Why Open Source? Why Free (as in speech) Software?

Some people who come across Ardour and start reading about its "odd" license (the GPL, used by thousands or tens of thousands of software applications and libraries around the world) end up asking "what is with this whole open source/free software thing anyway?" or "why don't the Ardour developers just release Ardour like a normal commercial, proprietary program?".

It isn't a complete answer to this kind of question, but as exhibit 13.1(a) part 8, I offer you:

http://www.zdnet.com/users-petition-avid-to-sell-sibelius-music-software...

Summary: Avid buys Sibelius, which brings in a nice $18M or so a year, then appears to layoff all of its developers, mumbling about how they plan to integrate its development within their California development team. Sibelius users are extremely unhappy and attempt to convince Avid to sell the software.

Now, its true that had Sibelius been an open source program, it would probably have never been generating $18M of cash flow/revenue. But its also true that its users would not face this kind of situation. The company would probably not have been acquired, and if it had, its developers would have been free to continue working on their own fork after being laid off by the purchaser.

Rest assured, dear reader, that this is not a fate that can befall Ardour. There are other issues that the project faces, but this is possibly one of the most serious future usage threats to the user of any piece of software, and this cannot happen to Ardour. Because of the GPL. Good night and good luck to Sibelius development team and its users.

The same logic applies to

The same logic applies to using Ardour with proprietary plugins, no?

The same logic applies to

The same logic applies to using Ardour with proprietary plugins, no?

I wonder if the same applies to any project once it passes a certain level of complexity or requires specialist knowledge. There are a lot of orphaned open-source projects out there which are either completely incomprehensible to anyone but the original developer, or simply lack anyone with the relevant skills to take them on. Ardour is among the best in terms of being a well structured, professionally coded application - so in 'theory' if the original developers quit for some reason, it could be picked up and maintained, but there aren't that many (read hardly any) developers with the necessary combination of skills (and I think most are already involved).

The same is true of plugins - proprietary or not - If one of the core libraries on which x / y / z plugin depends stops being maintained by the one developer who understands how that complex DSP code actually works, then to the average user the result is the same - they can't use it anymore.

It could be argued that in most cases a commercial developer with a good business plan will have a commercial incentive to keep a product working across future OS / host / API changes, and also some obligations to existing users, whereas perhaps a single developer / student who releases his work for free, in his spare time, has no obligation to fix anything and may just lose interest. So I think as usual there's no clear cut 'one is better than the other' answer - just a different set of advantages / disadvantages - but at least you have the choice to go with commercial or open-source software.

It's not like this kind of

It's not like this kind of thing never happens. Anyone else here remember Opcode's Studio Vision ? How about Tascam's Gigastudio or Voyetra's music software ?

The company owns it, they can do what they want with it. C'est finis.

Best,

dp

Correct, abandoned open

Correct, abandoned open source projects do not automagically come back alive.

But the number of projects which got picked up after the original maintainer and all of his/her team left is >0.
The number or proprietary cases where an abandoned project was picked up by someone with no connection to the original author is by definition 0.

Thanks for that Paul, though

Thanks for that Paul, though I'm sure there are probably some people still shaking their heads in disbelief. ;-) Your project is solid proof that the terms "high qualify" and "open source" are not mutually exclusive (Ardour isn't the only one, of course, there are others like Inkscape and Scribus to name only a few more). :-)

I wonder how many Sibelius users know about Lilypond and frontends to same? ;-)

> abandoned open source

> abandoned open source projects do not automagically come back alive.

Sometimes they do. At a rate of 0,01%, though :) LAoE audio editor would be one of them.

