h a l f b a k e r y
Think of it as a spell checker that insults you, as well.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
or get an account
I hate computer games in which you have to kill thousands of bad
dudes who all look like one type or another. There's such a limited
variety of textured models that it saps you of all your murderous joy
and staples your soul to the numbing reality that is your pathetic
computer game-playing life.
Samuel Beckett would have loved it.
Escapism, that's what I'm after. I want to kill these people on my
screen because, in actuality, I'm quite a nice guy; calm-tempered,
slow to react to violence, with no inclinations towards homicide.
Gruesome massacre is not a viable option in my career I work on a
business magazine correcting people's speeling. A world in which I
can lop off someone's head with a spade is an alien one to me. A
vast multitude of gruesome experiences that I'll never have the
(mis)fortune to experience in reality. No, escapism is the ticket.
Would it not be possible to randomly generate models of bad dudes,
perhaps using a hue change on their clothing, skin colour, hair and
eyes? The game could adjust the face and body shapes, using similar
technology to the player modifications in Tiger Woods or Pro Evo. A
range of clothing models could also be chosen from.
Particular colour ranges could be used for each specific set of bad
dudes, to ensure their authenticity. Otherwise, the whole illusion
would be dashed if a bunch of harlequin-coloured mafiosos came
charging at you. Or one pink-wearing Nazi homophobe in a pack of
otherwise neat skin-heads.
The hardest thing to randomise would be voices (which in days gone
by have been changed only in pitch, which just sounds obviously
contrived) and body movement.
However, for this latter problem, I have the concept of a solution. My
friend is an actor, he trained at Llamda in London, and he was telling
me how everyone walks from a particular part of their body. A bit like
the Donnie Darko tubes of fate that emanate from people's chests, if
you imagine that coming out of other parts of your body that pull you
forward, that's how you can replicate somebody else's walk. Some
walk from their knees, others, their foreheads, elbows, whatever.
(Apparently I walk from the crotch, who knew?)
Each walk is an impressive indictment of character as the focal point
relates to the rest of the body and affects your stature immeasurably.
Walking from the forehead makes you look glum or reserved, from the
chest makes you look proud or arrogant, knees - relaxed, elbows -
cockney, you get the picture.
Using the (albeit simple) rag-doll technology [linky] could pre-recorded
motion capture be modified by pulling the nodes from a focal point?
For example, leading with the shoulders would give a stockier feel to
a character, and would only require the nodes of the body be accentuated at the top quarter of the body.
I've no ideas towards the voice problem I'm afraid.
Bad Dudes vs Dragon Ninja
Two street fighters must battle their way past hordes of attacking enemies in their pursuit to defeat the evil Dragon Ninja and rescue president Ronald Reagan. Players must defeat end-of-level bosses, one of which is Karnov himself! [zen_tom, Oct 15 2009]
[theleopard, Oct 15 2009]
(?) Unique Enemies
Prior Art... sort of. [Jinbish, Oct 15 2009]
||Why should the system not generate Nazi homophobes wearing pink? Is that the single group of people you wouldn't murder?
||No, I didn't think Nazi homophobes would wear pink.
||They might if they were ironic, post-modern Nazi homophobes, but to be fair, there aren't that many about.
||The only effeminate Nazis I can think of were from Allo Allo.
||Er... What about: "Springtime for Hitler: A Gay Romp with Adolf and Eva at Berchtesgaden"?
||Schlagen, heir [Jinbish].
||If I'm not mistaken, in games like Oblivion and Fallout 3, the random encounters are all randomly generated baddies -- also in games like Dynasty warriors - I believe the faces of the enemies are all randomly generated.
||Not to the extent you're talking about.
||As for the voices, you could get, say, 50 people to say the entire phonetic alphabet, then have all the words be constructions of phonemes, and draw from one set randomly. That way, you could give each voice different accents as well -- for instance, out of phonemes, you could instruct the same voice to say, "Whale, hah theyer y'all!" or "Well, hello everyone." Each word would be a phoneme map for the computer to fill from a specific voice set.
||pitch adjustments could be used to simulate articulation -- final words having downward tones, question words trailing upwards, etc.
||It's complicated, sure, but it's totally doable.
||The voices could be done by phoneme recordings, but it'd be easier to just have fifty people (or however many) record the various stock phrases.
||You could also just record things once (either phonemes or clips) and generate new voices by changing the locations of the formants (perhaps by a process similar to additive synthesis) - but then, if you had this technology in place, why not extend it to the player character? Have the player speak some, analyze his cadence and harmonic content, and then re-filter catch phrases and dialogue...
||I was thinking that the phoneme method would save on space. I like the option for the player as well -- just have him read the phonetic alphabet, or a series of words that contain all the sounds.
||What do you think of the randomised animation adjustments. Slightly
to one side might produce a limp/swagger?
||//What do you think of the randomised animation adjustments//
||I'm sad I can only bun once :-)