Slashdot Mirror


Solid-State Mini-ITX Linux Recording Studio HOWTO

An anonymous reader submits "LinuxDevices.com has posted a project howto on building a dedicated music recording and editing computer that uses a CompactFlash card instead of a hard drive, to eliminate hard disk chatter. It uses the latest release from the Agnula (GNU/Linux Audio) project, and the newest Epia MII-12000 mini-ITX board from VIA. The method described in the article applies to embedding most any Knoppix-based Live CD onto CompactFlash boot media."

9 of 242 comments (clear)

  1. Re:Hmm by Anonymous Coward · · Score: 4, Informative

    >in just 650MB of Flash storage space that is mounted *read-only*, to maximize the life of the CF card.

    Looks like they thought of that.

  2. flash memory by SuperBanana · · Score: 3, Informative
    building a dedicated music recording and editing computer that uses a CompactFlash card instead of a hard drive

    Big problem with CompactFlash- you can kill the card. They have a very finite number of write cycles. It's in the millions, but you can burn through those VERY quickly if you aren't managing your writes. CompactFlash in a camera, for example, only sees sequential writes, so you can literally fill the card and erase it hundreds of thousands of times before it's zapped.

    The same may be true when recording, but when you start talking about editing, things get messy. God help you if you put swap on the card.

    CompactFlash also doesn't seem nearly fast enough for real time audio beyond maybe 1 or 2 channels.

    Really, I don't see the point. Use a laptop; many modern laptop drives are so quiet you can barely hear them in a dead silent room, and if they're too noisy, run your cables into another room, or put a pillow or box over it, etc. You can buy a ton of memory at decent prices and use ramdisks if you're really concerned about HD noise.

  3. Agnula is alive and well by FreeUser · · Score: 4, Informative

    Clicking on the "to enter the site click here" link, we find that the site (and the project) is up and running just fine.

    Software patents will either be recinded, or software development will come to a screeching halt and ALL free software will be killed, not just this project.

    In which case we can all just pack up and find another profession, or move somewhere other than the US and the EU (if current legislative trends continue). After the IT economy has been destroyed and innovation has moved to India and China, perhaps the US (and possibly EU) beurocrats and politicians will get their heads out of their asses and ban software patents ... assuming the West doesn't just bully the Chinese and the Indians into adopting similiar measures and crippling their own tech industries as well.

    I am quite frankly amazed at the EU's stupidity in this, as it clearly benefits Microsoft and other big American firms, to the detriment of European startups such as Suse, Mandrake, et. al. But that is neither here nor there.

    I will continue to develop and use free software (including this project) until such a time as $un, Micro$oft, or one of their stooges ($CO) kills free software dead, or reform occurs.

    At which point I will continue to use and develop free software, until such a time as their thugs pry my keyboard from my cold, dead fingers...but that is a rant for another day.

    --
    The Future of Human Evolution: Autonomy
  4. Re:Noise levels by Kiryat+Malachi · · Score: 3, Informative

    The solution, of course, is to use real audio interfaces, which will have an external breakout box and digital interface back to the computer from that point. At that point, the length of cables back to the computer become a lot less important.

    For example, you could use a nice standalone A/D box with a ADAT-compatible output, then string your digital cable the 15 feet into your nicely isolated computer closet, where it enters an ADAT card. Run monitor and keyboard cables the 15 feet, and you have a system that can be as loud as it wants to be without getting anywhere near your recording.

    Of course, for real recording, you're going to want to isolate control from recording, so you can have a somewhat noisy computer in control (so long as its noise factor is less than what you can tolerate during mix and edit).

    This project is neat for geek factor, kinda like sticking SSH on your cellphone, but there are a lot of easier, more useable ways to minimize recorded noise.

    --

    ---
    Mod me down, you fucking twits. Go ahead. I dare you.
    (I read with sigs off.)
  5. at least 20 GB free, depending by 192939495969798999 · · Score: 3, Informative

    If you're doing a studio project with 4 instruments including a nice drum set, and it's a live band, you can expect to have at least 16 tracks, meaning 16x5 minutes of audio, or 80 minutes for one take. Assume 4 takes, and that's 320 minutes of record time, or about 2800 megs, for one song. I would anticipate needing to have 8-10 songs on the drive, and then burn the rest off to DAT's for mastering some other time, so that figures to around 20 gb free. That's my experience from being in real (see: records artists you've heard of) production studios more than a few times.

    --
    stuff |
  6. not-so sequential writes by vlad_petric · · Score: 3, Informative
    Everytime you modify a file on the filesystem, the file allocation table gets modified ... And pretty much everybody uses FAT16/FAT32 for CF cards.

    jffs2 is much more conscious about write behavior, so I'd strongly recommend it for anything on a flash filesystem.

    Anyway, the main reason compact flash is rather slow is simply the fact that few people need high throughput. There are cards these days that sustain a throughput of 15M/s, but they're only meant for high-end cameras. While flash is slower than RAM, it's still considerably faster than mechanical devices, so I'd expect this number to go higher.

    --

    The Raven

  7. Re:Hmm by bsd4me · · Score: 3, Informative

    Dont flash cards have a maximum number of write operations? Or is that USB keys?

    All FLASH devices have a limited number of write cycles. Looking at the specs for a random device shows that modern devices support over 100,000 write cycles, and I think this is per sector.

    A good device driver will use various techniques, such a wear leveling, to extend the life of the device.

    --

    (S(SKK)(SKK))(S(SKK)(SKK))

  8. Re:I've got the ultimate silent PC right here. by silas_moeckel · · Score: 3, Informative

    External firewire/USB/Box on a cable has been the only way to get good recordings of analog inputs on a PC for years. As long as you do you A to D outside the box your fine.

    --
    No sir I dont like it.
  9. Re:NFS? by Kiryat+Malachi · · Score: 3, Informative

    If you want a simple, relatively small and quiet, two channel recording rig, with equivalent sound quality to this, I highly suggest buying a fucking Minidisc or DAT deck, a decent mixer, and a couple good mics. Then you can dump it to a machine with decent editing tools later.

    And the best part? It is silent.

    What the hell advantage does this system have over a DAT deck and a computer with editing software worth using? None, because its a two-track system using a consumer-level sound card. Any gains you might make in reducing hard drive chatter will be totally overwhelmed by the crap quality of your A/D subsystem.

    This thing is barely suitable for use as a two-track tracking machine, and there's no reason to edit on this thing as opposed to a decent PC which won't run into disk space or flash write limitations.

    --

    ---
    Mod me down, you fucking twits. Go ahead. I dare you.
    (I read with sigs off.)