Here's what I don't understand...
From the original Analog Industries archive. Some links and images from that era no longer resolve.
Now, I’m no whiz with Linux. I _can_ navigate my way around a typical Linux system, using my OS X chops (ironic, that) but quite frankly I find it to be rather tedious.
But that’s neither here nor there. Here’s what I’m curious about: Ardour, much beloved in the Linux audio circle and the recipient (from my lofty vantage point at least) of a more-or-less unending stream of kudos, is roughly equivalent in both operation and appearance to ProTools 2. Even with the MIDI, natch. Whereas EnergyXT is a far more advanced creature, both better looking and much more feature-laden, and assuming the Linux version and the Windows version are more or less the same, eminently usable. Why is it that I didn’t know EnergyXT was available for Linux until I went to download the damned thing last week?
This is what I don’t get about the Linux community. I mean, I’m not trying to be an asshole here, but if a DAW came out in 2007 for PC or Mac with Ardour’s capabilities (or rather, the utter lack thereof) it wouldn’t even make it out of the starting gate. I don’t mean to offend, but what the fuck is it with you people? Is the fact that something is _FREE_ so important? I hate to say it, but I believe that to be the case. Ardour is free (as in b33r) while EnergyXT costs a whopping $75.
49 comments
EnergyXT has only really been available for Linux since last week. Well, there were pre-release betas of version 2, but they were missing stuff from the final build. V1 didn't work on Linux. So it hasn't been that long for things to catch on. I've seen a few Linux-centric web/blog mentions about it, but nothing big yet.
My experience with Linux people is that they tend to hold all things open source in higher esteem than things not open source, regardless of quality. Maybe money is the issue, but it struck me more as a political thing.
I sense a lot of backlash from linux nerds cause the program isnt open source. Ill take a rad and prolific developer like jorgen over a 100 open source devs, who code when they feel like it. Id like to switch away from windows eventually, so at least on my composition lappy, ext linux may be finding a new home.
I feel like I should qualify my comments slightly. As Steve surmised, I wasn't aware it had been out such a short time, and that would explain for a large part the silence on the subject.
What set me off was, in my morning reading, I came across several posts by people that said they wouldn't buy it simply because it wasn't GPL'd. I just thought that was utterly ridiculous, and speaks to herrprof's point, I think. As I said in my post, no offense to the Ardour people, but it has been in development for years and is still barely more than a glorified tapedeck. (And not even that glorified.) Whereas e-xt, the work-for-commerce of a single person, is perfectly capable, far more capable when compared.
In my experience, and believe me when I say I at least _try_ every operating system, and attempt to make music on all of them, two things are true about Linux:
1) The software generally available (for free) falls somewhere between "really fucking bad" and "unusable." When an OS has been out for a year or two, it is understandable, but Linux is _well_ in to its second decade on this planet. I call bullshit.
2) This thread will no doubt engender responses that both agree with me and violently disagree with me. The latter will take the form of "yeah, but you forgot about Software X, that does Y." Well, do that, and I abso-fucking-lutely guarantee I'll come up with a commercial equivalent on the Windows or OS X platforms that does the job far better. Again, I call bullshit.
This entire thing is odd, inasmuch as people that use Linux are generally _in_ the computer or academic industries, and one would think they'd be quite crafty with the programming and such. Is the OS that hobbled?
-CR
Okay, as a long-time linux administrator (running heavily-used web sites and stuff), it seems like one of the biggest problems is that the OS is designed to be a _server_ OS. Unix in general is a server OS, and its methods of multitasking and handling stuff that requires quick responses is very different from a general-purpose or user-stuff-centered OS...
The thing is, you on your web browser aren't likely to notice hiccups in the data stream from a linux server. Even if you were on the console doing things, having your interactive session swap out for a fraction of a second isn't going to slow you down any, and you'll still be able to type and edit things and so on.
Also, the GUI system used by linux, the X Windowing system, has a weird client-server thing going on behind it because it was designed to totally abstract the screen away from the programs using it. So, you can very easily run X programs on one system and have them display on another, but you don't always get zippy responses like you do from, say, Windows, which can talk directly to the hardware but which is more of a pain for doing remote application stuff. Apple didn't ditch X as the primary graphics system when they moved to Mac OS X just to annoy people-- Quartz is far more responsive since it leaves out all those layers of abstraction between the software and the actual video hardware.
The only reason we're seeing useable real-time audio stuff show up now on linux is that CPUs have gotten fast enough and big memory is cheap enough that things like having your DAW software swap out is less likely to happen and faster if it does, so your audio keeps playing. And CPUs and video hardware are now fast enough (and usually supported through some kind of accelerated X drivers that take care of some of the problems of using generic VESA or VGA drivers to talk to your video card) that you can do reasonable GUIs without making your audio all choppy.
It's not that nobody has been working on this stuff, and it's not that the OS is _hobbled_, it's just that it wasn't designed for any of the things DAW software requires of it. Think of using Windows XP as a web server-- it'll run like crap unless you throw a ton of CPU at it, and even then it'll be pretty iffy compared to using a real server OS.
The stuff that people _have_ been working on for years _is_ good-- it just hasn't been audio.
Thanks for that post. I hadn't really thought about it like that. I mean, Apache and MySQL are certainly fine.
-CR
Yeah, solipstnation, that was totally enlightening. I will never, ever again indulge in wishful, what-if thinking regarding audio and Linux. It has been removed from my equation. Thanks for the house cleaning.
Great post Solipstnation. Very informative.
I'll take the bait.
First of all, a correction. Ardour is free as in speech, not beer.
You can complain about the lack of features or quality in an open source project, but the blame rests solely at your feet. You, me, everyone who reads and comments on this post has 100% total control over Ardour. If there is something about it you don't like, you have power to change it, and you don't even have to be a programmer.
You're doing the right thing. Complaining about open source projects helps them find direction, but it would be more supportive in the form of bug reports and feature requests, and even more so in code patches or financial support.
As a linux nerd, or more aptly, FOSS nerd, I'm not opposed to products like EnergyXT. Software developers need to eat, too. And it certainly isn't about cost. I could afford to buy commercial versions of the software I use. In fact I do buy PalmOS software where FOSS alternatives are not available.
