From: Alex <com...@gm...> - 2010-09-09 18:42:15
|
Tim, is it possible for me to hack the jackmidi buffer size in the src file somewhere? (At least to try it, and see if it's the culprit) Is the buffer size just an arbitrary number i can change , or is it a bit more complicated than that? Alex. p.s. Not sure what constitutes a perceived maximum size for muse in terms of number of track, busses, ports,etc..., but so far it's running 176 midi tracks, 78 wav tracks, 117 jackmidi ports, 68 inputs, 46 group busses (24 of which are jconvbusses), and so far, 32 EQ plugins (ladspa) as well as jconv in and out. Have to say i'm mildly impressed at that, particularly as muse is only chewing between 26-30% cpu for everything on my faithful amd dual core X64. I've killed off any other GUI's (it's all cli here now, apart from muse) and no xruns from jack. I still get the occasional mem cough when i shut muse down (64bit challenge i think), but that's every 3rd day or so, and i've been hammering muse every day, and most nights for a while now. Significantly, muse doesn't hassle Linuxsampler prompting error messages, which hasn't been the case with other linux daws and apps. (At least here, jack1, jconv, and LS are best mates with no errors) I'm going to push it further from here soon, and see just how many tracks and busses i can add, before muse starts spitting at me. I know this seems a bit extreme for writing music, for what is perceived as a rare use case, but these numbers are fairly modest/normal for scoring large works, using multiple sample libs. Muse seems to be doing well, and i've added the info so you have a clearer perception of another type of use case, and the numbers involved. So i guess any requests or problems i've noticed should be considered in this context. |