Track/Bus Inspector, redesigned to use tabs and the same context menu as the mixer strips. * rectified the plugin list handling in the mixer strip and the inspector window. * ardour_ui.rc is modified, so you might need to take action to get it used. The simplest way is to copy the new one to ~/.ardour/ardour_ui.rc * fix long standing bug where closing+saving a snapshot saved to the base version of the session. Thanks to Doug Mclain sticking with this bug until we finally got to the bottom of it. * make the "ask about saving" dialog use the snapshot, not session, name. * new metering design paves the way for 1.0 work on multiple meter points. the "pre" button is no longer a toggle, but a N-state button, where N is currently 3 (input/pre/post). Rec-enabling a track automatically switches to input metering. * if a diskstream cannot be rec-enabled because it has no known connections, check with JACK in case it was connected by starting from another JACK port. * catch "no more port" errors when using "Connections" * don't try to unregister non-Ardour JACK ports * encapsulate and rationalize the way diskstreams switch between audio files before capturing, and how they store capture pending info. This fixes several crashing bugs reported by Doug over the last 10 days. * handle block size changes for VST plugins. Note: some VST plugins do not handle this correctly, and will cause crashes. * remove call to VST "resume" method from process() thread. This *may* have some ramifications for certain plugins that don't understand the intended thread relationships for a VST host. Haven't seen them yet myself. * fix ordering of creating the "Master Outs" "Connection" and the Master bus ports. This bug has led to some potentially nasty stuff happening with session files that appear to drop all connections from tracks to the master bus. [ Not totally solved. ] * Ardour will build on macosx out of CVS now, with the right toolset installed on OSX and recent JACK CVS. * playlists now have "frozen" property. - frozen playlists cannot be selected by any track. they are accessible only by freezing or re-freezing a track. * rec-enable buttons and playlist selector buttons start out insensitive if a track is frozen at load time * some modifications to ArdourMessage to make it even easier and more consistent to use. * make "cannot record - no inputs" message more useful, informative, explanatory and use ArdourMessage. * remove middle click mutes regions behaviour (for safety) * add behaviour to the meter button that is similar to the mute/solo buttons: shift-ctrl-click adjusts all metering ctrl-button2-click is momentary ctrl-button1-click affects mix group * cleanup improvements (SIGNIFICANT) - fix stupid bug that led to sources having paths with "//" in them. this meant that string comparisons between paths found using a directory scanner and the source paths would fail to see them as identical, and would conclude that the file was not being used withing the session. - remove peakfiles for files that are moved into dead_sounds - actually unlink files when the wastebasket is cleaned out - search all snapshots to ensure that cleanup does not remove files used by one snapshot and not another. * sample rate indicator now includes period size indicator in msecs

19 years, 11 months

[linux-audio-dev] ANNOUNCE: BEAST/BSE v0.6.4

by Stefan Westerfeld

BEAST/BSE version 0.6.4, BSE-ALSA version 0.6.4 and BSE-PortAudio version 0.6.4 are available for download at: ftp://beast.gtk.org/pub/beast/v0.6/ or http://beast.gtk.org/beast-ftp/v0.6/ This is a development version of BEAST/BSE, the BEdevilled Audio SysTem and the Bedevilled Sound Engine. BEAST is a powerful music composition and modular synthesis application released as free software under the GNU GPL and GNU LGPL, that runs under unix. BSE-ALSA is an ALSA driver and BSE-PortAudio is an experimental PortAudio driver for BSE. The project is hosted at: http://beast.gtk.org A mailing list is available at: http://mail.gnome.org/mailman/listinfo/beast/ GUI skins, example sounds and instrumets for BEAST/BSE as well as screenshots can be found at: http://beast.gtk.org/browse-bse-files.html http://beast.gtk.org/screenshots/index.html This new development series of BEAST comes with a lot of the internals redone, many new GUI features and a sound generation back-end separated from all GUI activities. Outstanding new features include support for skins, many sample file formats, MIDI file import abilities, an improved piano roll widget, the track editor which allows for easy selection of synthesisers or samples as track sources, loop support in songs, mixer support, unlimited Undo/Redo capabilities and MIDI automation. Overview of Changes in BEAST/BSE 0.6.4: * Fixed storage logic that caused data loss when overwriting BSE files * Added -N and -n=<nicelevel> options to drop nice level priorities * Removed libbsw from package, glue code is inlined now * GCC-3.4 build fixes * Fixed high CPU consumption on scope updates * Added spectrogram view for module outputs * Minor GUI fixups and miscellaneous other buglet fixes * Updated Canadian English translation [Adam Weinberger] * Updated British English translation [David Lodge] * Updated Czech translation [Miloslav Trmac] * Updated Dutch translation [Tino Meinen] * Updated Spanish translation [Jorge Gonzalez] Overview of Changes in BSE-ALSA 0.6.4: * Build fixes. Initial Release of BSE-PortAudio 0.6.4: * Provide a PortAudioV19 PCM driver for BSE -- -* Stefan Westerfeld, stefan(a)space.twc.de (PGP!), Hamburg/Germany KDE Developer, project infos at http://space.twc.de/~stefan/kde *-