I don't know that a project being free or open automatically makes it better. But in my opinion, it gives the project more potential, potential that I as a user have direct influence over. A closed source, for profit project has more limitations than potential. As a non-free project matures, its direction is governed by the narrow mindset of a singular development team. If the direction differs from your wishes you have no recourse but to abandon that project for another or wait for your ideas to become commercially advantageous. With FOSS, you have a direct conduit to have your ideas entertained and explored by the project's community, and if still your ideas are refused, you can fork the code and start a different version. See XFree86 vs. xorg.
You seem to be focusing on the monetary cost of the product whereas I would assign value to the availability of source materials, freedom to distribute, as well as the monetary cost or lack there of.
If it makes you feel better, you could donate $75 to the Ardour project.
OK. Now i will take the bait.
_aa_,
are you suggesting it would be preferable to donate $75 to the Ardour project and HOPE that it would someday do what the end user wants rather than pay $75 now for something that does what the end user wants NOW?
In my day gig i run 12 linux servers and all is super swell. Going back to the Sun SPARCs i started on, Unix/Linux has always been a great server OS for me.
Once or twice i have thown together a Linux workstation and have immediately reformatted. There is always someone around to tell you that you can do everything with FOSS but can you? My favorite is the "use THE GIMP instead of PHOTOSHOP" tip. Have the people that recommend this ever used PHOTOSHOP. Come on.
To me, it is a little like the electric car/hybrid thing. For years we heard about how we should using alternate transportation, electric cars etc. were coming. As long is you didn't need to actually be anywhere in any kind of reasonable time and as long as you didn't need to travel more than 20 miles and didn't mind looking like a dork, unable to meet the minimum speed requirements to travel on the freeway, you were golden. It took an expensive, mass market vehicle to get it happening for the mainstream.
I really don't want to offend anyone here, and firmly believe everyone is entitled to their own opinion but i think that Linux fans want to believe so that their chosen OS is THE OS that they forgive (or justify) almost anything.
While I agree with your post's sentiment in general, myrnaloy, it kind of misses the mark. If oil suddenly stopped flowing, you could absolutely guarantee that we would have fuel cell vehicles inside a year that performed at or beyond current expectations. The market moves when it has to.
Which brings up another point: a paycheck and a person in charge are much more powerful motivators than groupthink. Bugs get fixed a lot faster when it's "fix this or you're fired." If AD made plugins for donations, we wouldn't make plugins, plain and simple, because I have other skills that can pay the rent and don't have to rely on the largess of musicians. (Being one, I can state for the record that musicians and largess are two concepts with only passing familiarity to each other.)
In short, the 1000 Monkeys Writing Shakespeare model applies. Yes, the 1000 monkeys would _eventually_ write all of Shakespeare's plays. However, Shakespeare wrote them first, and he wrote them because he was paid to. The GIMP is a Photoshop placebo. But the only reason it has something to copy is because someone paid a lot of people a lot of money to invent something that was worth being copied. Ardour sort of does what ProTools used to do back in the day. That's interesting. See above.
Basically, it's a dumb fucking business model. (Or anti-business model, whatever...) And a terrible way to get things done. But all of this seems moot now that solipstnation has given me an epiphany, and made it feel kind of sexy to not care about Linux audio.
-CR
moyashi-- I'm not sure if that was sarcastic...
I forgot to mention that Korg's Oasys is based on Linux, but with loads of custom software on top. Also, audio on MacOS X is useable, but again, by replacing lots of chunks of the basic unix bits and adjusting scheduling. So, audio on linux and unix is possible, and even quite functional, but mostly by ditching a lot of the general stuff and focusing the kernel more on low-latency audio. (Or by ditching the very general GUI and focusing more on a very efficient rendering engine.) At some point it stops being generic unix and turns into something else.
CR,
you completed my admittedly incomplete thoughts. My point was (at least in my pointed head) that money, not good intentions and self satisfaction, drives both innovation and the marketplace.
Being a user of EnergyXT for a while now, I get the impression that Jorgen doesn't give a fuck about the whole philosophy of current Linux user base. It seems more like he is on a mission to get people from the audio community on to Linux, or to at least get people thinking and talking about it. Are there even any producers actually using Ardour for serious audio work, or is it just a toy for Linux geeks to play with?
With the recent developments in Vista and now Leopard, who isn't a little concerned about the future of DAWs on the modern bloated commercial OS? IMO, Linux wasn't even an option for serious audio work until EnergyXT2 came along. It still isn't there yet, but I think Jorgen's intention is to jump start some development on the Linux side so that maybe it will become a viable option sooner rather than later.
Just check out the discussion on the eXT forums. Users are already talking about an email campaign to get their favorite developers to port commercial plugins over to Linux. They figure that there IS a market now and if they generate enough demand it can happen. None of these people are die-hard Linux users with some kind of anti-commercial philosophy, they just see Linux as the future alternative to an expensive, bloated, unfriendly OS.
The people snubbing eXT on Linux probably have never even tried it, have no idea of it's long history of development, and don't realize that there is a good size community of eXT users, some of which are just itching to jump ship over to Linux. They just think, "Ha! The Linux community won't pay for software! Who is going to buy that?" and leave it at that.
I for one just hope that all this Linux attention doesn't detract from eXT development on the Windows side of things because I was really hoping to get away from Cubase in the near future and EnergyXT was my ticket. I have no intention of switching to Linux anytime soon.
Interesting to read solipstnations comments on how Linux isn't made for DAWs. It makes me wonder though, shouldn't a Linux app be able to include a whole lot of the features the major DAWs carry anyway? I have to agree with the poster on Ardour, it's way behind.
I'm a pro musician, pro mixer and pro producer. Besides all the major DAWs I've tried a shitload of minor. While I'm not a Linux user, I have in fact several times looked at that platform for pro audio application but it just isn't there yet. I sometimes force myself to think it is, but it isn't, in fact, it's not even close. In fact, I'm not sure any of the apps can rival Cubase VST5 or Logic 5 that I started using on a PC. Not talking about audio hiccups but pure features.