I think something that often

I think something that often gets overlooked is the importance of open data formats and APIs etc rather than just a strict open or closed approach to software. As a user as well as a developer, I actually care more that my data - be that audio or text or a DAW session, or whatever I create is not inherently encrypted by being stored in a form only readable by one (closed) application. If the format is known, at least I can take my data and use it with another (open or closed) application, if the developers of the original either go out of business / lose interest etc. My own software is commercial - I can't necessarily give away a complete built, tested implementation of ideas which I have spent my own time and money developing. That would not make commercial sense to me (or other investors). But the technical specifications are published and there is a wealth of detail provided about the theory of how it works. But, as I said in my previous comment, the choice is there - at least on linux people can use open source or commercial applications, depending upon whatever suits their needs / politics

@linuxdsp: This is our

@linuxdsp:

This is our approach with Mixbus as well. We don't give away the plugin source code, but the user's data (in the session) is clearly preserved, and the fact that we use Ardour's session format makes Mixbus sessions much more durable than a proprietary storage format. I don't understand why "pros" are using apps with proprietary and even (in some cases) encrypted/obfuscated session files. Someday the owners of the material are going to ask them to re-purpose these old works, and it might be hard or impossible to open a session file from a long-dead workstation company. The execs might not understand why it costs a lot to reuse something that they supposedly already own.

Anyone remember what happened

Anyone remember what happened when Apple bought Logic? It's estimated that they kicked out around 60 000 Windows users.

@BenLoftis and @linuxdsp:

I'm a big fan of open source - both for several well known ethical and long term reasons, but realize that it can be very difficult to make a living by making open source programs. So I'm pragmatic and shamelessly buy and enjoy closed source programs and plugins - and Mixbus and the linuxDSP stuff are not difficult to enjoy - I'm probably one of your biggest fan boys. :-)

But i'm a much bigger fan of open standards and transparency - which I see as much more important thing than making program open sourced. Then it will for example work like this: If you don't like a car, then use another one, you can still drive on the road with anything you want because the road is standardized and everyone are more or less following the same rules when using it. I know that you know what I'm talking about here, but I'm saying this for all the people that are not aware of this Open\s.+ things (not that they will read this anyway...)

I was recently bitten by a mistake I did: I was stupid enough to buy a Presonus Faderport and it is a true mess when using it on Ardour and Mixbus (and probably anything else without special drivers for it). Some buttons work partially (MIDI CCs), and the fader and pan control (which probably have a proprietary protocol) doesn't work at all. And it seems to me that no one in the whole word can give a straight answer on how til fix it and make it work in a decent manner for Ardour/Mixbus - that is if somebody even bores to answer - I think Presonus is probably answering with a massive silence on this matter.

So I have my Presonus visible on a shelf, reminding me that I shall never ever buy a gizmo or gadget without making sure that it actually works together with my excising stuff. Hoping that a thing someday it will work should is plain stubidness and i did it again...

@josander: The fader/pan

@josander: The fader/pan control knobs probably send 14 bit NRPNs, rather than some proprietary protocol. Its fairly easy to figure this out with any MIDI monitor app.

@paul: Ardour/Mixbus'

@paul: Ardour/Mixbus' autolearn functions doesn't detect the fader/pan control controls.

Is usage of 14 bit NRPN something Ardour/Mixbus will or might implement? Is there something I can do apart from sniffing MIDI data? I'm willing to ship my Faderport (and also pay for the return of it) so somebody that is skilled and can figure it out and implement it or I will be happy to buy/donate one to Ardour, but can't afford it at the moment (maybe in two three months or so).

EDIT: I think the following link have a lot of useful information for Faderport support, but I do probably not have the skills to write a patch that can be sent to the Ardour project:

http://forum.renoise.com/index.php?/topic/26515-presonus-faderport-implementation-27/

..and this is probably all the technical information that is needed i an insanely great documentation the maker of Renoise's Faderport driver maker did:

http://airmann-faderport-driver.googlecode.com/files/faderport_manual_1_0.pdf

While the fader does send 14

While the fader does send 14 bit, the pan controls sends an increment and decrement command only. Going off memory both of these are supported in A3 with the MIDI protocol, but I would need to look it up to be sure.

Seablade

I am very grateful for this

I am very grateful for this product being opensource, it means I can be safe in the knowledge that no matter what happens to the project in the long term, my sessions will always be accessible.

@christophski: my sessions

@christophski:

my sessions will always be accessible.

I think that is as much because the format ardour uses to store your sessions is an open format as it is to do with the project itself being opensource. I think open standards are just as (or perhaps more) important than the open vs closed nature of different applications.

i agree with josander, i've

i agree with josander, i've found in Ardour, Mixbus and LinuxDSP i very confortable corner where i can "hide from the system" and im gratefull for that, the "horizontal structure" of this comunity is a big plus so even when i hope and wish you all devs can keep on dedicated to this and get better and bigger incomes from an eventual user base increase... i also hope it keeps being this horizontal.

so.... Long Live "OUR DAW"!!

