i finally have mscore working with linuxsampler as the sound engine.It's taken a lot of work because of Mscore's insistence on capturing and clutching tightly to its chest, the RTC timer. I've mentioned this before, and there's been some discussion already on the 'right' of an app to exclusively possess the timer, so, enough said. However there's the smell of success here, as a problem is overcome. And the answer's fairly simple.
For those like me who cringe at the general quality of generic soundfonts (Strings usually suck by default), and desire a more pleasing playback, the first step is setting up a generic midi soundbank, using GM format, in linuxsampler, and assigning all your quality instruments.
Assign in mscore edit/preferences 'midi output' as the, obviously, midi out device.
jack needs to be stopped.
Start Mscore first, so that you don't get the accursed 'RTC Timer is mine, all mine, and someone else has it, so i'm not playing' message.
Start Jack, with alsa seq as the default midi device. (I'm using jackmp, so this is modified with the 'Jack_control' mechanism in a terminal. And it's from the terminal that i start jack, with, you guessed it, 'jack_control start')
Start Linuxsampler, and load your brand spanking new GM default lscp template. Cable linuxsampler's audio outs to your output of choice, be it system, or another app for recording or playback.
Either in patchage (my weapon of choice) or qjackctl, cable mscore midi out to linuxsampler midi in. (and the midi port in linxsampler is alsa. I've got this working with the mighty a2jmidid bridge 'a2jmidid -e', but if you're scribing notes in mscore, the simple setup works ok.)
And you should be good to go.
Toby at launchpad has done a good job of packaging Mscore 0.9.4 for UBstudio hardy and Intrepid (testing), and it all installed well. So you get all the new features to enjoy as well. (Thanks Toby.
)I've long been a fan of Mscore and that enthusiasm continues, hence this quick post. This workaround is for those who starting on the mscore journey, and may be struggling with the whole RTC thing, and crave a better quality playback option. Thanks to Werner and the team, and in addition to the wonderful Rosegarden, we are fast getting spoiled for notational choices.
I made mention in another post of a 'dictionary' in RG, where the user could assign ports and channels to an external device like Lsampler, and select sounds for articulations by port and channel for playback, as well as set timing parameters for symbols, like fermata. We have the first bit in abundance in RG, and the second would be useful, imho, in the future at some point. (And i say this acutely aware of the formidable task the RG team have at the moment, porting to QT4.)
I would hope that something like this, or a uniquely 'linuxish' framework, that puts this control in the hands of the user, could also find its way into the Mscore project at some point.
Alex.
_______________________________________________
Linux-audio-user mailing list