Another thing I'd like to mention is the lack of two things: proper plug-ins and drivers for high end sound cards. I've heard Ardour can work with VST plug-ins today, but I've also heard it's really buggy.
Sad thing is, I can't see Linux getting anywhere in the audio world the next couple of years if some major player port their DAW, and I don't see that happening at all.
solipstnation, you explained the *nix issues very well, thank you.
I've been working with various flavors of *nix since about 88 and other than for VERY specialized applications have never found it enjoyable as a workstation. Oddly enough the kinda workstation centric flavors like NeXT and IRIX never gained much popularity amongst the unix zealots and the server centrix versions were a major pain as workstations for anything besides specialized apps and network admin work.
My current theory about the cult of FOSS is that it is just that, a cult; a modern substitute for religeon. Some people will obviously praise even the crappiest FOSS software with all the vigor of a missionary trying to convert the savages.
I've always wondered the same thing about Ardour. you have guys cranking out stuff for linux like EnergyXT *relatively* quickly. Heck, given that Tracktion is built on the JUCE framework, I bet Mackie could whip out a Linux version pretty quickly if they wanted to.
But Ardour? They've been at it for YEARS, it's not very mature in terms of features, and it's frankly ugly to look at.
Why is that?
For stater I'd like to say that I know nothing about Linux.
My question is what about Asio drivers for audio interfaces,I don't think there is even one pro or consumer company which offer Linux drivers for its products (for the reasons mentioned above).
so basically is there a way to run an interface with no drivers on Linux?
Isn't this the point where some smartarse reminisces about BEoS? :)
@Tomer:
There's the ALSA (Advanced Linux Sound Architecture) driver in Linux that supports a whole load of cards, pro and consumer. Pretty much the standard I believe so widely supported.
Not that I use it, pretty much in agreement with CR on this topic.
Didn't SSL announce giving a bunch of money to the development of Ardour a few months ago? Have their been fruits from the deal yet?
I used to love Linux back when I worked in IT. For doing server/network security (which is what I did), it was just infinitely easier and better to use than any other system. I even happily used Linux as my generic desktop computer, just because I liked using it.
Once I left IT, and stopped obsessing over computers like only a computer guy can, I kind of gave up on Linux. It's very much something for people who really love computers. Once you just want to make use of computers, but don't really care about them one way or another, moving to a system that is better equppied for the task at hand makes more sense (and Windows and OSX just have way more advanced music software available). I must admit, though, I do miss plaintext config files now that I only ever really use WinXP.
One thing that's always disappointed me about Linux for audio or any other high-level user functions is that so much of what's available is very much designed to imitate something done already on other systems. So much of it seems like an opportunity missed. If you want computers to work differently from the Win/Mac way, why not go whole hog with it and completely define what you want, from the ground up? Invent stuff instead of just making a lot of me-too software. Linux is going to be the underdog for the foreseeable future, why not do it in style?
I can't think of a single audio app that is Linux only that I've felt I was really missing out on no longer using Linux. Which I think is a pity.
solipstnation, I was sincere - no sarcasm in sight.
As an OS X user who gets into Terminal to do things that are best done on the command line, I recognize both the special power of *nix and how OS X, as you crucially noted, is highly mutated away from *nix norms (in this case, all the better). I had ignorantly thought that some day Linux would be the leaner, meaner pro audio champ, and this fuzzy anticipation had clouded my thinking about future working methods.
Three's a crowd; kick out the runt and let the two brutes fight it out in a proper one-on-one.
I tried to make Ardour work for all the reasons anyone would want free software to work, but damn if it isn't the clumsiest thing. It strikes me as a case of too many cooks.
One thing I think you get with most pay-for software is a concept. If you can get your head around and use, say, Ableton Live 5, you'll probably understand and be able to use any new features in Live 6 without a tutorial or readme.
I keep a Linux notebook around for Pure Data, but PD is kind of like camping. It's fun, but I know I'll be tired and dirty at the end of the day. When I actually want to make something useful out of boxes and wires, I opt for Bidule.
I will chime in and also thank solipstnation for his thoughtful response. It is a very valid point that OS X works as an end-user desktop-PC [PC in the original sense, not the specific sense] because it's a variant of Unix that was modified exactly to suit that task.
When I worked for Sun Microsystems for a spell, about 10 years ago, I was amazed at how backwards their end-user-oriented stuff felt compared to either Windows or (pre-Unix) MacOS. It was kind of like the designers and engineers had looked at photos in magazines of contemporary GUI systems and built their own based on those photos without ever having actually used them. It was really kind of pathetic, but hey, they were paying me. True story: while I was working there, some big, glossy posters appeared in one of the conference rooms. These posters lauded the multimedia power of Sun workstations. I was like "huh? my workstation can barely even beep, never mind play an actual sample..." I looked closely at the artwork in the poster because it looked faintly familiar. Sure enough, all of the screenshots of this multimedia wonder were taken from Mac apps. Expressing this discovery out loud to a roomful of engineers and middle-level managers probably didn't win me any points.
So, getting back to my point: My brief experiences with Linux have left me feeling much the same way. It's an OS that was designed to do some things, and it does those things very well, but being a single-user desktop-PC OS was not one of those things. It's getting closer and closer, but it's still behind the big players.
Now, regarding _aa_'s post, two points: first, I have no blame for anyone else's project. That's utter bullshit. Just because I can work on some project doesn't mean that I have any blame for not doing so. That's like me blaming every plug-in buyer for not buying AD plug-ins. Second, potential isn't worth much of anything when it comes to practical issues like actually getting music done or the bills paid. What can you actually do with potential, right here, right now? OTOH right here, right now I can pay for energyXT and write some music. All of the potential in Ardour doesn't add up to one implemented feature in energyXT or anything else.
As an outsider, it seems to me that Linux is approaching an interesting crossroads. Dell has started shipping machines loaded with Linux; that's pretty big news in that a big commercial player has put down a big vote of confidence in Linux. There are very good reasons to consider Linux as an alternative to Windows or OS X because of their own drawbacks. There's a chicken-and-egg problem facing software developers such as AD, however: there's no point in developing commercial software for Linux if its users are not a viable commercial market, but on the other hand the Linux world won't become a viable commercial market until there's a critical mass of useful software that will draw people into that market. Or, putting it more concretely: we can't sell plug-ins until there's enough people using Linux DAWs who are interested in buying plug-ins to make it worth our work, but on the other hand there won't be enough of those potential buyers until Linux has enough pro-music activity to attract those buyers.
