Monday, August 8, 2011

Does FL/OSS Have the Tools to Compete?

Some topics get talked about so much that the rhetoric becomes the topic.  FL/OSS versus proprietary audio is like that to me.

However, in this blog I would like to explore (and maybe inspire discussion) about FL/OSS and proprietary audio systems by comparing critical functionalities throughout the typical audio chain. Note that the comparisons will exclude hardware that does not interface with the OS or audio applications.

Q: Why functionalities?

A: Because I believe it is more important for people to understand what is possible rather than if FL/OSS mimics the same methodology or work flow. The question should not be Does FL/OSS have the foomatic-2000 reverb? but rather Does FL/OSS have a good reverb?

Interfacing the Problem
The first step of the audio chain is the audio interface which converts analog sound into digital bits, also know as the analog digital convertor (ADC). A low quality ADC can adversely affect sound by coloring it, introducing artefacts, or even inaccurately converting it. Closely comparing the original analog sound and the "converted" digital sound can reveal these issues. One term that is used often to describe quality ADC's is "transparent"; you don't hear the conversion, you just hear the original sound.

Luckily, and speaking holistically, audio interfaces with quality ADC's are available for the inhabitants of FL/OSS environs. From the very expensive (and very good) stand-alone rack converters to PCI cards, such as the RME series, quality converters are quite available.

Quality ADC's can be prohibitively expensive for many, but the goal of this discussion is to compare the functionalities available, not as a budget guide.

Lastly, I will mention bit depth and frequency only to explain that I will not mention them. If we are discussing transparency and high end converters then we have long crossed the threshold for adequate bit depth and frequency.

Now that the analog sound has been transparently converted to digital we would expect it to be recorded.

Work flow aside, I would posit that any modern DAW will properly record incoming digital signal to hard drive comparatively and adequately.

Plugging In
Disclaimer: this is a subject for which I am neither an expert nor have moderate experience with proprietary plugins.

The next major stage that effects the audio signal are plugins.

Thousands of proprietary plugins exist and more are being made every day. I would like to exclude the "black box" plugins (those that just "make your music sound better") for this discussion and focus on the "classic" plugins like reverb, delay, flange, etc, although I concede that this is slightly unfair for comparison purposes.

Hundreds of LADSPA and LV2 plugins are available for free to the FL/OSS musician. Additonally, many are available for purchase, like the excellent LinuxDSP plugins. VST plugins are also available, both free and purchased, although some DAW's may require hand building them with VST support enabled.

FL/OSS plugins might not have the numerical advantages and the breadth of spectrum that proprietary plugins have, but the base line functionality is definitively available.

Mixing and Mastering
I will define mixing as the consideration and manipulation of signal levels, frequency, dynamics, and panoramic position between instruments on a particular track.

Can FL/OSS applications handle this?  Absolutely.  Ardour, for example, does this easily.

I will further define mastering as the consideration and manipulation of signal levels, frequency, dynamics, and panoramic position between tracks on a particular album.

Again, can FL/OSS accept and defeat this challenge?  Again, absolutely.  Using JAMin coupled with Ardour is a very powerful, yet uncomplicated, method to accomplish this.

So here's the summary where I ask a probing and insightful question all while summarizing the blog and finalizing it with a witty answer.

If FL/OSS operating systems and audio applications can provide the same fundamental functionality as their proprietary counterparts, then why aren't more people using them?

I believe the answer is multifaceted, yet simple, and includes people and knowledge.

If Bob Katz was using FL/OSS to master albums I believe many would have already explored this options as well.

But what if a well recorded, mixed, and mastered album was made with FL/OSS and actively and robustly publicized as such? Would that be enough to quieten the naysayers and energize others to explore it?

Would you?


Ungleichklang said...

I think the answer is very simple.
Linux is a very good OS, mostly used by some "geeks" (like me) on servers.
This makes linux-users a minority group. Within this minority group there are some other "geeks" (like me) that make audiostuff, which makes them a minority group of a minority group.
If other non-geeks even can not use OpenOffice instead of $MS-Office, how do you expect them to install and make music with an "unknow, mystery, geeky" OS??
But exactly this users want make music and work all day on "the other" OS, they became familiar with (they know where to click to ...).

Pro Audio apps on linux do run but are not as numerous as on other OSs and we had a big stagnation in audio-app development the last years, luckily since 2010 this has changed. Because of this only recently some "commercial" and well known DAW-devs offer or still develop a linux version of their products. So we have a big latency in developing some "fame".

Will it help if some famous people use it and say that they love it .. yes it would a little bit.
Maybe the time will come ...

lsd said...

I think the biggest gap in functionality between open-source and proprietary apps is in the area of software instruments. Qtractor (as of 0.5.0) and Ardour (as of the upcoming 3.0 release) both include quite solid MIDI sequencing tools, but there's a lack of instruments available for them to drive.

It's also frustrating that many of the best soft-synths, such as PHASEX and Yoshimi, are available only as standalone JACK synths. They do work that way, but it's not a great user experience: it's much more effort than simply adding a plugin, it causes problems with automation, and it can require manual effort to ensure patches are loaded and connections are made (unless something like JACK session support is in place). It's especially confusing for new users.

Not everyone needs that stuff, though -- with Ardour, a bunch of LADSPA and LV2 plugins, and a decent audio interface, you've got a great audio recording and mixing environment, and if you're using mostly hardware instruments, that's all you really need.

I do think a high-quality album made (entirely or mostly) with open-source tools would help; I'm not sure exactly how much it would help exactly, but it still sounds like a good project to me. The bigger question may be whether such an album could be given a big enough profile to really reach the kind of audience that might be swayed by it.

Scott said...


You make a very good point about work flows with your example of Yoshimi as a standalone JACK-enabled synthesizer.

I do admit that I ducked considering work flows in this blog. But the purpose was to show that FL/OSS has the fundamentals in place.

LSD, do you think if the tools were consolidated into a more coherent framework thus presenting a unified work flow to users would progress FL/OSS to the next level?

And what else could be done to make it more appealing to the audience at large?

lsd said...

I just think it'd be nice to have some good synth plugins :) Having a good selection of soft-synth plugins available out-of-the-box would go a long way toward making the tools more accessible to new users.

In the proprietary world it's typical for a DAW to come bundled with a selection of synth and effect plugins. Under Linux, I think it's the role of the distributions to provide those plugins. That's already happening, of course -- there's just a lack of plugins to begin with!

It's unfortunate, too, that LinuxSampler runs awry of most distribution licencing policies. It's a really powerful tool, and having it available for easy installation, or even installed by default, along with some good CC sample libraries, would help a lot.