Tracks

[mdlbear]

Vocals for "I Wanna Be a Webmaster" and "Daddy's World". Between taking my car in for service and having to go to the Kaiser Campbell pharmacy to pick up a prescription, the morning was going to be shot anyway. Two guitar tracks left to record, some editing, a couple of choruses, and I think I can stick a fork in the recording phase.

Then comes tweaking the mixes, and the booklet and tray card layout.

Comments

[technoshaman.livejournal.com]

w00t. I was just thinking to myself this morning on the bus, "he's gonna have to level the mixes, but some songs are *supposed* to be louder or softer than others, so wonder how he's gonna do that..."

Oh, apropos of recording... cflute and I have realized that we're going to begin recording individual tracks, probably sooner rather than later, and were wondering about what sort of front-end equipment (mikes, preamps, etc.) we'd need to get started...

And while I'm thinking along those lines... I know you use Audacity to record. Do we need a special low-latency kernel to go with that, or...? And what would we need in terms of RAM/CPU/disk space?

[mdlbear]

I believe the low-latency kernel I'm using comes from http://archive.64studio.com/ -- you'll have to grep around to figure out what to put in your /etc/apt/sources.list; they seem to have reorganized since I last looked. These days you can probably get along without the patches; most of the low-latency stuff was part of the SMP code, which is part of the main line now.

My current lineup includes a CAD large-diaphragm condenser mic, a couple of AKG C100S small-diaphragm condensers, Presonus tube preamps, and an M-Audio Delta 66 PCI interface. I use an Edirol UA25 USB interface for field recording; I've been using it with the Mac so I have no idea how well it fares under Linux. Marshall and Behringer also make good, low-cost microphones; I used a Behringer C2 stereo pair for the last couple of concerts. I'm probably going to step up to a multiple-pattern studio mic. There was a good discussion of gear a couple of months back on billroper.

In terms of CPU... you probably need at least 1.5GHz -- dual-core and more speed would be better -- and at least 1GB of RAM. You can get away with about half that if you're recording a single stereo track and running the UI over X. I always record 24 bits for the extra headroom -- Audacity stores it as 32-bit floating point -- but only 44.1KHz because there's no point in downsampling it for a CD. That combination is 10MB/track/minute -- twice as much as "CD quality" audio. And you're going to want multiple tracks, possibly multiple takes, and eventually exporting a mix to a floating-point .wav file and from there to a 16-bit .wav file for the CD. An intermediate floating-point file for normalization and post-processing, in my case (feel free to use my Makefiles and scripts).

My recording directory currently contains about 17GB worth of track data and 1.8GB in the "Coffee, Computers, and Song" album directory. Disk is cheap. An audacity project compresses down to about half size in a git repository, which is probably your best bet for version control. Haven't tried subversion; the way it does a checked-out working directory would at least double your storage requirements.

Thinking about an album?

[mdlbear]

BTW: the UA25 includes preamps, as do most of the current crop of USB and firewire interfaces. Probably not as good as the Presonus tube pre's (which are $100/channel), but certainly good enough for field recording, and powered off the USB interface, which is why I use it.

[rowanf.livejournal.com]

Totally cool. Thanks for the update. :-)