It's been fairly well demonstrated that software development and advancement for end-user products--and this is an important distinction--is more successfully accomplished by commercial business entities than open-source collaborations. There are a number of very good open-source projects that are quite good but they are not end-user products, e.g. Apache, SVN, MySQL, stuff like that. But it seems that if you actually want to motivate people to create finished, polished, nicely packaged (as in installers, etc.) and documented products for end-users, you have to organize and pay them. (Even some of the most successful open-source projects are funded, it turns out.) If this weren't the case, then Ardour and the GIMP wouldn't be such a joke compared to their commercial counterparts.
A year or two ago Chris said something to me about developing Linux products and I basically laughed in his face. Today it's starting to look like it may one day be viable. Maybe in another year or two I won't be able to laugh in his face when he brings it up again.
--Adam
"Isn't this the point where some smartarse reminisces about BEoS? :)"
Yes, it is! w00t! BeOS ftw ZOMG!
I was a huge fan. Then JPG stabbed me right in the neck. For what it's worth BeOS is _still_ better at audio than Linux. Don't make me pull out my iOpener and show you.
-CR
Steve, SSL did donate a bunch of money to the Ardour project, but then two things happened:
1) SSL bought Soundscape which gave them all the same tech but without the mass of teenage ADD-inflicted coders
2) Harrison decided to use Ardour as the basis of their consoles.
A lot of people point to the fact that Sony uses Harrison consoles to say "Ardour has made the BIG TIME, Beeeyatch!!!" When in actual fact, the consoles at the Sony dubbing stages are just front-ends for PT and/or Nuendo.
-CR
I think it's interesting, in this discussion about audio and Linux, that no one's mentioned the Muse Research Receptor ( http://www.kvraudio.com/get/1000.html ), which is quite possibly the most high-profile VSTs-on-Linux success story (the Receptor uses a Wine-based DLL wrapper for VSTs and ALSA drivers for it's sound output).
Rather than expecting developers to come out with DSSI and LADSPA versions of their plugins, it would probably be more productive if they sought compatibility with these wrappers, and if the Linux audio guys spent more time developing for the open-source versions of these wrappers (ie. FreeST and dssi_vst, both of which seem to have been abandoned sometime last year).
I'm a long-time UNIX person (I cut my teeth on NeXTSTEP and Irix, and was a big fan of the SGI Indigo Magic desktop, which was, at the time, the best desktop OS I had ever worked with. I loved having an Indy on my desk between 94 and 97), and at least one machine in my house has run Linux since around 98. I have tried really really hard to appreciate the work that's been done in the Linux audio realm. At the end of the day though, if I want to get audio, video or graphics work done, I use my MBP or my XP box, and not the Linux machine.
As someone who was a fan of the original EnergyXT (some of you may recall it was the only way to use VSTis on versions of Ableton Live prior to 3 or 4, I believe), I'm tickled to see a Linux version. I can't say I'll be using it much though, and I even already own a license.
in reply to CR when you said: "If oil suddenly stopped flowing, you could absolutely guarantee that we would have fuel cell vehicles inside a year that performed at or beyond current expectations."
im not trying to change the subject by any means, but couldn't resist throwing in that: if oil did stop flowing, economy would come to a halt and there would be chaos on the streets (especialy in cities) as everything from food production to dilvery to maintenance of the machines that process it all (including the factories and machines that need oil to produce those fuel cells) could not function because of how our socitey is so dependent on oil....
more realisticly what will happen first is that oil wil get so freaking expensive (hundred-fold) that it will be unusable/not profitable... unfortunately depression will follow eitherway
but thats enough doomsday talk ...
back on subject, Linux seems like a developers thing, while most musicians just want a feature-set and stable package (and Linux is definately not leading in that category)
Isn't how linux processes everything in a given install up to the kernel? A kernel which for all intents & purposes is actually still under control of a single person's vision? (Subject to overall advances in the field and Linus's awareness of them and/or willingness to go along with a new idea or two yadda yada yada). So aside from kernel level prioritization of process sets, it seems to me that how much crap is actually running in a given install is up to the distro packager, yea? And of coruse the end user doing the configuration of that distro install. With the advent of 'live cds' and easier configuration tools (Aptitude on the command line & the ubuntu-esque GUI frontends for it) most of the configuration can even be 'steered' by the distro packager. So it seems to me that there's really no reason that a well tuned distro couldn't do an acceptable job for Linux audio (there was a time when the Stanton guys claimed that Final Scratch could ONLY run on linux due to having more control over "their" distro of linux than they had over a given person's XP or OSX install).
As for the relative boringness of the 'mainstream' Linux audio apps like Ardour, it's been my impression that the Linux world is either driven by research (people exploring technology for college degrees & the like) or by people's interests in their spare time.
So on one end in Linux-universe you wind up either with code that IS really advanced, but nearly unusable to someone that isn't a major player in that codebase (on that same project in that same school or whatever. Think csound in the 80s & early 90s, PD & Max (the commercial variant), all the newfangled 'interactive art' constructions seen so frequently on CDM, etc.
Then on the flipside there are projects (Ardour for instance) with the majority of the contributors to the codebase that are primarily interested in older (established) paradigms such as recording their Blues/Funk riffs and some other middle aged pal's Jazz/fusion/funk band, aka the glorified tape deck.
Somewhere in between of course the research end (some of it commercial even) matures enough for others to hook into the codebase & extend it and say 'look now we can run [feature Y] too!'. ALSA & Jack are an examples imo of this. But for the most part aside from the 'getting my hardware to work' driver level stuff, everything is is still fiddly-er than Rewire & VAC (or the recently orphaned Wormhole codebase for you OSX guys).
