another day of procrastination..

an election was just called in canada. true to form as the nerd that i am, the policy discussion [rather than the name-calling that will dominate the news cycle] is of great interest to me. under the proper circumstances, i could see myself running for office, one day when i'm reaching retirement age.

i'm pretty sure i've got it stable, and i just need to see, now, if there's any conceivable way to get the files to null - i need to determine whether i'm dealing with corruption at the cubase level, or if the drivers were just right fucked. i can't shake the logical flow charts that combine before the out, and the pattern turned out to be relatively weak in the end, so a broken cubase really seems like a more rational problem. but that might mean i can't recreate it at all.

on that end, i went through and made sure there were no leftover file access connections to the previous install [by taking ownership of the secondary drive i'm using], and i'm going to want to do some things like try and render from other disks, and maybe defrag it.

i may find myself in the annoying position of having mixed everything in broken cubase projects and being unable to recreate the results. not there yet...

...but i do honestly think that the drivers are really finally clean.

this is the first time in years and years that i've seen such a thing. honestly. the last time i saw any kind of worm or virus or anything of the sort was about 2004. i don't run anything to counter it. it's actually very hard to get infected, but those library machines are kind of going to be swarming with shit and i should have been more careful...guess i got complacent...

as mentioned, it couldn't do anything on my machine. the instructions would just error out. that's by design. no, bill, we don't need telnet. but, it can replicate itself and keep going. and that just might be what's been pissing me off....!

there's no sign of it in any of the registries. no remnant files. i formatted the key. i wiped the device. it's gone. but, it might be it...

i just found a worm on my usb key. i didn't think to check the date on it before i deleted it, which i immediately regretted. but the last time i brought it out somewhere was about the same time things started going haywire - to print odsp forms at the library at the beginning of july.

it's apparently a browser hijack. it wouldn't work on either of my xp pcs, because of the way i've stripped them down. i don't know how it would react on the 7 laptop, but it's probably going to come up against user privileges. i've seen absolutely no physical signs of malware, and really tend not to worry about it much. but i did notice that the drivers went bad, the last time, after i went online.

firewire is a networking device. it's not entirely insane to think it could have squirmed in there and imploded. the firmware update should remove this as anything to think about further.

it was probably sitting there for a month...wow...i need to be more aware of that...

i'm now convinced that there's corruption in the cubase files and that i can null the results by replacing the tracks. i've experienced this before, but it was with midi tracks that were easy to recreate. i'll have to convince myself with some testing. if i'm right, this might be time consuming.

i found the firmware updater on a backup drive, and it does seem to have made an immediate obvious difference, although i'll have to see if it holds. i'm going to try and rerender the files that aren't nulling, because the plot just thickened - a july 12 re-render just nulled, and that doesn't make sense because it's definitely after the break.

ok.

so everything else is nulling. this isn't an old problem that i didn't notice. i can have faith in my ears, which seem to be quite sensitive. i don't have to go back and fine comb anything.

i'm running the null tests a little further past where i did before, and i'm noticing that a pattern is developing: it's the first two tracks. always. the rest always nulls.

it's also becoming apparent that the break didn't really happen until about five tracks in. in hindsight, i think i dropped some effects work on the third and fourth, but i'll have to see if i can recreate it; i vaguely recall playing with a quadrafuzz. they're not fitting the pattern - they just seem to be missing distortion. i remember re-rendering the second because it didn't sound right [that would have been the first time i noticed it was broken], so the earliest version i have of this isn't clean. from the fifth on, i get the pattern with the first two tracks.

...which means it couldn't be the codec install, because that was a lot earlier.

i still can't explain why the first tracks are always broken, but given that it's consistent it's starting to seem more like a broken cubase.

but i'm now starting to wonder if the clock issues might be related to the constant stretching i was doing. i sent these files in in such a way that they're subtly out of sync. i then had to carefully stretch them back in sync through trial and error. could i have broken the clock in the machine?

a firmware reset in the device is a useful troubleshooting step, i just don't know if this is possible.

i'm going to finish running through nulling the sequence, and then see if i can at least get consistency over the set of new files.