@linuxdsp: I totally agree. I

@linuxdsp: I totally agree. I have a folder of Protools sessions which I might as well delete as I do not have protools and am not willing to spend that much money to save the sessions. Thankfully I have been able to convert them at university, but that kind of thing would not happen had they used an open standard. It seems to me that open source and open standards tend to go in hand though, so I am always more likely to pick the open source choice, not just because I advocate what it stands for but also because it has saved my arse more times than I can mention.

I have 20Gigs of audio in a

I have 20Gigs of audio in a single Apple's Logic Session that i wont be able ever ever again to open, i have to rely on a friend to do a stem export or bounce the most edited channels so i can pick it up from it....

Not happy... not happy about it... not at all.

@fernesto : Could

@fernesto :

Could AATranslator help ? I'm only guessing, I know nothing about Logic or its file formats, but I see that the converter does support OpenTL and lists Logic as an import/export target.

http://www.aatranslator.com.au/

Best,

dp

I think that would certainly

I think that would certainly help, sadly my country has a control over the amount of money we can spent over the internet to other countries... i reached my limit this year so i'm just using whats left for Ardour's suscription.

I could buy the software next year or maybe succeed opening an account on the US this year since im traveling on november for a week to miami, going to buy a couple of hardware pieces, anyone up for a beer? heh...

AAF and OMG support will be very important also for my work so i am very interested in aatranslator or ardourxchange... yet to see what works best for me.

Best.

Felix

I tried Sibelius a while back

I tried Sibelius a while back and I thought it was quite poor. It might be alright if you are concentrating on purely printing scores but for composing with the ability to write a score, I think Logic was best.

As for Avid. I don't like what they do. I don't like being locked into certain formats and the hardware (although that end of things is better now). I wish everyone could be using open source or free software. However, Avid have absolutely no competion with Protools. It is without question the fullest featured while being easy to use DAW. This is why they have a monopoly. Logic is probably next best. I would use Protools HD in a pro environment if I was making money to cover the cost of it even although I'd have a gripe at being forced to do it.

However, Avid have absolutely

However, Avid have absolutely no competion with Protools. It is without question the fullest featured while being easy to use DAW. This is why they have a monopoly. Logic is probably next best. I would use Protools HD in a pro environment if I was making money to cover the cost of it even although I'd have a gripe at being forced to do it.

I would strongly disagree with this, and I think it depends strongly on what you are doing. For example composers might enjoy Logic more, Film Scorers for a time tended to prefer DP, etc.

The reason Avid has a stronghold with ProTools has more to do with the fact that ProTools is everywhere from when they were without much competition. That means I can start a PT session recording in one studio, walk down the street to another with a better mix room, finish it there, and send it to a mastering house in a third. The other options have not reached this prevalence yet, and it is in part because they don't have this prevalence so they are in a catch-22 type of situation there, and Protools continues to dominate as a result.

Seablade

I noticed a slowdown in A3

I noticed a slowdown in A3 development for 3 months... I hope the stable version will be released this year.

Don't worry about it. Svn

Don't worry about it. Svn still get updated :).
I regulary do an svn update and there are always updates, but not always showed in issue tracker i think.
It' were holidays also, for me it's normal it was little bit slower.

Its mostly burnout on my

Its mostly burnout on my part.

Can one endeavour for a near

Can one endeavour for a near perfect (bugless) version release?
The logical answer is probably "yes", but the real world answer is probably "no".

Mmn... i don't think that

Mmn... i don't think that reaching a bugless release would ever be the ultimate goal as every DAW is always evolving, that bugless release would eventually have to be replaced by a better version, however probably buggy release .... what i think could be a main goal regarding to buggs is to ensure every release is free of "freeze - session corrupt - frustrating hours of work wasting" buggs, and even that is aiming high maybe? Devs are working hard on that, i know it.

I haven't experienced that type of bugs from either Ardour 3 or Mixbus 2 for a long time now and thats great, and every time there is not much "news" on Ardour's site i know something good is being cooked for A3 or for Mixbus and sincerely i cannot wait to see what brings the future of A3 to a eventually A3 based Mixbus... and i know its not going to be soon but still....