In summary, Linux is finally there for Uncle Jimmy on bass with his friend from work on the sax (yay?). Then there's also that weirdo up the street with the crazy hair who you suspect has a full chemistry lab tucked away somewhere, doing csound infused meta-block controlled compositions for the local planatarium (linux is great for his experimental DIY noodlings). But not much in between still, you're right... The upshot here is that Uncle Jimmy was running Cakewalk 3 years ago (excuse me Sonar, not that he uses any of the new functionality). So the fact that he can use linux now (albiet at the expense of his experience as a software dev and some spare time) gives hope I guess.
It's interesting to look at how & where Linux has gone over the last few years. I made the switch back when my Windows 95 box got too unreliable, and these days I stick with Linux because it's what I'm used to plus it's got all the apps I use. I do use Ardour, and I do wish it was a bit more smooth. The plugins are generally just embarrassing... good thing I only bother recording acoustic music which needs little treatment.
To me the big difference between the FOSS software and commercial software is the area of coding that's emphasized. Commercial apps tend to start with mock-ups of what the software will look like, and then there's development to actually make it do that stuff. Open-source, on the other hand, tends to start with building the low-level infrastructure and then finally sticking on some sort of user interface. So generally the open-source stuff is very solid in what it does... if you can figure out how to use it. Commercial software tends to look good but often works unreliably because the infrastructure had a lower priority than the user interface.
Another interesting difference is in getting custom stuff done. Back when I was a full-time serious software developer nothing was worse than finding bugs in the commercial tools that might be fixed on the next upgrade cycle, six months from now. Meanwhile I had a customer who needed something within a couple days or their warehouse would have to shut down (in retrospect that's probably about 75% of why I burned out on coding). In contrast, a handful of times I was able to use an open-source solution and could directly contact the developer for changes. That saved one particular project, and the developer was happy that somebody was providing useful feedback about features.
Most of the apps I use daily are now available for Mac and even Windows now, so choice of platform isn't quite as big a deal. I'll still use Linux mainly because I don't have to pay Microsoft a couple hundred bucks, then pay more for antivirus software because Windows is still pretty weak on security.
Now, let me ponder this thought out loud:
From the various learned posts in this thread, one could make the assumption that a lot of audio tools for Linux are mainly the result of people that don't know any better trying to make their operating system do something that it isn't very good at.
Yet, you have something like EnergyXT get ported over, and it's a for-sale program. One can use Windows VSTs with it, with Wine, but it can also use built-for-Linux VSTs, I gather from my research.
So, here's a thought: IF there was an easy-to-use and relatively comprehensive DAW, and IF there was a reasonable selection of plugins, and IF the only thing you needed to use the computer for was music, AND you wanted to do the whole thing on the cheap, wouldn't it logically follow that IF a company was to man up and make a product (e.g. Jorgen porting his DAW) the market would follow?
I mean, if all you need is a simple DAW and a 'verb and a delay or whatever, you're looking at $200 and you can slap it on any old computer, and the OS is free? (As in "free to gouge out your own eyes with a rusty spoon," I'll grant.)
-CR
Well, what's emerging here seems to be a "fact" that Linux-Audio - much like embedded Linux for gadgets - has only worked, and will only work, when a for-profit company takes the bull by the horns and sculpts a distro to meet the specific needs of their product, closing the distro off from the public and keeping tight control over it. That makes a lot of sense if the company wants to save money (and what company doesn't?) and can leverage the know-how. This is what I've been waiting for: a commercial, professional-oriented, semi-customizable, turnkey solution comprised of an excellent DAW, an excellent interface, a low-to-moderately priced computer, all knit together with a lean-n-mean, tightly controlled, audio-optimized (proprietary) Linux distro. It seemed like a no-brainer to me, but since it hasn't happened yet I was clearly mistaken, and as far as I'm concerned, solipstnation has pretty much explained why.
>From the various learned posts in this thread, one could make >the assumption that a lot of audio tools for Linux are mainly the >result of people that don't know any better trying to make their >operating system do something that it isn't very good at.
I wouldn't put it that way. I'd say that a lot of audio tools for Linux are mainly the result of people who know what they want to do at the low level - perhaps a really interesting filter model or something. But they often do it to satisfy their own need initially, so the user interface is something primitive but which they created just to solve their problem. So they know what they're doing, they just don't often get around to making it easy to use by other people, or making it a real-time plugin vs. a standalone to process WAV files or whatever.
Here's the crux of it, though. Most all successful companies have one thing in common: someone like me. I couldn't give two shits and a rat's ass whether the interior of our product is crafty and whether we had to invent something, or there was some information available that we could learn from or whatever. My main concerns are two-fold: (1) How does the product look and feel, and (2) does it do what I want it to with the least possible drama?
This no doubt drives Adam up the wall, when I'm nit-picking over the range and such of the knobs, or when I ask him to add distortion to a block because it's too clean, or moving things a pixel up/down/left/right. But those are the reasons people _buy_ a product when there is a similar one available for free. One wouldn't expect Replicant and dBlue Glitch to be able to exist in the same market, when one costs almost $50 and the other is free (as in b33r), and they both do similar things. That just doesn't make sense. Yet Replicant is far and away our fastest growing product, and is about to over-take Dubstation as our best-selling. So there you go.
As Mies Van Der Rohe so eloquently put it, God is in the mother-fuckin' details. (I'm pretty sure he put it just like that.) The Linux world has too many cooks, and that isn't necessarily a bad thing, but there's just nobody running the fucking restaurant.
-CR
Chris Randall wrote:
>The Linux world has too many cooks, and that isn't necessarily a bad >thing, but there's just nobody running the fucking restaurant.
I think thats a fundamental misunderstanding about Linux. There -is- no restaurant. What you get is a potluck with all those cooks bringing stuff to the table for nothing, and you get to pick and choose what you want, but none of them feel any obligation to make sure there's something for the vegetarians, something for the lactose intolerate, something for the diabetics. They're all busy crafting their own specialities, and it almost doesnt matter to them whether what results is even a balanced meal.
At the end of the day, things tend to balance out, and most of the bases get covered, but its a happenstance process because they're not out to -serve- you meals you might prefer, just make the best examples of the meals they enjoy making.
To get away from that stretched analogy, at the end of the day, Linux is developed by and for a series of overlapping special-interest (developer) groups, with the assumption that anyone who wants something not provided will build it themselves just like what they done. If you're in a special interest (user) group that isn't served, nobody's going to jump to sort that out for you; you're supposed to assume the mantle of developer, wait in case someone else does it, or forget about it.
