Slashdot Mirror
Linux Audio Developers Conference
paulbd writes "This weekend sees the first Linux audio developers
conference at
ZKM in Karlsruhe, Germany. Gathering together many members of the
Linux Audio Developers mailing list and others, the conference will feature 2 days of in-depth technical presentations and demonstrations of many
cutting
edge Linux
audio and
MIDI
applications." Desktoplinux.com has a related story about using Linux in a professional recording studio.
This all appears to be text, are they streaming the presentations, which would make sense at a conference like this ?
An Eye for an Eye will make the whole world blind - Gandhi
Sound support is one area where Linux has consistly trailed more important Operatin Systems such as Microsoft Windows and Macintosh OS. Where those systems have had Professional quality support for Professional quality hardware that works well, Linux has been stuck in the background.
Perhaps this Conference can identify and deal with such issues as:
1. High Latency when performing other tasks such as opening windows or moving windows around. This leads to stutters in Audio and MP3 Playback.
2. Poor compatibility with Professional and New hardware. Realistically, although most people use SB AWE64 and SB Live! sound cards, most Professionals use newer cards and many new computers have other cards. Linux is not compatible with hardware that is newer, cheaper or more expensive.
3. Poor feature support for Linux, because it is good support for features such as 3D Sound and MIDI Music playback.
4. Best Stability on Linux audio drivers. Other Operating Systems have drivers that crash less for Audio Hardware. Linux is a very much more stable Operating System in most respects, but the lack of stability in audio drivers is Irritating.
If these issues can be addressed then Linux could be a top quality audio platform!
But how many hardware manufacturers are actually putting out low-latency ASIO drivers for Linux?
(On the plus side, Linux does have CSound and PD, which are excellent progs for electronic musicians.)
I recently installed Gentoo Linux with modest optimizations for my processor (athlon-tbird 1GHz at -O2), expecting some pretty snappy response. Every app and driver was compiled from source with compatabilities built in for ALSA and OSS. I though it would be better than pre-compiled binaries.
:)
I've been quite disappointed. Maybe I layered in too much.
Noatun plays MP3s with only modest smoothness. mpg123 suffers similar problems. Skips are common when switching or redrawing windows. Real users stick to command lines, I guess.
I haven't tried recording from a live source, but I'd be wary -- is that weird pause in the music because of the recording skipping, or the playback skipping? Which system do I trust?
Anyway. Perhaps I tried stuffing in too much compatability, and instead should have picked one system over the other. But then who knows which apps would work and which wouldn't?
Please please please -- can we have a standard layer that's easy to install?
GMFTatsujin
Wow. I've never read such a vague article in my life.
Here's the synop:
We used Windows. It crashed and got viruses. We didn't want to upgrade to XP.
We played around with Linux. We decided on Mandrake. We went Ogg Vorbis. Life is grand.
Nothing on the implentation, nothing on what programs/hardware they used in Windows or Linux, nothing in regards to performace of said hardware and/or ported software.
Linux is great for them, but being too vague doesn't help small time studios understand how to use it in their shop, or how best to go about it.
Why not get a little more in-depth, such as what utilities they used, what hardware settings needed to be tweaked (if any), and how difficult it was to train for.
For example:
What was the hardest part to train/learn?
What features are you hoping Linux audio programs will add in the future?
What advice would you give to a small, struggling studio in regards to using Linux in a studio?
Do you know of any other studios who have utilized Linux?
The list goes on.
a) it's great to see the various specialized summits and meetings that have happened in the last few years especially (distro-based, or "Desktop Linux" or kernel summit, or ...) -- it's impressive that the various subsystems are independently good enough (and independent enough, if that makes sense) that improving one system does not (usually) kill the others.
...), I tried q -1, and am shocked at how good it is! I ended up with a compression ratio of about 35-to-1 (350-some MB total WAV files; approx. 10MB squashed to -1) and sound that in non-critical listening environments would be jes' fine, thanks. I should try compressing to q -1 and making it *mono* first, too.
b) Speaking of Linux sound, a nice thing: the other day, I compressed some music (passengermusic.com is the band's site, though no music is on the site) for a musician I know, because I'd like to convince him to post some music in ogg vorbis format on the band's website.
