crisis of incessant white noise averted
i was greeted this morning with a rather disturbing development regarding the plans for the future audio facility here: somebody had decided that putting a machine room up against our outer wall was a good idea. a machine room. the kind of thing that goes “GHGHGHGH” at about 60 dB, 24 hours a day. the kind of thing that probably has to be cooled by HVAC, which also goes “GHGHGHGH” as it vibrates the floor / walls / ceiling. the kind of thing we spent a month of floorplan back-and-forth making sure we avoided (can’t be near the bathroom! can’t be near the elevators! can’t be near the central HVAC! gotta be on seperate electrical!). what the deuce.
then just as i’m trying to figure out how to deal with this – hey, let’s float the machine room! – word comes back that no, the machine room is not up against the studio after all, but *across* from it. whew. for a minute there i was afraid i was going to earn the eternal disgust of my pro audio smooth buddy who’s been giving me advice on this project. don’t worry greg, i’ll still manage to screw it all up somehow. stand by.
the cart before the chicken & the horse and the egg
i seem to have a lot of posts that follow the format “ah, i remember when… / so, nowadays…” or sometimes the other way ’round. so in that fine tradition…
ah, i remember when. working at Kesmai with a relatively small team of a few dozen folks, there were two ways of getting your pet feature (or desperately needed functionality) into the game. The Right Way – having an approved design, getting your tech docs in order, getting implementation on the schedule, and let engineering take its course. and The Ninja Way – ascertaining the programmer’s favorite scotch, procuring a bottle, and presenting said bottle in exchange for getting the work done after hours (and not while consuming the scotch…that part is key).
so, nowadays. we’ve got a pet feature that we want to get in, and we’re really psyched on it. thing is, Zenimax is a much bigger studio, working on much bigger projects, with much bigger development processes. almost everything has to be done The Right Way. it’s much harder (and more frowned upon) to do stuff The Ninja Way. loose cannons are bad for well-oiled machines, to mix my metaphors. but our pet feature is of such a nature that one almost has to see the finished version in the game to be convinced it even belongs on the feature list. that is, we have to get it done before we can get it started – the best way to build momentum for design approval is to actually show the damn thing to people. holy catch-22, batman.
hm. what’s even more stealthy than a Ninja?
night of the living baseheads
once upon a time, perusing my sound effects libraries went something like this: search for the sound i needed via some crude FileMaker Pro thing, pull the physical audio CD off a shelf, stick it in my CD-ROM drive, spin it up and RECORD IT INTO MY AUDIO EDITOR. record. the audio. in realtime. granted, this was 1995 and it was also the first year of my doing things like “perusing sound effects libraries”, but still. kee-ripes.
eventually i wisened up and ripped all the SFX CDs to my hard drive (boy did i ever think i was clever) and suddenly had all those sound effects right there at my finger tips. needless to say, that sped things up a bit. that also afforded me the option of doing clever things with file renaming, so i could use Windows search capabilities to find files my own way.
also i found myself doing a lot of database prep work, using Audition batch processes to strip out the silences between cues, hi-pass stuff @ 15 Hz, even do a little compression & limiting across the board. so while i was using a combination of FileMaker Pro & Windows to navigate the sound libraries, i was simply using the Audition “file open” dialog to, yes, audition them. with auto-play enabled in the dialog, i’d click around all the files until i heard what i wanted. i got to know the first few seconds of every file very, very well. :) and thanks to the prep work, a file that originally might’ve been 20 seconds long with five explosions at different volumes all seperated by a few seconds of silence became a 10 second file with all the explosions kinda sorta “mastered”. BOOM BOOM BOOM BOOM BOOM went the auto-play. that was me perusing my libraries in the early 2000′s.
now of course software like Soundminer and (my fave) Basehead combine uber-thooper-dooper database search & retrieval and keyword management (buh-bye FileMaker Pro) with on-the-spot auditioning (buh-bye “file open” dialog) in one sexy package. type in “horse fart” and get hundreds of…hm? what? oh, just the one? okay, “horse fart” gives us just one hit – a real gem, mind you. but type in “metal rattle” and get hundreds or even thousands of hits, depending on how many libraries you’ve got lying about. you can listen through them all as fast as you can arrow up & down, or jump into each file a bit, and even pitch it up & down to see what the effects of transposition would be. then you just drag whatever selections tickle your fancy right into your waiting audio software. lovely.
every time i do this it puts a smile on my face. because i’m old enough to remember how much time and effort used to be involved every time you thought to yourself, “you know what this SFX session needs? a little bit of metal rattle, about two seconds worth, right here on track 2…” i bet that used to take me about 15 minutes. now i can do it in about 15 seconds. freeing up lots of time to pen this dreck and refresh facebook.