Ive been using Linux since about RedHat2, and Im tired of hearing that GIMP replaces Photoshop (or that Blender replaces Maya for that matter). I dont actually bother with Linux any more, the gains were outweighed by the hoop-jumping required to obtain them.
However Jorgen's effort on getting energyXT2 onto Linux is indeed admirable, and incredibly altruistic, but the Linux converts he'll predominantly get will be new ones without the 'if its not free we dont use it' baggage. And unfortunately, unless that 'market' of new converts is large, plugin developers are just going to run into that same 'commercial software is baaaad' barrier. Freeware and shareware might subsist there, but I dont yet foresee a point where it'll be worth the effort for full-time commercial devs. Even now, people will be saying 'why should we pay for eXT2 when we can just run Rosegarden and Csound via PD with JACK, and the previous version of ALSA drivers with a customised low-latency kernel from this Finnish teenager as long as you hack the config file a bit, and thats -just- as good -and- free'.
By the way, the main developer behind Ardour, Paul Davis, has spent an awful lot of time and effort trying to get some semblance of 'proper' VST support for Linux off the ground by getting Steinberg to properly open their 'open standard'. If things get any more funky with the moving-standard that is the VST spec, his stance might be a bit more understandable on that one.
I just remembered something screamingly obvious. I'm pretty damn good with unix but even if I could get something decent for free, I'd rather have something good that I paid for. Making music is enough work without some buzz killer of fiddling around with kernel tunings etc. Aint nothin ruins the studio mood like equipment failures and I count fiddly software as a failure.
On the positive side, its nice to see other people who remember csound or who also learned unix on a NeXT.
You know, I'm amazed that nobody has mentioned all the good research that folks like CCRMA, IRCAM, and CNMAT are doing with Linux. While it's not commercial work per se, these three organizations are *the shit* when it comes to new audio technology. They've been using various unices and Linux for quite some time. I remember when the first inverse FFT additive synth was done on IRIX by Adrian Freed and Rodet (open-source IIRC), which provided a basis for a lot of new additive softsynths down the road. And never mind the libsamplerate code that's EVERYWHERE, and that wouldn't be if Julius Smith hadn't open-sourced it.
I remember hanging out in #musicdsp on IRC a ways back, and hearing commercial plugin developers scouring open-source code for little nuggets of wisdom to toss into their kit. I don't know for sure, but I would bet even some of the AD code could be based on stuff like filter models from Robert Bristow-Johnson's biquad filter cookbook, libsamplerate, or any of the code generated by the wealth of papers published by Julius Smith (physical modeling, filter modeling, sample rate conversion, virtual analog synthesis, band-limited impulse trains, etc.)
While linux isn't usable by itself per se, it's provided the foundation for a lot of code bits that have driven development of the marketplace. And since 99% of this code is open-source, you won't see it on anyone's list of software credits unless they're feeling VERY altruistic.
This whole discussion makes me realize that I'm not like most of the people who read this blog. Seriously, I understand almost nothing of what has been written. As my friend Bob says, it's all "Swahili on shortwave."
I am SO not a computer person, I'm a musician.
Downpressor and beauty pill nail a big part of why Linux isn't ready for prime time: musicians--and end-users in general--are not computer persons. They just want to get their job/music/word-processing/whatever done. They have no inherent geek fascination with the computer itself, they just want it to get out of their way and work right. Linux, at least in my limited perception, requires way too much futzing around with the computer itself before it gets out of your way. Just because you _can_ recompile the kernel to suit your needs doesn't mean that most people _want_ to do so. Now, tuned and polished distros like Ubuntu go a long way to ameliorate this, but for musicians it's still far easier to get a Windows or OS X machine together and running smoothly, and there are far more choices for host apps, plug-ins, etc.
--Adam
I'm back.
I would like to explain why linux may be the ideal operating system to develop an audio application on.
First let me explain the shortcomings. There isn't much in the line of APIs or standards for audio development on linux. There are some, like ALSA, jack, umm.. the KDE audio system that slips my mind presently. These are/were all good solid projects. Unfortunately, having a half dozen or so projects all seeking the same goal causes lack of focus in the community.
A development environment that provides total freedom doesn't always result in ideal situations. That is the price you pay for no patents, no copyrights, no proprietary IP. You don't get ASIO, DirectX, and VSTi.
What you do get is powerful kernel that you have complete power over. Maybe that's not important to some of you. But if you're going to build a program that is revolutionary, if you want to get down into the nuts and bolts of audio development and start from scratch, linux is a great place to start.
Someone had mentioned that Linus has complete control over the linux kernel. That is not true. If you disagree with the direction he is taking the kernel, then you have every right to fork the code and start your own kernel. People do it all the time. Again, see XFree86 vs. xorg for an example of a major FOSS project that was forked.
I'm going to make a musical analogy, bear with me.
In response to myrnaloy's comment, I suspect that most of you are typical musicians. You go to a brick and mortar music shop and buy a factory built Fender or Korg. That's fine. You want a guitar and you want it now.
I see most linux users as being the craftsmen that would rather build their guitar in their basement workshop than buy one off the rack. Maybe it turns out like shit, but they built it and that makes it important to them. I'm sure you've all built a rubber band banjo at some point in your life. Of course it's not a Stratocaster, but that's not the point.
But that's where discoveries are made, in the basement workshop. The circuit benders who take a mass produced instrument and say "screw that", crack it open and make it their own. You can't crack open closed source software and make it your own.
My opinion; you can't buy revolution off the rack. We need linux nerds to seek different directions, even if they fail.
Maybe you guys can't see the potential in linux, but I can. And it's perfectly fine with me if you don't want to use FOSS, but you shouldn't rain on my parade and disrespect the efforts of people who volunteer to make these tools available to me and you for free. I'm not saying anyone has been disrespectful, but I'd like to see people who try Ardour or any FOSS project and not like it, submit their complaints in the form of bug reports and feature requests through the appropriate channels instead of saying, "Project X sucks because of this and this." That's the only way Project X is going to get any better.
