No, not from me, from Alternativa.
Their new demo is just brilliant!
http://blog.alternativaplatform.com/...9/04/democity/
Printable View
No, not from me, from Alternativa.
Their new demo is just brilliant!
http://blog.alternativaplatform.com/...9/04/democity/
well no its not impressing me anymore- its basicly what they showed before already.
I guess when seeing many of the 3d engines I get used to it and only pay attention to the ones that actually invented something new or different.
I've got to agree, it's not got the same wow value that those demo's did first time around, although it's still a lovely thing.
I wonder if F10 will allow for a Sega Rally clone...
Squize.
like for the others, maybe even more, there´s zero wow factor there for me.
Well, i´ve said it before, nothing realtime 3d made in flash can really wow me cause it only could if i´d put my "only see what´s possible in flash and ignore actual technical evolution in the last 10 years" googles on.
So to me flash 3d stuff can only be really nice if it features a really creative/fun implementation/usage.
I also wonder how many companies would actually use a flash 3d engine for something like this. I´d guess only around 50-100 and even those only until they notice their result looks 10 years old compared to what´s possible in other tech.
To me the eco zoo is still the nicest use of realtime 3d in flash, original, fun to use and just lovely, this in comparison just feels sad to me.
I understand them trying to pimp the engines being used for something like that, but seriously, this looks awful.
interesting responses...
The main thing I don't like about their engine ( Perhaps the only thing looking from it as just an end user point of view rather than as a developer ) is the way they change the quality when it's stationary.
I'd rather it always stayed in low-quality rather than it changing like that. It's like "Here's how it should look if you weren't busy moving". You get the whole jarring pixel move with it, and it makes you aware that the quality could be better.
Squize.
yeah,i totally agree. To me it already looks barely acceptable in the "high detail/quality" version (super low res textures, ultra low poly models, lack of any modern more quality/detail impression giving effect like bump/normal mapping effect, lack of any modern post processing effects etc) but just as Squize said, seeing the static image in their high version just makes the low version feel even way more ugly.
to Cortana: I´m sure many flash view focussed people will praise this demo as if it is the next huge achievment, but yeah, one has to be realistic, most people who have seen any 3d computer graphics in the last 10 years and don´t care whether they are made in flash would think this looks pretty rubbish.
With the state the flash player is in right now it just doesn´t make much sense to go for realistic looking realtime 3d, it will only look/run rubbish compared to other things.
Although if you reveal the triangulation, I find it interesting that the polygons shift around so that the textures get stretched as little as possible. Is that new?
No,its not particularly new, its what the other 3d engines had as one of the main "selling points" when compared to pv3d for a while now.
I agree as well the freezing effect annoid me as well- (not just in this demo)- i´d rather have it look jaggy and pixelated as long as I have the feeling that its constant.
Imo a noise cam would be nice - so it feels more natural, just slightly.
have you guys actually used the engine though?
those features can be enabled / disabled. Alternativa engine is probably the best solution for a game developer in flash atm. Imo The API is 1000* better than bloated papervision.
I highly suggest that anyone deciding to develop a 3D game in flash go download the new 5.0.4 source right now. They have developed a flash 10 specific version to :D.
As to the demo, its ok I guess, looks like there is some mouse event system finally in place.
already started a another new one myself: rendering real polygons and billboards instead of triangles ;)
i guess our comments weren´t really directed to the way alternativa is structured but rather how their own demos are presented. It would only be logical that one can disable that behaviour in code.
I liked your experments with Duke Nukem as sprites walking all over the place. And if i´d really had to stick my nose in someone else engine it would be propably rather be alternativa or the one from the japanese (ecodazoo, pictabs, ...).
Papervision is used by to many and makes it less attractive for someone like me (almost a must to avoid ^^)
ah thanks :D
good luck on you're new engine, is that number 3 now? ;)
papervision is the devil. Sure its got a nice feature set, but the API blows.
Well this wasn't a game related demo and I don't think we have nearly enough polygons or processing power available to even start doing fully 3D realistic looking scenery. That's why this looks like their worst demo yet because you just can't get realism with those texture sizes and polygon counts.
Now that we have some 3D technology, we just need some really creative people to go at it, like the Eco Zoo guys.
Those guys get what the strengths and weaknesses of low end 3D is and have very good designers.
is that a requirement? - if you dont have the power design something that fits the power- I think there is plenty of space to design something with such a engine.Quote:
Originally Posted by hatu
And agree its the creative minds often like ecodazoo that combine it perfectly and make it much much more impressive that just a tech- demo.