and thar she blows…our favorite SFX DB search return: “horse fart.wav”
like pulling the fire alarm in school, but better
the plan for today was: map out the forthcoming milestone, do a little postmortem assessment of how the last one turned out, shoot a lot of shit about the state of audio tech in the game, get through a meeting or two, then work through the night on music or something fun while the rest of the world sleeps.
the reality of today was: the power went out in the building early in the afternoon. we stepped outside to play some ball in the gorgeous, humidity-free, summer-in-april weather. we got word that the power wasn’t coming back on anytime soon and the buliding would be locked until the next morning. with heads hung low (ha), we reluctantly (ha!) convened at a favorite watering hole well *before* happy hour. the next six hours were spent in an adirondack chair at “the beach” – the secluded, sand-covered, outdoor section of said watering hole – with the sun pressing me further and further into that chair until nightfall, at which point i had to be forcibly peeled from it.
adopt, adapt, improve…right?
hands at ten and two o’clock
during my morning commute i was thinking about thinking during my morning commute. since i live 100 miles from the office, i’ve got that kind of time.
used to be, i lived 10 minutes from the office. but that was a different office, at a different studio, in a different town. i regularly rode my bike and rejoiced in doing so (except when it started raining halfway through the day). but the time came when i took a new job, and came to live 60 miles from the new office in a new town. bike riding was out of the question, and i figured it was just cosmic yin & yang, evening things up for all those years of easy-commutin’. at least my new job afforded me the luxury of driving off-rush-hour, and the trip was through beautiful farming countryside nearly the whole way. it was 60 miles in as many minutes. plenty of my co-workers drove longer to cover less ground, so i didn’t complain.
it was during this time that i began to master the art of death-wish-multi-tasking. well, let me back up a sec. in my past life as an amateur musician, i logged literally a couple hundred thousand miles of highway driving by the end of my 20′s. so i’ve long since mastered the art of driving with one knee while eating / writing / phoning / typing / drumming with my free hands. cast the first stone if you must, but i’ve now surpassed 400,000 miles of highway driving without incident.
anyway, during the years of my jaunty little 60 mile commute, it got to the point where i was doing a couple hours of solid work in the car every day. driving with my left hand while using my right to plunk out spreadsheet & doc work on a laptop, or work through melodies & chord progressions on a keyboard, i usually got a leg up on the day’s work before i punched in. and if i still had it in me, on the way home i’d wrap up any business left unfinished when i punched out.
and now it’s up to 100 miles. but i still drive off-rush-hour, and my drive is even more serene than before – comprised nearly entirely of the rolling hills and farms of virginia, then west virginia, then virginia again, then maryland. (also, i don’t have to make the trip every day, which contributes greatly to the bearableness of things.) if the trip weren’t one big 5th gear open road free-for-all, i wouldn’t be able to use it as my Time Of Reflection. it serves as the “chewing on the end of a pencil and staring at the ceiling tiles, deep in thought” part of my day. so once in office, i can spend most of my time *doing*, executing whatever my brain has worked out while flying obliviously over all those state lines.
if all this sounds like absurdist rationalization for coping with a ludicrous situation, well hush. it’s *working*, so don’t ruin it for me.
in celebration of middleware
being a bit reflective in the wake of my recent birthday, i’ve been looking back and taking stock. not about the choices i’ve made in this life or anything heady like that. but at stuff like this.
what the hell is that, you ask? that, my friends, is audio implementation in the era before FMOD or Wwise. that…is Windows “notepad” as middleware. this is the file that controlled all the audio in Kesmai’s BattleTech 3025, our only “tool” for talking to the Miles Sound System underneath it.
go ahead, read through it and try and make some sense of it. it took me a minute too, and i came up with the damn thing (though it *was* a decade ago). let’s see what high-powered functionality we managed to drive with our text editor…
- various sound object definitions: single sounds, random sound groups, and start-loop-stops.
- sound object parameters that could have discrete values or value ranges set, including: sample rate, volume, pan, and offset time.
- attenuation distances.
- priority groups, and their sound cutoff behaviors (new overrides old vs. old trumps new).
- meta-sounds that could reference one or more of the above and override their parameters (ie crude scripting).