Personally, I use linux (and Audacity) because I made a personal vow to stop pirating software and music several years ago. I've ran into road blocks but with trial and error and research and effort I find ways around them and have learned to use linux audio tools very effectively.
Anyway, it saddens me that the consensus of this conversation has become, "audio will never work on linux." Audio will never work on linux because people like yourselves are unwilling to support it. And yes, studio nebula, you are to blame. If you were made CEO of Microsoft today, what changes would you make? Would you do nothing?
Everyone who commented on this post is the CEO of Ardour. You have the power to change every and any aspect of it at your whim. If you don't think that's true, by all means, tell me what is stopping you. Not a developer? Most CEOs aren't.
Anyway, good conversation. Thanks.
It's extremely unlikely that I'll be made the CEO of Microsoft, so speculating on what changes I'd make if I were in that position is a waste of time. I am, however, the CEO (or rather the President) of Audio Damage, Inc. As such, I get paid to write audio software. I do this in part because I love writing audio software, and have loved it ever since I started doing so over 30 years ago. However, love, like potential, doesn't pay the mortgage or buy groceries or pay the electricity bill to keep the computers running. I need to get paid for what I do. If I can't get paid to write audio software, then I have to get paid for something else. If I'm getting paid for something else, odds are very high that I won't have the time and energy left over to work on a not-for-profit project like Ardour. Believe me, I speak from experience; I spent most of those last 30 years doing music software in my (increasingly limited) spare time while I held down a "real job" during the day.
Since you asked, here's what's stopping me from working on Ardour: there are only so many hours in the day. I work full-time on Audio Damage's products. You think I'm going to spend my spare time working on Ardour? Hah. Besides the obvious danger of burning out altogether on programming (and yes, I've done that also), I have, like, a life. I have a house to maintain, a significant other to interact with, a lawn to mow, bills to pay, cat boxes to scoop, etc. etc. etc. In what little spare time I end up with, I usually try to not do anything even remotely like programming. It's even dicey for me to use AD products when I write music because doing so can make me think of work, and then poof! out the window goes my creative flow. You think I'd rather work on Ardour than use a commercial off-the-shelf product like Cubase to write music in my spare time? Why, pray tell, should I? For potential? For the advancement of the greater good? When do I get to stop advancing the greater good and just write some music, which is what I actually wanted to do?
I never said that I wasn't willing to support Linux. If you re-read my posts you'll see that I said it was a possibility in the future. We even have people using the Windows versions of our products on Linux right now, and we do what we can to help 'em. Today, however, it's not commercially viable for us to officially support Linux. Supporting other platforms is extremely expensive for a small company, and there is no evidence that selling Linux products would pay for itself. We're not just picking on Linux; the same is true of DirectX and RTAS. (Actually RTAS is a more complicated story, but during the time we did support it it certainly was a net loss for us.)
So, like I said, it's something we'll revisit in the future.
--Adam
_aa_ you more or less supported my point though, that much of what exists on linux for non-server/web/database type activities is either of the mad-scientist variety or the coder-in-his-sparetime variety. There's absolutely nothing wrong with this but for a non-mad-scientist-and-non-coder who just wants to write a tune or two it's just not there yet. Also I didn't get the sense that the prevailing opinion was that it will never be there yet, the fact that we're even discussing this now means that it's edging closer. At the same time for typical 'mainstream' tasks it still tends to lag a few years behind. And since I can't even get someone to take a look at the issues Synergy (keybd/mouse app I use a lot) has shown on my dual xeons since 2003, I wouldn't hold my breath for it to be patched up for the mainstream use of multi-core cpu's. Meaning that while for some projects feedback is listened to, for many others you wind up with either no clear heirarchy to prioritize external requests & bug reports or simply the fact that many end users haven't a clue how to move around on Sourceforge.
And as for Linus exerting his control, if someone were to fork the kernel under the gpl would it still be linux? Not to mention that his influence even extends beyond Linux to the OSS community in general (for instance the draft revisions of the GPLv3).
To end my contributions on a positive note, I'd agree with you. The upshot of Linux is that if you DO want to get into experimental audio biz you'll probably find a lot more available code & tools to fiddle with on Linux than on the more commercial OS's. Open source tends to work that way =]
_aa_,
I came to the opposite conclusion, that there are many computer literate people here, alot of them with good unix backgrounds who choose not to use linux for music because they want to actually get the music done. As far as I can see, most of the regular readers and posters here are interested in doing music much more than doing computers even though probably all of us use computers as part of making music.
Valis,
I think that musicians are by nature mad scientists. I think that tweaking, customizing, and ultimately designing from scratch one's instrument is a natural progression for a musician. For a computer musician, his/her instrument is the software.
If you did a true fork of the linux, I believe that legally it would have to be named something else, Valix for instance. Alternatively you could simply supply and maintain a patch to existing kernels, which is what most people do.
Downpressor,
I'm sure most pianists don't know the first thing about piano repair, but I bet most piano repair people are excellent pianists. I think understanding and working with the inner workings of your instrument is important to developing musical maturity, and definitely crucial to developing individuality on any given instrument.
If you apply that logic to computer music, it means a computer musician will benefit from tinkering with source code and the inner workings of his/her computer.
I don't discount the importance of an instrument simply being able to work out of the box, but I think there are levels beyond that that are important to explore.
studio nebula,
When I said "support linux" I didn't mean your products specifically should operate on linux, though I would encourage that. I meant support it through bug reports, donations, etc.
Money is a problem. Of course you can technically sell GPL'd products for money, but you're also required to make sure the consumer is aware that it is available for free. This is sort of shooting yourself in the foot, but one of the goals of FOSS software is making it available to all regardless of wealth. This is why the bulk of income for FOSS comes from donations, both corporate and individual.
I for one have always been a proponent of compensating developers in FOSS projects first and foremost. I'm strongly opposed to campaigns like SpreadFirefox that I view as taking money that should be going to developers and throwing it away on fruitless advertising campaigns.
Just like other not-for-profit organizations, open source projects pay a lot of their employees, developers included. Volunteering is crucial, but well supported projects can afford to hire extremely talented developers to work on the project full-time. It's entirely possible to work full-time on an open source project and still put bread on the table.
