interview: Richard Bown and Rosegarden
Rosegarden is one of the main usable MIDI sequencers for Linux. Originally a university project and written for SGI/Irix, it was ported to Linux some time back and now has a KDE GUI. Rosegarden 4.0 is out now and if you're around London England on Oct 9-10, 03, you can catch the project members at Linux Expo at Olympia. Here we talk to project member Richard Bown about it all.
Richard Bown's homepage
Rosegarden
A couple of years ago when I was searching around for sequencers for Linux, Rosegarden was one of the few things I could get to work and there seemed to be a nice whiff of anarchic feelings about the website. What's the history of rosegarden?
Rosegarden started as a final year project at Bath University in late 1993. My friends Chris Cannam and Andy Green had John ffitch (Linux audio and Csound guru) as their project tutor. I don't quite know who's enthusiasm for the project was the greatest but what started out as just a piece of project work turned out to eventually become something of a labour of love for both Chris and Andy. The early design documents from this time are still available on our project home page as is a longer version of the history of the project.
Rosegarden (named after a Bauhaus song incidently) was originally written for the SGI IRIX platform and it wasn't until late 1995 after I had got involved (i.e. bought a PC and a copy of Slackware 3.0) that we decided to start a Linux port of it. Andy had been working on Windows version at that time so I just took over his responsibilites for the Linux sequencer and took it from there. We implemented the sequencer in the OSS/Voxware drivers and when Guillaume Laurent then appeared with a big patch of code and a lot of enthusaism things moved quite quickly from there. It's nice to know that Rosegarden 2.1 is still very much out there and still being used to this day.
While the main developers are all musicians and fans of (sometimes quite challenging) music we're also quite passionate about how our project is perceived. I think that the slightly "anarchic" feeling you sense is down to us trying to challenge the average user's (and above all else the average developer's) perception of just what makes music software good anyway. Our intent (and this is especially true of the latest version of Rosegarden) is to make a really usable piece of software - there's no point us having a host of features if the user can't find how to get to them. So we'll always listen to criticism and always try to improve our application if we can.
How have you organised the coding? Have people just signed up for their area of interest and then gone ahead and done it or has there been some kind of heirarchical approach ... or a mixture?
Not really a hierarchy as such. For Rosegarden-4 theoretically it breaks that Chris does mainly Notation, Guillaume does GUI, framework and infrastructure stuff (printing, scripting languages) and I do the sequencer (MIDI, Audio etc). In practice the GUI work is pretty evenly shared and where we all spend a good portion of our time, arguing! We now have another guy with CVS write access (Hans Kiesermann) and a few more occasional contributors and bug fixers but the core deveopment team of the three of us do most of the talking and most of the coding for the moment.
Chris and I have both had good amounts of time off in the last year to work on Rosegarden which has enabled us to move quite quickly. This has been really useful to get a good idea of where the project is going. As the number of features left remaining to be implemented is falling we can now start to think about getting a community of beta testers too.
Do you think MIDI is going to survive as a music making tool?
I think MIDI is still pretty much underrated as a music making tool. The power and control that the user of a MIDI sequencer can have over a soundcard, a rack synth or all manner of MIDI controlled equipment is quite simply stunning. Having said that the major turn off for a lot of users is the complexity of the routing, connecting, programming and also keeping and restoring system state between sessions. With the advent of the home studio or PC studio these concerns have become even more valid - the average user with an average PC or Mac can make some pretty intersting music if they know what they're doing. Half of knowing what they're doing is to keep the complexity of the software and the hardware down as much as possible. Pieces of software such as Rosegarden should aim to offer this trade-off between getting results and offering power and simplicity. There will of course be no one solution to any one user's wishes but this is what makes music software generally so interesting - the same as with graphics packages.
Part of my attraction to MIDI is that you can change instruments and configurations so quickly. The same signal routed to your soundcard can suddenly be re-routed to a box on the other side of the room just with a few clicks of the mouse. When you want to find the right sound for a track quickly it's much easier to sit there with a nice list of patch names in a drop-down on your MIDI sequencer than it is to step through them on the hardware itself. The MIDI studio might take a little bit more setting up but when you're composing it this time spent pays dividends. Of course if you don't like MIDI instruments you don't like them and nothing is going to change that.