I have a humble suggestion

I have a humble suggestion for Paul and the other developers of Ardour. Before I impart my idea, I would like to say I'm really happy that Ardour exists and is pioneering open standards and good software practices in the music world. However, progress on the project is very slow due to there being limited resources available. There are many features in proprietary music programs which are not yet implemented in Ardour 3, and so it is hard to use it for production-quality work (in my limited experience with Ardour 3).

My suggestion is this: create a usable audio software bundle that "just works", offering basic features and plugins on par with contemporary professional DAWs. I imagine something like Ubuntu Studio, but with *no* assembly required. An important priority should be the creation of high quality, stable plugins, which are glaringly missing in the Linux audio universe. If accessible to the technical skills of an average sound engineer, you could promote it and charge for it. And if priced similar to the proprietary programs in the market today, you may be able to fund a full-time development team as well as a marketing budget.

The program would still be free (as in speech), but there would be a revenue stream which doesn't depend on unstable donations.

My intention is to provide constructive criticism from a practical end-user perspective, because I feel that Ardour has the potential to become the de-facto standard in audio production. As a Pro Tools HD user, I am tired of wearing Digidesign's (now Avid's) golden straitjacket.

@Auser: i suspect you are not

@Auser: i suspect you are not aware of how hardware dependent good DAW performance can be. There are companies such as Rain who do nothing but build PC's for Windows DAW software. And then, of course, there is Apple, who do more or less the same - picking very specific hardware components for their machines that are known to work well together. Even Digidesign, when they first released PT for Windows, certified only a single model PC for use (they have expanded that but are still fairly explicit about the potential problem with using "unsupported" hardware). Several other efforts have been made to create a ready-to-run Linux DAW platform. All have essentially failed, and one of the most fundamental reasons is that if you cannot control the hardware, it is very hard to control performance.

I think that you are also not aware that the major reason people offer as a reason for NOT moving to Linux as a DAW platform is the unavailability of essentially all of the plugins they have access to on Windows. There is nothing the Ardour project can do about this that makes much sense (we have provided a way to run Windows VST plugins, but it is fragile and not really supportable, given that it is essentially dark magic).

An important priority should

An important priority should be the creation of high quality, stable plugins, which are glaringly missing in the Linux audio universe.

google "linux plugins"

Paul, you are right, I'm not

Paul, you are right, I'm not very familiar with internals of DAWs; I'm a mere user, though a programmer by profession. But if performance is so hardware-dependent, there's no reason you can't make the same recommendation that Digidesign made when they released their Windows version. Maybe you could even assemble and provide hardware configurations that work well with Ardour. I understand that this doesn't fall within the scope of the Ardour open source project in its current state, but for an "Ardour company", it could be a viable idea.

If you focused on supporting specific hardware, at least some people would actually be able to use Ardour in a professional setting, and what's more, they would pay you for your hard work. Besides, why is Ardour abstracting over the all-important hardware? If you know hardware is as important as the software, why is Ardour development so bent on a breadth-first approach? The consequence I observe is that the project isn't able to leave the programmer's sandbox and become accessible to the general populace. Rather than spending extra time building a DAW that runs on all hardware, more value would be created by providing a professionally usable product today, hardware-specific if necessary, for which there is huge market. If the problem at hand can only be reasonably solved with a vertically integrated platform, then so be it: tackle the real-world problem, rather than solving an idealized version of the problem that is much harder. Anyone who wants to do serious production with Ardour is going to be looking for specific, good hardware anyway; you could point them in the right direction, or even better, sell it to them. And if the Linux that came with that hardware were a customized Linux, I don't think they would care. I would personally be more willing to buy an integrated, pro-quality Ardour hardware+software music workstation than donate to the Ardour project.

You say in your article, "...had Sibelius been an open source program, it would probably have never been generating $18M of cash flow/revenue." Being open source and making a lot of money aren't mutually exclusive: take MySQL, which was valued at $1B by Sun, or Red Hat, currently valued at ~$10B. Ardour too can be a commercially successful product, if executed with a user- and profit-centric strategy.