I don't see how Ardour is stopping you from writing music. Perhaps it's not as easy as Cubase but there's obviously many people out there that think Ardour is adequate and can be creative with it. You decided that you would sooner pay for Cubase than spend time learning, configuring and using Ardour. Since time is obviously very important to you, I think that's perfectly reasonable. I on the other hand don't mind spending a few hours reading a man page or wiki. Money is more valuable to me than time. Personal preference I guess.
Of course you still have to read Cubase's manual and help file. Is the learning curve for Cubase really so much shorter than that of Ardour, or is it instead that you're already familiar with Cubase?
I've been using linux and FOSS exclusively for about 7 years now. I don't use Ardour, I use Audacity, but it would not be a big deal for me to apt-get install ardour and figure it out in a matter of hours. For someone not accustomed to linux environment, that might be much more difficult. But I would contend, that it would be equally difficult for someone like myself who has never so much as touched a Mac to use Protools.
It's easy to try Ardour for 20 minutes, decide you hate it and delete it because it didn't cost you anything. If you buy a copy of Cubase, try it for 20 minutes and hate it, you're not going to delete it because you have something invested in it and what retailer is going to take back open software? You'll keep working with it until you figure it out and after weeks of working with it, you learn to love it.
I think we should all be giving the same consideration to Ardour and other FOSS projects.
That's an awful lot of typing to base on a false premise. Musicians are not, by nature, "mad scientists." Mad scientists can occasionally be musicians, but the other way around isn't necessary or even that common. Since most of your following text is based on that argument, which is demonstrably false, and easily so (I wouldn't have a job if it wasn't) I don't think I need to address any of it.
-CR
I use Cubase because it's familiar; I've been using it for years. I suspect that, for a user new to both, the learning curve for Ardour is shorter than that of Cubase since Ardour is a smaller and simpler program.
From the perspective of a musician, there's no reason for me to try Ardour. The tools I have work, and I have yet to see anything about Ardour that presents a compelling case for abandoning the tools I use in favor of it. Yes, Ardour and its host OS are free, but the cost of the non-FOSS tools I use is not prohibitive to me.
--Adam
Chris Randall,
You seem to be suggesting that your clientèle are incapable of experimentation. I'm certain that's not the case, so feel free to rephrase. I was using Valis' term, but if you're going to disregard my entire argument because of it, then I must insist upon an explanation.
Are you suggesting that there is no science involved in music? No experimentation? You buy a keyboard and use nothing but factory presets? Are you and your clients incapable of innovation?
I believe you're suggesting that an individual who can sight read Rachmaninoff but couldn't compose Mary Had a Little Lamb is a musician. Technically true, and perhaps I should have used the word "composer" to be more clear, but I think you are wrong that musicians are not mad scientists.
A mad scientist would be an individual who studies a science, and then proceeds to perform radical unorthodox experiments, do you disagree? Are you telling me you never push the envelope, never take your equipment to the limit and wish it could go further?
All the musical instructors I've had in my life have always encouraged me to experiment. Perhaps you were trained in a marching band where the focus is uniformity as opposed to exploration. Hell man, even Sousa invented a fucked up tuba.
If you're a musician, and you're not a mad scientist, then you're missing the big picture, in my opinion. Who the hell gives a crap about an painter that can only make perfect duplicates of the Mona Lisa?
I think I'm starting to see a pattern developing here. Why do you hate freedom, Chris? Just kidding. But seriously, explain the mad scientist jab.
Thank Jeebus for jazz.
Man, you really have to stop dealing in absolutes. I don't recall saying "EVERY MUSICIAN USES GARAGE BAND TO DO PINK FLOYD COVERS SO SHUT THE FUCK UP," but that seems to be what you drew from it. Experimentation, musically, doesn't hardly ever equal rebuilding your kernel. It equals finding uses for discordant note combinations, or trying to find a way to fit a bouzouki solo in to a dancehall track. You could either spend a week building a bouzouki sound in CSound, or you could just get a fucking bouzouki off ebay. The same net effect, ultimately.
Ultimately, this conversation is spiraling off in to the land of Shit I Don't Care About, so you can save yourself the trouble of typing a reply. The bonus of having this site is that I always get to be right.
-CR
...uh, _aa_....as a composer who graduated from the New England Conservatory, where there were a great many people who could sight-read Rachmaninoff, I hate to break it to ya, but essentially none of them had any interest in composing, and approached the activity only under duress. Also, you seem to be denigrating Mary Had a Little Lamb. Sure, it's a simple child's song - but one whose life has spanned centuries and that millions of people know. Has your kernel tweaking helped you match that recently?
I respect your passions _aa_, and certainly share some of them, but dude, "different strokes for different folks" ain't just a cute phrase, it's bloody reality. Giordano Bruno, burned at the stake in Rome for heresy, had a theory called something like "congruence of opposites," in which opposing ideas inhabited a circle, like beads on a chain. At 180 degrees, dead opposite on the circle, the ideas were said to be completely opposed, in complete disagreement, perfectly different. As they became more extreme, moving farther to the left or right, they eventually meet at the back of the circle, in other words becoming essentially the same thing. The classic example is fascism and communism both resulting in totalitarianism. You sir, by absolutely mandating your idea of creative experimentation, are in fact demanding total(itarian) conformity that is diametrically opposed to what you purport to stand for: freedom of thought and action.
So lighten up.
aa wrote>
>Money is a problem. Of course you can technically sell GPL'd >products >for money, but you're also required to make sure the >consumer is >aware that it is available for free.
Incorrect. You radically misunderstand the 'free' in free software. You can sell GPLd software, but the requirement is that you make the -source code- available. Money is not even an issue; the GPL even stipulates that you can charge media costs for the source code.
In your supposed 7 years of using Linux and FOSS you really should have learned the meaning of the phrase 'free as in speech, not as in beer' by now.
>This is sort of shooting yourself in the foot, but one of the goals of >FOSS software is making it available to all regardless of wealth.
Incorrect. The philosophies of FOSS have -nothing- to do with wealth.
>This is why the bulk of income for FOSS comes from donations, >both corporate and individual.
Incorrect. The 'this' is based on a provably false supposition.
Comments are closed on this post.