Usually, I have used grip to do such compression (nice interface, easy), but this time I wanted to try a wider range of qualities without going in an changing grip's preferences several times, so I started up oggenc instead.
Compressed at q6, the sound was predictably good, and my tin ears on my low-end equipment could not tell from the original. Sadly, same is true at 3. Probably most of the other available integers, too.
For kicks (and since this is for web use, and since most people are still on dialup, and since long downloads are a pain in the tuchus
timothy
jrnl: http://tinyurl.com/c2l8yr / foes: http://tinyurl.com/ckjno5
However, supporting pro hardware, and syncing MIDI with real audio, at 24 bit resolution, including a reasonable GUI to do it all with, is the domain of Mac/Windows only as far as I can tell. Cakewalk Sonar leads by a long way on this. You can add digital effects in real time, chuck in a canned drumbeat while you lay down the first couple tracks, export to a single compressed file (lossless) and all sorts of wonderful stuff. That's what a pro studio needs.
Editing single stereo files is NOT what professional recording studios do. Radio stations have very low requirements in this regard, they just pre-record shows and interviews, compress them in a lossy format, and send them out. Since FM and digital radios have analog or digital compression anyway, then OGG at high bitrates is fine. However, artefacts from SEVERAL mp3/ogg streams all in a multitrack environment is not acceptable.
Conversion Rate Optimisation French / English consultant
There's a recent project that tries to integrate all the best audio software into Mandrake Linux 9.1, including a patched multimedia kernel for the low latency, the ardour sequencer and other stuff.
It's all explained in this howto.
Having spent the better part of a weekend trying to get RedHat 8 and ALSA to jive with a fairly standard pro-consumer soundcard to no avail, I'd say the first thing Linux sound developers need to do is get shit working. I mean ALSA is absurd: every time you upgrade your kernel, you recompile. Explain that to any pro studio user who's used to the Mac/Windows "install driver, reboot, get working" way of computing. Then explain to them about all the command line stuff they'll have to learn, about two differing and conflicting user interfaces, about all the different distributions and package formats. I'm not sure pro users are willing to sacrifice hours of thier time for a 5ms drop in latency or whatever Linux audio developers can promise, and they certainly won't give up tools like Logic, Reason, and Pro Tools. Get stuff working for average users, Linux audio devs. Then we'll talk.
Another alternative, is NetJuke. Did I mention it's free software, unlike Andromedia?
This is my own contribution to Linux audio programming. Its pretty basic, and possibly buggy, but it was fun to write.
The name and the idea was inspired by the infamous Mac application "Back to Basics" that was used onstage by the late, great, band Man.. or Astroman? Its actually much more basic than that, though.
It uses curses, and one of the intended uses is that the "musican" could be on stage controlling an unseen computer with vt100 or other such hardware.
http://hacked-2-basics.sourceforge.net/
Too bad it doesn't seem to be jack [sourceforge.net]-aware yet. Though that should change with the new portaudio [portaudio.org] version, which already has the beginnings of a jack-output.
Yes, Jack support is definitely coming, as soon as PortAudio v19 is finished.
- Dominic [Audacity developer]
Audacity on Linux (well, the 1.0 series) is only good if you don't need full duplex recording. Which means you cannot do any multitrack recording with it. It shows promise, but it has a long way to go.
Version 1.0 is based on a codebase nearly 18 months out of date. Try the latest code from CVS, you'll be pleasantly surprised. Full duplex on virtually all platforms, floating-point samples, real-time resampling, and lots more. We're hoping to release version 1.2 in a few short months.
- Dominic [Audacity developer]
I love Linux but it is WAY behind in professional audio (sound blaster cards are not pro quality). Digital studios need at least 24 tracks of 24 bit sound and a badass software sequencer like Pro Tools, Digital Performer, Cubase, Logic, etc. I can't name a single pro audio interface that supports Linux and is compatible with a high performance sequencer. Just so you know before you tell me why I'm wrong, I'm an Audio Engineering major at Berklee College of Music. Most people here have never heard of Linux. Don't get me wrong, I'm all for open source, but Linux doesn't have a chance in the pro audio domain. It's way too far behind. The article is about everyday things like mp3 support and cd burning, not professional recording.