A simpler answer to the question "Why MIDI" is also that Chris' passion is really for providing and using professional quality notation editing and mine is for providing and using recording and production facilities. MIDI is a good half-way point between these goals and indeed Rosegarden-4's internal structure is geared around providing Notation support, MIDI support and audio support within one framework. The fact that for me it's also an incredibly useful compositional tool is a great (and of course intentional) bonus.
and MLAN? I agree that MIDI is a wonderful tool. One thing I wonder about is how many people use it in a way where its speed capabilities become a problem.
The raw MIDI protocol is pretty speedy anyway - it's capacity that's the problem. You can go around sending samples down MIDI through SysExs if you want but it's not recommended if you want to do anything else on the MIDI network. mLAN or whatever alternative there is sounds great but I question the need for it yet. Thing is though, I would - I'm slow to take on new technologies and digital mixers scare me enough let alone a new protocol to connect them. I just like to record music using the tools I've got available at the time - I'm not particularly interested in the latest gizmos.
What do you think of the current state of affairs in Linux audio? The future?
Things are getting very exciting at the moment. Not only do we now have great sound drivers and support from ALSA now but there's JACK and there's LADSPA and a lot of projects looking to make the most of these enabling technologies. On top of this professional quality sound-driver-level support KDE and GNOME are maturing more and more every day and therefore so are the applications themselves. There seems to be plenty of projects that are getting very usable and this helps spur everyone on to do better.
While everything bodes well technically I think now is the right time to change the focus from the more mundane concerns of the application authors and testers to that of attracting potential musicians to Linux because it has simply has great sound support and great apps. This can work on two levels - both with the attitude of the developers to the users and the attitude of the packagers and distributors to the platform itself. AGNULA (GNU sound distro project with RedHat and Debian and EU funding) is doing something to address the second concern but its organisation appears rather opaque and its goals, while admirable, depend ultimately on a set of developers who they don't and can't control. Bottom line is I think that unless they open the project up some more and involve people like us who actually write the apps then they could find themselves chasing their own tail. The opportunity is there now to make Linux sound apps that work for the average user on the average Linux distribution. As distro makers hopefully follow SuSE's lead and start to bundle ALSA (and of course we head towards the 2.6 kernels) we should start to see sound apps becoming easier to install. This means that the average user/ installer can then potentially be impressed with out-of-the-box sound applications - ones that just work and work well.
It's definitely an interesting time. We're showing Rosegarden for the first proper time at the Linux Expo in London on October 9-10th and we hope not only to be able to demonstrate some nice features but also that we can make music pretty easily and persuade people that we're about the musician (whether they be amateur or professional) and not just the technology. That's a bit of a cheesy line but I think I can safely say that our angle is letting the average person with a PC think feel that they can make and record music with Linux and that it doesn't have to be a pain in the arse to set it all up. This is kind of where I've been coming from for a couple of years anyway with my guide to recording: www.bownie.com/guide
and I hope this is where we end up coming from with our software: www.all-day-breakfast.com/rosegarden
That's interesting that you don't mention that these audio apps aren't costing them a small fortune either. I guess, as far as use is concerned, the app you create wants to be able to compete as a good tool not as a free one. Now I think about it, there are so many copies of fruity loops, reason, etc out there that the "free" marketplace is quite full. Still, with Linux, people know that they're not breaking the law and most people think that's a good thing!
Quite. It's relatively easy to get your hands on a "free" copy of many of the commercial sound apps for Windows but of course that's illegal and the very fact that the software has to be modified to bypass its copy protection can sometimes make it unstable. If Windows is unstable enough by itself then unstable, CPU intensive music apps on top of that holding your precious music could be something of a recipe for disaster. If we can make Linux the stable and reliable alternative for the desktop and studio musician then we've got a potential winner on our hands - sticker price or no sticker price.
Thanks Richard.
Bookmark:
post to Delicious Digg Reddit Facebook StumbleUpon
Recent on Mstation: music: Vivian Girls, America's Cup, music: Too Young to Fall..., music: Pains of Being Pure At Heart, Berlin Lakes, music: Atarah Valentine, Travel - Copenhagen, House in the Desert
front page / music / software / games / hardware /wetware / guides / books / art / search / travel /rss / podcasts / contact us