- a change log! haha.
of course all this stuff is in any decent audio middleware package these days. but i thought it was pretty spiffy at the time. DIY, BITCHES!
keep it simple, stupid
it all started off innocently enough, with the TheAudioProgrammer saying, “hey i’ve got a little time in the next day or two…you got any odds and ends you need me to wrap up?” after a little thought i replied, “yeah…maybe some UI stuff i’ve been meaning to get in…the absence of which keeps getting called out in the playtest feedback…let me mull it over for a minute and i’ll shoot you an email…”
so i start to write the email. which prompts me to jump in game to check out a couple things to make sure i know what i’m asking for. which prompts me to drag TheSoundDesigner into the game to assist me in this endeavor. which leads to 30 minutes of playtesting. which turns up a number of related UI things that need some lovin’ and should be mentioned in the email. which prompts me to open my UI audio design doc and start rifling through the half-baked material therein, revise it, and add my new findings. which forces me to update the layout of the whole doc. which means i need to update the corresponding elements in the audio project files themselves, and notify TheAudioProgrammer of this via my ever-lengthening email. which leads to consolidating a bunch of one-off UI audio bugs, also to be included in the email.
and now it’s six hours later and i’m hitting send. and the email is…long. but i feel like i’ve really whipped the current crop of UI audio loose ends into good shape. all that remains to be seen is if TheAudioProgrammer is now going to take one look at the email and say, “fuck THAT. i said i have a *little* time, remember?”
be careful what you wish for
as i’ve made mention before, the sound system on my last game was…ah…less than robust. as such, the list of things we couldn’t do was longer than the list of things we could. and though that sucked by-and-large, it also kept things simple, albeit in a rather perverse way. we often didn’t even attempt to support Complicated Audio Feature X, because odds were Complicated Audio Feature X was well outside the capabilities of the tools at hand.
hey, what a time-saver! we don’t have to write up a design, interface with other teams, hire additional personnel, produce the assets, integrate the assets, iterate and troubleshoot the assets associated with Feature X! we can simply shrug and say, “sorry, Feature X is a non-starter”. huzzah! bullet dodged! defeat snatched from the jaws of victory! /sadface.
nowadays, i have quite the opposite problem. the combination of sound system, internal tools, and smartypants code monkeys means we can do *anything*. i can soar as high as my imagination dares! i can dig my grave as deep as my unfocussed vision will allow! so now…i gotta be careful.
i asked for – and received just this week (oooh i’m giddy as a schoolgirl) – the ability to edit animations directly. so now the audio team – not animators – is responsible for placing audio-triggering tags on specific frames, auditioning them with sound in realtime, scrubbing through the anim, soloing up tags, etc. so now a given “idle” anim for a given critter can be laden (or overladen) with numerous “step”, “shuffle”, “huff”, “growl”, “limb”, “everythingbutthekitchensink” tags. henceforth as that sucker idles, all sorts of lovely, subtle racket can be heard. awesome. it’s what i’ve always wanted.
and i’m sure animators are as joyous about not doing this work as i am about taking it on. no more will they have to hear me going, “hey animator guy, i need a ‘jump’ tag here…(a week later)…no, i need to you scoot it back a couple frames…(a week later)…no, wait, scoot it *up* a couple frames…(a week later)…no, wait, can you change it from ‘jump’ to ‘leap’?…(a week later)…hrlghhrlghhrlgh (the sound of my windpipe being crushed by the hand of an exasperated animator)”.
but now i have to step back for a moment and wrestle with the Big Picture. how far down this rabbit hole are we going to go? how many anims are deserving of this level of audio detail? which ones? what’s the list of supported tags going to look like? who’s going to create all the sounds for those tags? and *when*? what’s the pipeline between the animators & audio going to be like, so we’re in constant sync regarding animation version control? i was looking at the animation list for a typical character today – the loooooooong animation list – and it brought all this into stark relief. hey, you asked for it. you got it. deal with it.
still. i *am* giddy as a schoolgirl…cuz it really is so suh-weet. and i’ve been waiting for it for oh-so-long.
hey, funny coincidence. “option paralysis” is not only what i’m in the throes of, but it’s the new record from Dillinger Escape Plan, whom i saw last night. well…i didn’t so much “see” them, as i was rendered senseless by them. this morning i was nursing a mild case of post-traumatic shellshock…half deaf and replying to coworkers with only blank stares. but by lunch it had worn off and i was gleefully building an all DEP playlist…
voiceover is hard, part 37
i couldn’t have said it better myself.
well, maybe i could have tried, but it would have taken a great deal more effort to do so, and it’s much easier to just link to someone else’s strong work and go, “yeah! what he said!” i have a bit of experience with what he said, having been involved in voiceover production for a so-called “triple A” MMO. and everything he says rings true. tragically, depressingly true. getting VO into a game of that size typically follows this sequence of potential pitfalls…
1) a script is written. and when i say “script”, i mean a collection of lines & phrases pulled together from hundreds if not thousands of seperate moments of gameplay that may or may not be interconnected. and when i say “written”, i mean a collection of words & punctuation concatenated in some more-or-less readable fashion, often by people with little training as writers. so the script ends up being tens of thousands of fragments of dialog, penned by dozens of non-writers, seperated by team divisions, over the course of years. sure, it’ll all feel like an organic whole. and we’re off!
