Subject: Re: [linux-audio-user] Pro Audio Software RANT!!!
From: Paul Winkler (pw_lists_AT_slinkp.com)
Date: Fri Jan 25 2002 - 21:13:58 EET
On Thu, Jan 24, 2002 at 08:37:38PM -0800, Ken Locarnini wrote:
> Honestly,
> I've been in the Cakewalk windo$e camp for awhile, and for the last year
> or so I installed Mandrake 8 to give linux a go. I was so excited to get
> away from microsoft. I began to explore all the linux apps out there for
> linux to find my equivalent tools to work with, but the sad fact is,
> -Unless you are into hard-core computor music composition, i.e. coding
> software algos for music
> -Unless you are willing to wrestle with your hardware, far beyond getting
> your sound card to work with asio drivers to get that perfect low-latency
> -Unless you are willing to age 10 years compiling, re-compiling
> applications, drivers, libraries etc., and doing nothing musical
> STAY AWAY!!!
I hate to admit it, but in this case, I agree. If you want to run a
professional studio *now*, there are no applications that will satisfy
you. It still requires way too much fiddling around to get anything
done. I spend more time fiddling than I do making music on linux, and
I've been at it since 1997!
For example, probably the closest thing to what you want for
recording/mixing is Ardour, which *will* be professional quality -
Paul Davis has been working on it full-time for years and he's not
just going to stop, and he's attracted a small crew who help out. But
MIDI support (other than MMC) isn't planned for version 1.0, so even
once the remaining issues with the recording & mixing app are worked
out, you'll need something else for MIDI. Certainly Jazz++ is not in
the same league as the current generation of Windows / Mac commercial
MIDI sequencers. MuSE is supposed to be promising (haven't had time to
check it out myself) but still kinda unstable.
Ecasound is more stable than ardour, but a) I don't know if it's
designed to scale up to, say, 48 tracks (maybe it is?); b) the
currently available user interfaces would make a Cakewalk user go
crosseyed and blink rapidly. It's mostly text mode, for god's sake! I
think it's cool, but I would be amazed if anybody ran a studio with
it.
There are other issues... for instance we have a growing collection of
plugins, but no standardised way to give them nice GUIs. So they tend
to look a bit ugly. Doesn't matter to the sound, but it *does* matter
- consciously or not - to paying clients if they look over your
shoulder in the studio (and they do).
My advice, for now, is keep using your previous OS for studio work. Be
prepared to do so for at least a few more months (maybe years,
depending on what you need). Keep playing with Linux and keep checking
back with us every few months. You'll sooner or later find something
that fits into your studio needs.
Finally, remember this bit of the Ardour FAQ:
"2.4. Ardour needs to have this feature now.
Paul accepts donations through his paypal account. The feature will
(probably) be added more quickly if you pay him. Contact him first."
I suspect this applies to many linux audio projects. Studios
routinely pay to have their hardware customized; why not do the same
for software?
> I have found that there is alot of people in linux who get enthused, get
> big plans, begin writing the new monster app only to loose steam later. Ya
> you have the source code but are YOU willing, or have the skills, to get in
> there and code what needs to be done?
Usually not... Part of the problem is that most apps that find their
way into professional studios (ProTools, Digital Performer, Logic,
Cakewalk) represent a truly enormous amount of complexity compared to
the average open source project. The user interface alone represents a
staggering amount of work.
For instance, a little experimenting with find and wc tells me that,
unless I missed something, Ardour is about 67,000 lines of C++, about
half of which is in several custom libraries, 30,000 lines is in the
main GTK version of ardour, and 5,000 lines is in an alternative
keyboard-driven interface.
Or take a look at one of our oldest and most advanced sound editors,
Snd. It's 171,000 lines of C plus 46,000 lines of scheme!
These are not the kind of project you can dive into over the weekend
and make significant changes. (Well, Snd might be, if you can
read/write Scheme.) Getting involved takes patience and probably some
commitment.
This leads to the laughable / lamentable situation we have with MIDI
sequencers and with soundfile editors. Say a talented young
programmer, or a pair of them, wants one of these. They have a little
spare time, so they look around at what's available. It's all
unfinished, with design or interface decisions that "aren't the way I
would do it", and there's always some important (to the user) feature
missing. Full of ambition, our young programmer decides that all those
thousands of lines of code aren't worth the bother - "I can do
better!" All this is easy to understand. Two years later, the project
has either been abandoned, or it's struggling along at version 0.8.2
with the author(s) wondering why nobody's helping out.
Compare this to the situation with Cakewalks, where over the course of
several years, a company with secure funding paid has paid for
thousands and thousands of man-hours of development time. It gets done
because it has to get done - it's a job.
Don't get me wrong - I think the linux-based pro-quality studio is an
inevitable reality. It's just that there aren't that many people
working on it, most of them are working in their spare time, there's
no funding for people to work full-time (with the exception of OSS's
commercial venture, and SuSE's support of ALSA), and there are basic
human tendencies that make it "easier" to strike out on your own than
to throw yourself behind somebody else's project. For all these
reasons, things are moving slowly.
But they are moving.
--paul winkler home: http://www.slinkp.com music: http://www.reacharms.com calendars: http://www.calendargalaxy.com
This archive was generated by hypermail 2b28 : Fri Jan 25 2002 - 21:01:55 EET