What about licensing issues? Alternativa3D seems more restrictive (source is closed AFAIK). Is this true? If so, does the cleaner API justify the additional cost of licensing? With free-er alternatives, Alternativa3D will struggle IMO. This isn't really a knock on their quality - I still think it's a best looking 3D flash platform around.Quote:
Originally Posted by mr_malee
I think the demo is pretty neat. The only thing I would change is the transition to the skybox. Where the ground of the city means the ground of the skybox is a jarring transition (one scrolls, the other doesn't). This can be solved by getting rid of the bottom panel of the skybox and using a Mode7 approach to extend a (scrolling) texture to the horizon. Then I think it would be easier to blend to the vertical walls of the skybox, you could make both more or less the same colors and it would be hard to notice the scrolling/non-scrolling transition.
about demo, it almost sucks. the biggest turnoff for me is that it runs anything but smooth. if the engine cannot support X polygons, than demo featuring >= X polygons is not the way to pimp the engine. I was working for some time on marriage between alternativa and Tim Knip's quake parser, but it doesnt run smooth so I dont go around pimping it like, "look what flash can do", because it cant.
about engine, it is nice, and I definitely see possible uses, but it lacks many things at the moment, so I personally would at least wait for 5.1... and surely play with 5.0.4 in the meantime :)
Ok how about this demo? I never cease to be amazed by what these Russian guys do...
http://blog.alternativaplatform.com/.../inneryard.swf
Its amazing. But the textures really let this one down. I remember in the other demos the blur setting in when you stop moving but for some reason it looks like it does two passes. Water looks cool though :)
but still it doesn't feel fresh to me anymore. The mirror thingy is nice but almost a side effect in the demo. The scene is very tiny and the textures are very low res - with the given default resolution I get close to 15 fps on a Athlon 64 3500+ processor - so its lagging and freezing with the default auto-quality setting. Only with 1/2 of the screen size and auto-quality turned off I get 20-30 fps and a smooth feeling.
oh well I guess my real disappointment is rather the general pace everything took now with flash and 3d - it seems that many of the key- archivements in perspective,texture mapped engines is already done by the major engines. Its not the same bang anymore as with the first times one could see the nice demos showing speed, gimmicks and sometime a sense of humor. But when things keep repeating it becomes boring or average in expectation.
recently I stumpled upon (thx to squize) on a free voxel editor similar to msPaint but with voxels in 3d space
http://forums.revora.net/?showtopic=45798
and some maxscripts to convert existing 3d polygon objects into voxels
http://www.geocities.com/cnc_central/utilities.html
now that would be interesting to use in Flash imo. esspecially if used just for objects only or other fresh mixes.
It is a lot better than the last one, as it plays more to the strength of the engine.
I could see a game being made out of that as it stands, a point and click or an escape the room. The textures aren't great, but that's just a file size thing ( I never really care about textures, they can always be made to look better, it's how the engine handles is what's key ).
Voxels do have quite a bit of potential, there's a port of Wolfenstien done using a voxel engine which looks really cool.
Squize.
I don´t like this alternativa demo either. Not sure what you mean with it plays more to the strength of the engine Squize, is it that it handles a smaller scenery than the last one?
To me its just no good idea that they try to create a sorta realistic look, going for a realistic look will always look sucky until flash´s capabilities are raised a lot because the audience has an idea for how the things should actually look so then it looks really bad in comparison.
I disagree with what you said regarding the textures: Its not just a filesize thing and they don´t just look sucky because its ugly textures. It looks sucky because they use textures with very low dimensions and they don´t only do that for saving on filesize but mainly for saving on cpu and ram usage (which would be way higher when pushing around larger dimensioned images)
and then they scale those up to fill large areas with em. And this has a lot of impact on the sucky look of the textures. So yeah, one could make it run nicer with higher res and bigger dimensioned textures but then it would run even way slower.
Overall to me it would be much nicer if they went for a non realistic look and then rather use texture dimensions which fit to the scale the objects are shown in.
hey, looking at it again now: should this be a remake of an area in a game? gears of war? well, can´t sort it in now, but yeah, it looks so much like the pre snes 3d brother of it that it took me looking at it several times till i even realized knowing the place from a game.
"is it that it handles a smaller scenery than the last one?"
Yep pretty much. The last one looked poor because it was mainly a skybox, where this is nicely tight and only shows what looks good rather than filling the screen with content ( Such as a skybox ) which will never look that great.
It's like the bunker demo they did, that was really good.
"It looks sucky because they use textures with very low dimensions and they don´t only do that for saving on filesize but mainly for saving on cpu and ram usage (which would be way higher when pushing around larger dimensioned images)"
I wasn't sure if there was a direct correlation between texture size and pixels plotted, in that if you bitmap fill a rectangle 100x100 I didn't think it would take twice as long to fill that with a 200x200 texture ( As the source ) than a 100x100 texture.
I know it will be more costly in terms of cpu ( Memory isn't really a worry ) but I didn't think it would be that great a cost per pixel, that's why I assumed that there was more left in the engine that this example [ In terms of texture quality ].
In a real life example you wouldn't have it running full browser window size anyway, so as an example it's a bit of a worst case scenario.
Don't get me wrong, it doesn't make me damp, but I think it shows a lot of potential. It's just finding a use for that potential.
Squize.
its just for a typical game level it would need ~ 100 times more geometry and bigger environments.
I guess I already wrote that before but all major engines so far rather specialized on rendering unique (texture baced, non symetrical/ instanced,...) objects. That's nice for product placement but will suffer from more complex interaction or scenarios.
I personally though that with the screen FPS rates we saw so far in demos it would be possible to create a 3d racing game similar to destruction derby 1 & 2 but what makes most of them useless it the way scnery is handles- or I simply haven't seen a demo that does that well. Because all I have seen were rater some smaler scenes or objects.