Interest in developing a high quality SoundFont for MuseScore?
It seems to me there are a few competing formats for samples, and the SoundFont (SF2) format has not been the flavor of the month for some time. Most of the sites I see sharing SoundFonts are pretty old. Does anyone have a sense of whether there really is still much active development in this format? Is there anyone here that has any experience in SoundFont development? I'm wondering about the possibility of seeing a high quality SoundFont that integrates really well with MuseScore - one that starts with GM instruments, but also includes the other common orchestra and band instruments that were left out of GM (bass clarinet being a personal pet peeve of mine), program change info coordinated with MuseScore's instruments list. Maybe a master SoundFont with every instrument in MuseScore's instrument list, and then a series of subsets tailored for specific uses, so we don't always need to load a gigantic SoundFont.
Right now, FWIW, I'm "mostly" happy with FluidR3 as the best sounding soundfont overall for my purposes (jazz and orchestra), but it's GM-only, and also the saxophones leave something to be desired.
Comments
Also see this .
In reply to Also see this. by chen lung
Yep, sorry, forgot about that thread, which I had even commented on. But I hadn't forgotten the Sonatina Symphonic Orchestra (SSO) project referenced there, and one of the things I am trying to get a handle on is, which would be more viable - adding SFZ support to MuseScore to make use of that sample collection, adapting that collection to SoundFont format, improving an existing SoundFont like Fluid, etc.
One deceptive thing about the Sonatina Symphonic Orchestra demo - that clearly isn't just the default playback of a notation program, but rather, something that has been very heavily massaged to really nail the dynamics, articulations, etc. See this discussion on the SSO site, in which several people familiar with SSO, including the developer, comment on the viability of integrating SSO with MuseScore:
http://sso.mattiaswestlund.net/forum/viewtopic.php?id=153
It's pretty obvious from the comments there that much of realism we might perceive in a piece produced with SSO is at least as much a function of the painstaking work done by hand in DAW software as of the sample collection itself. I also once tried an experiment of building a SoundFont from the samples in SSO, using a tool I found that could do this (http://www.extranslator.com/), and the results were mixed at best on my initial attempts, but I gave up pretty soon. I heard enough to suggest it was entirely possible that the samples themselves are not significantly better than those in Fluid. So the work involved in re-architecting MuseScore to work with SFZ just for the sake of SSO might not be worth it. Someone else might want to put more effort into that, though.
Another way to test this would be if anyone here is really good at doing the sort of DAW work that went into the SSO demo, and see what kind of results can be coaxed out of Fluid that same way.