it finally sounds right again.

i think it was the bios...

let's see what happens when i wake up. i really, really want to get back to work and am going to get upset if it's still broken.

you know, that's not sitting with me well. the bios revision would have to be at least 2008. this version of xp is slipstreamed up to the end of the product cycle. i can't think of anything that would default to jan 1, 2002.

it fixed itself on the reinstall. but it's just weird.

the reinstall came up dated to jan 1, 2002. bios date. i had to restart the process, because i can't have my system files dated that incorrectly.

but it got me thinking.

setting the clock in windows changed it in the bios, indicating these things are in sync. if the bios was corrupt, it could have been counting slightly wrong. and, that could have resulted in the whole system being out of sync, leading to the fading i was getting.

we'll have to see what happens. but i've all but ruled out a driver conflict, and i'm certain the driver clocks were installed right. yet, i was still getting what seemed like timing errors.

we'll see how this works out. but that cmos reset may turn out to be the fix.

that battery has never been replaced. it's only eight years old. but the board shipped with two. it's worth a shot if it's still bleeding.

hrmmn....

went i went to reinstall, i noticed i was missing a drive at the bios stage, which came back after i reset the cmos.

you know, it's ok if you don't know what the fuck i'm talking about. you signed up for music updates, not computer troubleshooting. i dunno. maybe you're learning something.

i can't know what caused that, but it's something to take note of. if i really did get attacked by something, that's the last thing i can think of resetting.

i had a bad flash last year that cost me months. i ended up getting a bus pirate to fix it, so it won't be nearly the problem it was then. but, i'm still wanting to avoid that....

let's see what happens on the reinstall.

well, that was pyrrhic. cubase wouldn't load the dongle, so i tried to reinstall and now it's crashing; this is a known issue, the setup files require something i remove in the wipe. i need to reinstall.

i think i've determined that the only logical cause of error is the creative emulation driver. it's the only other thing that the system is accessing.

we'll see what happens, but i've been getting downright *annoyed* over the last 48 hours or so. this is not fun.

my odsp is late. i'm going to flip out if i get denied, and spent my last few months doing fucking computer troubleshooting. it makes me want to vomit.

it's stable with the mixer, the h20 drivers, the dongle and the guitar synth - on a full boot. i'll try the m-audio drivers in a few hours...

gah. no. the creative emu blew it up again. but then it fixed itself when i turned it off rather than restarted. has that been it the whole time? i'll wait until i wake up.

it didn't like it when the soundblaster drivers set themselves as the default device, which seemed to load something in the background that nulled the drivers. setting the default device to the mixer resolved it. i would normally have the default device set to the m-audio card. i'll have to see what happens when i get there. but the next thing to turn back on is the guitar synth drivers.

i'm stable with the dongle. it's obvious to hear when it does click. am i stable with the soundblaster?

i've been saying for a while that connecting to the internet has been changing the tone, and that sometimes it seems like the bandwidth is cut.

i've deleted the firewire network port. and, in fact, i remember doing this once before. it's made an immediate difference in terms of opening the sound back up and making it sound less compressed. loading the tcp/ip layer may have been interfering with the stream some way or another.

before i did that, i cleaned the thing right out - deleted all the sound related drivers, wiped the registry, removed any executables. when i got the drivers back on, i had to flash the firmware again. this is something you don't want to do all of the time, as it's kind of dangerous; if there's a power outage, or just a fail, the device gets bricked. but, it's the only thing that is consistently getting me back to square one.

i then loaded the h20 drivers and it seems stable without the dongle running. but i'll have to give it another hour to see what happens.

all three wmi services are on the reinstall, along with wia. that would have to be it. in fact, the script tries to delete the wbem folder, but it gets an access denied because wmi is running.

i'm not going to rip it out, yet, i'm just going to turn it off and see what happens.

wmi is mostly useful for network admins. it's one of those face-palm microsoft moments, where they integrated driver software into network monitoring software. why, bill? why do you do these things? normal users don't want the network admin aspect of wmi, which is why i originally removed it. like, literally. you do NOT want that shit on your machine. but, it was built in such a way that certain hardware support requires it. it's a strong argument for linux.

....which i haven't moved to, btw, because i can't get drivers.

i'm reinstalling on a virtual machine to figure out where wmi even came from.

when i initially built the scripts, i had wmi disabled - deleted, in fact. i vaguely recall bringing it back in when i rebuilt the machine, probably because i meant to get a scanner, which i did, which i need to run in a virtual machine, anyways.

at this point, i can't think of a single reason why it should be installed in the first place.

with wmi driver extensions on: fail.

the wmi driver extension is just error logging...

i'm going to try and reinstall with it on. and then reinstall with it deleted, and delete it after. it makes more sense to me to think it's interfering than that it's required. but we'll find out.

this is still fading, but not nearly as badly. fail.

i'm going to try with the wmi driver extensions turned on. then do the reinstall with it on, & etc.

this is a mild enough fade that i may have not noticed it before. but it's there.

it's becoming clear that i may have to re-upload some material from last year. i've known since the start that the files sound slightly different on the laptop, but i just assumed it was because it's a laptop.

uninstall h20 driver, alesis drivers, reinstall h20 driver, reinstall alesis drivers ----> apparent success. i'll have to test further when i wake up.

if this turns out to be a fix, i can only conclude that there's a third driver causing the conflict, because i don't touch the h20 driver in the script. i'm still leaning towards line6.

reinstall drivers with h20 driver disabled ----> success

reload with h20 driver disabled: fail. fading highs.

i don't think it's the driver, exactly. way back when i started this, i said something like "how did the user mode driver framework get turned on? i wipe that in the script.". the answer is that cubase installs it. i must have reinstalled cubase at some point since the last reinstall.

first, i need to ensure this is repeatable.

h20 driver load: fail. fading highs.

ok. good news is it's totally stabilized, convincingly, and i'm certain it could stay like that if i left it that way. sounds just like through the laptop - clean drivers, stable, no ambiguity.

bad news is that means there's a conflict somewhere, because i turned off every single sound-related driver and reinstalled the mixer drivers to get to that point.

i'm hoping it's the line6 drivers, because there was already a known issue with them. but, what i'm going to have to do now is turn each driver on one by one and see where it breaks. it might even break twice, who knows.

see, i'm not reproducing this..

so, whatever busted the output resolved itself.

see, what i did was cut the program while the mixer was on, then restart the mixer. and that fixed it in real time. then i rebooted.

so, it could have been some kind of crossed path. the hardware mixer might have picked up something connected to the m-audio card, then dropped it and recreated a correct path, which it's now holding to.

hopefully, it's that easy.