2) the actors are cast. but due to budgetary contrainsts and the basic laws of physics, there cannot be a one-to-one correspondence between actors and speaking roles. so the talent search becomes complicated as it becomes more of a ten-to-one correspondence, with each actor being responsible for as many roles / voices as you can get out of him or her. this of course means you want to find actors that not only have a great voice, but have ten great voices. and for the love of god, this means no more “in-house talent” (read: co-workers who are up for anything). stop it. we’re done with that.
3) the actors perform the script. which usually amounts to a lone figure in a booth reading hundreds of lines once or maybe twice, straight down the page. this proceeds for hours, with only brief pauses between lines for a sip of earl grey or to look over the exhaustive (one word) stage direction for the next line. “angry”. “happy”. “purple”. so much for in-depth character study and the proverbial “what’s my motivation?” no time for that. we’ve got 12,000 lines to get through by 5pm. hustle, hustle! it *would* be nice if we had more time to spend on each line. it *would* be nice if the actors actually got to interact with one another. someday, maybe. but even so / in the meantime…mad props to the actors who manage to make the dialog work 95% of the time, despite all the limitations.
4) the audio editors edit the VO recordings. no matter how awesome batch processing is these days – and it’s awesome – some of that stuff just has to be done by hand. truncating files, chopping out bits you don’t need, needledicking every last mouth click and breath noise, etc. and once you introduce the possibility of human error into the production of 50,000 audio files…stuff happens.
5) the VO gets integrated. if you’ve really got your shit together, you can just drop the files in and the magic will happen, right? you won’t have any instances of multiple lines simultaneously coming out of a single mouth, stepping on each other. you won’t have different actors voicing lines spoken by a single character. and each actor will have read every line for a given character in *exactly* the same voice character, maintaining perfect consistency across those hundreds of lines, perhaps even spanning multiple recording sessions and re-do / pick-up takes. and of course all the VO will match the onscreen text perfectly. yep. what could go wrong?
going by the above, it’s a wonder *all* game VO doesn’t suck eggs. and it’s a testament to the professionalism of everybody involved with those games with great voiceover.
stay in school. don’t do drugs. (part 3)
i believe the children are our future. teach them well and, let them lead the way.
whelp, Career Day has come and gone, and i think i managed to aquit myself with a minimum of embarassment. i did my 30 minute “what it’s like and what it takes to work in the video game industry” spiel six times in a row today, to as many 3rd grade classes at my neighborhood elementary school. some of the highlights:
my presentation included a video montage of Warhammer gameplay footage (edited-for-content), and some behind the scenes footage of music-making in Prague, VO in LA, etc. i figured i’d better run it by the school admins ahead of time, lest the game footage prove too violent for general nine-year-old consumption. i got an email back from the principal saying she enjoyed it and thought it was fine. damn. i should’ve left more violence in.
the boyfriend of one of the 3rd grade teachers came in for one of the sessions. the teacher said he was a big RPG’er, and would be psyched if he was permitted to come hang out. the more the merrier!
my first presentation began with the video not working. the teacher was assuming *i’d* know how to run her A/V equipment. i was assuming she’d deal with it. we were both eventually saved by the school A/V dude. go A/V dude!
after three hours of hearing myself prattle on about my life story, i never wanted to hear myself speak about anything ever again.
the kids questions were awesome. as expected, i got lots of, “did you make the sounds for <game X>?” sadly, no. hi kids! disclaimer: unless you’re a hardcore MMO or RPG player, you’ve never heard of the games i’ve worked on. does that make me less awesome in your eyes?
also, “do you like Metallica?” why do you ask? “because of your tattoos and rings.” oh. um. well, see back in 1986 when they were at the peak of their creative powers…oh nevermind.
also, “do you like Dragonforce?“. i shit you not.
also, “what was the first video game ever?” at which point i drew Pong on the whiteboard. which i enjoyed immensely.
also, lots of “what are you working on now?” sorry kids. i can’t tell you that. “seriously?” seriously. for grownup silly reasons.
also, after explaining how foley is the act of recording the real thing itself, “how do you make the sound of someone dying?” well kid, we go out an kill somebody. i actually said that with a smile, followed quickly with a j/k. still, i don’t think teacher found it all that funny.
at the end of the day, the whole experience was a hoot. dunno if i sparked the minds of any future sound designers out there, but based on their enthusiam it’s a safe bet that 95% of them have a future as gamers. and let’s hear it for teachers! how they stand up in front of all those little buggers all day and get them to focus on anything is a mystery. somebody should be paying them a lot more money.