19 years, 11 months

[linux-audio-dev] [ANNOUNCE] rtc based MTC generator

by Florian Schmidt

Hi, i have written a small MTC generator that sends MTC via alsa_seq. It is based on the RTC running at a frequency of 8192hz (this introduces systematic jitter) and currently only sends 30 FPS (nodrop? i think :) It seems to perform fine on my 2.6.x RP kernel (with the RTC irq at highest prio above soundcard irq (jackd with ardour running at prio 60)). Here's the webpage (with tarball): http://www.affenbande.org/~tapas/wiki/index.php?rtc_mtc_gen Maybe someone finds it useful. Let me know what you think. Is there any software other than ardour on which i can try it? I don't have any hardware here that speaks MTC either.. Flo -- Palimm Palimm! http://affenbande.org/~tapas/

19 years, 11 months

[linux-audio-dev] Linux found to have less than 1% of the "normal" occurrence of bugs!

by Cornell III, Howard M

http://www.embedded.com/showArticle.jhtml?articleID=55301858

19 years, 11 months

[linux-audio-dev] adm: list archives at eca.cx going away

by Kai Vehmanen

Hello, this is not an urgent issue, but I'd like to warn in advance that the laa/lad/lau list archiving at www.eca.cx/la[adu] will stop some time next year. I'm not sure of the exact date, but sooner or later anyway. Note that I'm not going to take the existing archives offline, so all links pointing to mails at eca.cx will still work. Fortunately we nowadays have many other archives of the lists. We just now need to pick a new official archive and update the links at: http://www.linuxdj.com/audio/lad/archive.php3 I propose that we use the archives at columbia.edu: http://www.music.columbia.edu/pipermail/linux-audio-dev/ ... and so on. Very good archives can also be found at gmane.org: http://gmane.org/info.php?group=gmane.linux.audio.devel ... and so on. -- http://www.eca.cx Audio software for Linux!

19 years, 11 months

[linux-audio-dev] Re: rosegarden, alsa_seq and usx2y rawusb

by Stefan Turner

Rui, Thanks for your reply. I am trying to build Jack first just to see if it'll work with the CCRMA Alsa I've already got. So far I've done: [root@localhost jack-audio-connection-kit-0.99.35.3usx2y]# aclocal [root@localhost jack-audio-connection-kit-0.99.35.3usx2y]# autoheader [root@localhost jack-audio-connection-kit-0.99.35.3usx2y]# automake configure.ac: required file `config/mkinstalldirs' not found I can't find anything about this file, not even in configure.ac, and I really don't have a clue about this kind of thing (sorry if it's something really obvious). It certainly seems to stop the rest of the build working though. Any ideas? Thanks Stefan Turner ___________________________________________________________ Win a castle for NYE with your mates and Yahoo! Messenger http://uk.messenger.yahoo.com

19 years, 11 months

[linux-audio-dev] [Fwd: Graphical dataflow programs violate patents]

by Xavier Amatriain

-----Forwarded Message----- > From: Xavier Amatriain <xamat(a)iua.upf.es> > To: iua-mtg(a)iua.upf.es > Subject: Graphical dataflow programs violate patents > Date: Tue, 14 Dec 2004 10:57:56 +0100 > > I read that National Circuits sued and won a case against Mathworks for > their Simulink product infringing a number of paterns (September this > year). > > According to a summary of the claims (which you can read here > http://cafc.bna.com/03-1540.pdf): the dispute was about the term "data > flow diagrams" (which are interpreted) as a graphical computer program > whose execution follows a set of semantic or operational rules as > follows: > > 1. The order of the operations is not completely specified by the user > 2. The order of operations is implied by data interdependencies > 3. a node may only execute after all its inputs have become available, > and, > 4. outputs are generated after a node completes execution > > It turns out that most graphical audio programs such as Max, Pd, OSW or > CLAM's Network Editor may be infringing these patents. Luckily this is > still not valid in Europe and we hope will never be. > > Read more about this story at: > > http://patentlaw.typepad.com/patent/2004/09/federal_circuit.html > http://digital.ni.com/worldwide/bwcontent.nsf/web/all/03B9A7C68030C42D86256… > http://www.mathworks.com/company/pressroom/articles/article7951.html > http://news.thomasnet.com/fullstory/16801 -- /********************************** ********* Xavier Amatriain ******** ****** Music Technology Group ***** ** IUA- Universitat Pompeu Fabra ** ****** www.iua.upf.es/~xamat ****** ***********************************/

19 years, 11 months

[linux-audio-dev] Re: [Fwd: Graphical dataflow programs violate patents]

by Juhana Sadeharju

>From: Xavier Amatriain <xamat(a)iua.upf.es> > >> I read that National Circuits sued and won a case against Mathworks for >> their Simulink product infringing a number of paterns (September this >> year). The first patent should expire soon as their software Labview was launched 1986. It contains some essential items. I collected papers on this topic recently. In Labview user creates the circuit editor and the control panel. Both existed alone earlier. Labview includes special GUI elements in the circuit (see Reaktor, for example). I'm in belief that this connection can be done slightly differently without violating the patent. When Csound type software was first written? 1960? 70? There are other Labview patents. One patent patents what I have invented prior them. I remember posting it to a mailing list, but have not yet found it. I have not read the patent but my idea was that when the module is a subgraph, a click on the module opens up a graph editor; and when the module is a script, a click on the module opens up a text editor. Obvious? Juhana -- http://music.columbia.edu/mailman/listinfo/linux-graphics-dev for developers of open source graphics software

19 years, 11 months

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

Linux-audio-dev December 2004