Messages

If you need to add inertia to make a shmup fun then it doesn't have enough bullets.  Case closed 

I was quite enjoying it myself, even though I didn't really have a clue what was going on story wise and I was a bit overwhelmed by the sheer number of gadgets and weapons and stuff so didn't really use any of them except for the basics.  I'll admit though that I only got to act two myself, started playing something else (Civ Rev which is amazing), and haven't gone back to it yet.  Then again that says more about me than MGS4 I think.  I love shiny new games and just cant resist delving into fresh ones before I've finished playing the ones I've got.  My game stack is stupid big and there's hardly any I've actualy finished.  That just inspired me to do a count.  I've got 86 console games in the piles in the front room at the moment spread out across PS3, PS2, DC, DS, and PSP, and out of those I've completed 4 

PS:  Then again 25 of those are shmups which I dont consider complete till I've 1cc'd them which is well out of my grasp for almost all of them so that skews the statistics a bit .

That looks pretty cool.  Never been to a demo party myself so I'll probably try and make it along.  Depends on being able to get some time off work though since it's the other end of the country.

Indeed, dont get me wrong I have a lot of respect for people doing the old school too.  I guess the point I was trying to make not very successfuly is that the modern scene hasn't seen a reduction in skill required, just that different skills are required.  That of course is as long as your trying to push the boundaries anyway, but those boundaries have changed too.  The hardware is now less of a limiting factor and instead you're coming up against the boundaries of knowledge.  Anyway I dont want this to sound like I'm making myself out to be some kind of guru or anything, I'm nowhere near pushing any boundaries, I'm still trying to churn out the stuff that's been done a hundred times before.  But then that's why I'm here, because I'm trying to learn rendering techniques.

Maybe try telling Stonemonkey that using a modern API requires more skill than software rendering.

Well now you're just putting words in my mouth  .  Please dont get defensive, I'm not attacking anyone, I just find this an interesting debate.  Anyway my point was that the advanced stuff you do is stuff beyond what the API supplies.  The API itself is just a tool to quickly get you to that level.  Creating that new advanced shader or creating a software renderer both take skill, that is my point.  This also kind of leads onto what I was saying about changing skill sets.  Creating a software renderer (to the old school level anyway) would require maths and hardware knowledge, where as creating a new advanced shader is more focussed on maths and scientific knowledge.  You also dont have the advantage of walking a well trodden path, you're trying to do something new.

@rain_storm: You seem to imply that API's have made it all easy and thus old school demos required more skill.  I dont think this is the case.  The API's are tools which can be used as stepping stones to work on more advanced techniques.  So yes, if all you want to make is the same spinning cube as seen in a retro demo then that requires much less skill now using a modern API.  However using the API's lets you skip ahead so you can get a head start with the spinning cube but then concentrate your efforts on creating some new global lighting pixel shader technique to render that cube.

I think some of what you say rings true stormbringer but in my opinion there's a flip side to this which you've missed.  I'm going to play devils advocate here so please dont be offended, I'm just throwing out ideas. 

Ok, so in short, perhaps the reason you're not pleased with the new scene is because you're still hanging onto the concepts of the old scene which long ago evolved and left you behind, you just haven't evolved with it?  I'll address some of your points to try and make what I'm saying clearer.  You say that people dont feel the need to get the most out of their machines.  Well I dont believe this is true, hardware has always been growing and expanding at this rate, in that respect it's no different now than it was back in the c64 days.  Today there's still an upper limit to the technology you can get commercialy, and that people dont always have the most up to date tech at home is no different than it was back then, not everyone upgraded to the more powerful versions of the BBC Micro or Amiga or bought the ram upgrades you could get.  The spectrum of possible hardware and configurations is wider now than in the past but there was still a variation then all the same and it didn't stop people from wanting to push the limits and instead just resign to getting better hardware. 

