just for future reference:  if you've installed an atheros-based wifi card, and linux is complaining that it can't reserve the pci memory region, you need to update your bios.  it's not a bug in madwifi, or ubuntu, or the linux kernel-- in fact, xp froze at startup for me while the card was installed.

so how do you update your bios if you don't dual-boot?  first, try flashrom, which is packaged by most major distributions.  if flashrom doesn't recognize your hardware, or if your board is unsupported, this page shows how to make a freedos boot cd that contains your motherboard manufacturer's dos-based flash utility.

i noticed something weird yesterday: my four-year-old delta 44 generates a high-pitched whine, with a frequency and volume depending on the samplerate it's running at.  i don't mean that it's adding noise to the outputs; the actual physical card itself is producing it.

it's most noticeable at 96khz and 48khz, where it's an f# (around either 2960 or 5920hz, possibly making it a subharmonic of the clock frequency).  the 96khz whine is by far the loudest.  it's also there at 44.1khz, at a lower volume and a higher frequency (too high for me to easily determine pitch).  you can hear it the same way you can hear whether a crt is turned on.  also, it goes away while i'm changing the gain on any of the adcs or dacs.

it's not a huge problem; it doesn't seem to be leaking into the i/o at all, and with the case on my computer, it's only even remotely audible at 96khz.  but i'd be interested to know if any other delta series owners have noticed this.  is it just mine?  is it just old?

as fond as i am of idaho's more recent work, nothing will ever top 'three sheets to the wind' for me.  it seems to be the perfect balance point in jeff martin's career..  the band's distinctive sound feels fully-developed, but unlike the subsequent releases, harmony is still given as high a priority as texture.  and given his recent shift towards piano-based songs, it seems likely that the world will never hear anything like it again.

one of the most memorable tracks for me is the closer, 'get you back'.  despite the fact that it clocks in at about 55bpm, its dynamics and intensity are enough to keep even the most ardent slowcore skeptic interested.  and as it turns out, the tempo is part of what makes this song worth looking at.

it opens on a DbM7, and stays there long enough to make you think it's the tonic.  but it's not!  the second chord of the verse moves up a fifth to the Ab, and suddenly you realize that what you thought was the I was actually the IV, and Ab is the real tonic.  but it's not!  because then it moves up to the Bb, and it turns out that the Ab was actually a VII all along.

except, of course, that it wasn't.

on its face, this probably seems ridiculous.  all three chords can share Bb as a tone center, given a modulation from the parallel minor, something rock songs do all the time.  so why isn't this just III/i VII/i I?

because musical cognition is affected by tempo.  in particular, the longer you hang out on one chord, the less relevant the previous chords are to the way your mind parses the music.  this song would simply not be possible if it were faster.  try it yourself:  play these three chords at twice the tempo, and see if it still sounds (...)

for no particular reason, i've started an ongoing log of every instance of the B♭-D-C squeak i come across in tv and film.  it's only a few weeks old, and i already have no reason not to believe that it's the most comprehensive squeak database on the web!

i don't know much about the squeak, except that it's everywhere; it's been catching my ear since i was about 10, and hardly a fortnight goes by that i don't hear it at least once.  i also don't know for sure that this kind of prevalence is unusual for sound effects..  it may just be that pseudo-melodic sound of the squeak makes it more distinctive than most, and easier to recognize.  if you know more about the history of this sound, please let me know.

see also: http://en.wikipedia.org/wiki/Wilhelm_scream

ascii rave in haskell.

as i understand it, he's using the karplus-strong algorithm (that generic plucked string algorithm you find everywhere) for the consonants, and a formant filter for the vowels.  it's put together in hsc3 (supercollider with haskell bindings).  the source code is available here.

(via zealan on #psycle)