What is different is that finding some new low-level hardware programming trick on the old machines could make a large difference since the power of those machines was so limited.  Although the increase you got might have only been small, it was a significant percentage of the total available performance.  Now however such tricks are mostly just drops in the ocean.  Therefore the modern scene evolved and the focus more on algorithmic, artistic, and scientific improvements.  The first two should be pretty obvious what I mean, but by scientific I mean advancing current models or developing new ones based on real world science to create more realistic effects such as better lighting models for example.  A lot of skill is needed first to understand such things but then also to come up with sensible approximations and optimisations which when combined with hardware knowledge and algorithmic skill allow them to be simulated in realtime.  It's true that NVidia as well as the Microsoft graphics research team often make the first steps in these areas, but they aren't dictating to you what you can or cant do with the hardware.  This is only the case if you allow yourself to be spoonfed by them and shut of your mind to new possibilities.  There's nothing stopping you from going out and creating new models to simulate effects that haven't been done yet, or to imprive existing models, except for your own skill.  This is probably where a lot of the old sceners get left behind, because their skill sets are focussed heavily on hardware and not mathimatics or science which is now increasingly important to stay ahead of the field.

Finally with regards to DirectX, I feel it has it's place in the new scene.  Since like discussed the focus has shifted from low level hardware to scientifically complex effects which often require much experimentation.  In these conditions productivity is more important than absolute efficiency and directX is the most comprehensive graphics development library out there.


PS:  Just a reminder I'm not attacking you or trying to demean your skill.  Just playing devils advocate to put forward an alternative point of view  .  All discussions welcome, please pick it apart .

Congrats everyone, cool competition.

5 Star for Assassins Creed?  I thought the tech was absolutely amazing, they just forgot to put an actual game in there.  There's like 4 extremely simple objectives which you just repeat over and over and over and.....  There's a very dissapointingly little amount of actual assassination too.

That's actualy really cool, especialy when it hits a coner.  I love how you've managed to make something funky even though you restricted yourself to only having one square.

Its looking pretty cool so far.  One thing that destroys the believability of this being a sketch though is that the lines are all very straight which rarely happens when you actualy sketch something for real.  You might want to use perlin noise to add irregularities to your line drawing as described here.

Thanks for the comments.  Indeed it took a lot of tweaking and iterations to get them to start looking natural.  What I finally settled on for the demo was to layer several iso splines ontop of each other for each branch and slightly perturb the node positions using a gaussian distribution.  This gave the branches a more natural shape rather than looking like pipes.  I also used combinations of gaussian distributions and white noise the generate the positioning and curvature of the branches and used repulsion forces to seperate the branches out.  It took a lot of tweaking but I think it was worth it  .

Really nice effect, congrats.

Out of interest benny and voltage, what are your computer specs.  That might give me a hint to what's going on.

It behaves a bit strangely on my dual-core at work.
Most of the time it just hangs after "loading", sometimes it crashes, sometimes it works.
Probably some threading-issue?

Very possible.  The program multithreads and I wasn't able to test it on a multi-core machine.  My machine at work is quad core so I could possibly try to fix it, but I dont think I'll find the time .

Cracking stuff, I like the shading gradients.

Says I need d3dx9_36.dll

Try installing the latest version of DirectX 9, hopefully that will help.

Sweet, glad it's working for you all.  Yeah it needs some serious optimising.  There's some tricks I could do such as partitioning the space to cull chunks away from the iso surface generator for example.  I'd definately have to spend some time on that if I wanted to include something like this in a full demo.  A nice thing though is that the generation runs in a seperate thread so I could hide the load times with other less intensive stuff in the demo such as the greets.

Well I'm glad it works for someone  .  Anyway Hellfire, I've changed the mesh to be dynamic which might fix the issue for you.  I've also changed the error reporting code so it should hopefully now spit out a DX error code if anyone gets any problems with DX stuff like that.

Now that's strange.  That means you got passed the DX mesh creation ok which implies you can run DX stuff.  I cant see why it would fail to lock the vertex buffer for you though if it gets that far, I dont really see how that can be dependant upon someone's machine.  Then again I'm no DX expert.