Hey guys.

So eventually I'd like to have a platformer created using the Separating Axis Theorem.  Good news is I have it pretty much working, however the bad news is that I've seem to run into a minor issue that I don't know how to resolve.

Check out what I have so far here: http://fortheloss.org/misc/sat-problem/swf.swf
(To see the problem I'm talking about, move the player into the area under the triangle and above the bottom rectangle and move to the left. The player undergoes a quick elastic movement and is pushed into the wall as long as the left key is held down.)

I think what's going on is the player checks for a collision with each of the 4 polygons in a set order.  After checking against the other 3, the final collision check is against the triangle, which detects a collision and simply pushes the player out, into the rectangle wall below.  The player simply is nudged into the rectangle wall with no problem, as the collision check with that shape has already occurred that frame.  Then the same thing happens next frame, and the next, etc, etc - the player will appear to stay "stuck" in the wall until the left movement key is released.

Hope that was clear enough. Anyway, I don't really know how to go about fixing this.  The SAT algorithm simply checks against 2 shapes, it doesn't take into account other shapes in the area.

Any ideas appreciated!

@X-Tender - Yeah, "lots of loops" was probably the wrong way of saying it. What I meant is the  code would just be messier in general. For example, instead of checking if he's just on a wall tile, you also have to loop through the moving platforms and check those too, then if I add movable blocks, loop again to check those, etc, etc... Things like that.

Anyway thanks for the replies. I just wanted to check in that there isn't some easy "magic" way of handling moving platforms in a tile-based level.

Hey guys.

Just planning out my next game, but I'm having difficulty figuring out how to handle moving platforms when everything else uses tile-based collisions.

I'm certain that the way to go is to use tile-based detection for everything in the world except for the platforms, which will be rectangle-rectangle collision tests.

However, that just seems like a recipe for lots of for loops (and possible slow-downs) and generally messy code.

I guess what I'm asking is, what's the general consensus when it comes to handling moving platforms, specifically where everything else in the level uses tile-based collision testing?

Thanks!

[Description:]

Description:
  I've never made a topic here on TIG following the progress of one of my games, I usually just keep it to the weekly game update posts I write on my site, but I figured I'd give it a try this time around.  Who knows, maybe this will get the game more publicity :D

  Defective is a Flash game I started back in May (2011) and I'm coming close to wrapping up now in August - I'm looking at a mid-to-late October release (factoring in the time to get some music + find a sponsor).

  I was trying to go for a retro feel with the graphics and gameplay (with the classic 8-bit boops and beeps for sound effects).  The story / plot hasn't been revealed yet, but it's not too much of a big deal.  All you need to know is you're a small little ship-guy and all of these bad little ship-guys are trying to stop you from getting any further.

  The game features 12 levels (3 of them boss levels), 11 enemy types (not including boss-types), 5 different areas to upgrade (by finding upgrade crates in the levels), custom music, particles, and action-filled & fast gameplay - all at a smooth 40fps.  The main gameplay element that will set this apart from other games will be the ability for the player-ship to "steal" the souls of downed enemies.  After collecting 3 of a type, the player can transform into that enemy and utilize it's unique attack with a substantial performance boost.

Stuff:
  Here are a few screenshots from the game:

• Main Menu
• Level Select
• Gameplay 1
• Gameplay 2

  And here's the trailer I released yesterday:

• 
  
  

  I also did recorded myself working on game art for the game a few times:

• 
  
  (old art, unused)
• Drawing Yellow and Minion Enemies (old art, unused)
• Drawing All Enemies (decided to start the game over, go in a different art direction)
• Drawing the Minions (redrawing them, too, in the new art style)

Follow the progress:
My site: http://fortheloss.org
My twitter: http://twitter.com/ftlralph

Wow, nice post there, lol.  For some reason I never thought of combining the two methods (well, rectangle-based instead of SAT like you said).  Really not a bad idea.

But yeah, I remember attempting to get slopes working with a tile-based method in the past.  Never did get it to work right.  However, I'm hoping I'll have better luck this time.

Anyway, great suggestions.  Really appreciate it.

Hey guys.  I'm coming along nicely with my latest project (working in AS3 in Flash/FlashDevelop, though that's kind of irrelevant to the question I have).  It's going to be a platformer with the typical suspects - moving, jumping, ducking, slopes, enemies, etc...

I already generate the level and the objects in each of the worlds with arrays (vectors, specifically).  I'm getting to the point in development where I have to start moving off of working on the main core of the engine and instead start working on some actual gameplay (player movement) - here's where I start having questions.

What do you suggest I use for collision detection in the game?  I'm semi-new to both the tile-based and SAT (separating axis theorem) methods.  I mean, since I already have arrays holding tile positions, I might as well use the tile-based method, right?  However, I still haven't figured out how that would be applied for moving objects that are never completely in a single tile cell (enemies, moving platforms, etc).

That leads me to believe I should use the SAT method, right?  However, even though I understand the concept of it and have gotten it working (as in, it detects collisions) I never figured out how to resolve those collisions.

So both methods to me have their ups and downs.  I just want to get some opinions from you guys.  What do you suggest or recommend?  Have any links to any good tutorials of either method (excluding that "tony" tile site or the n+ SAT tutorial page)?

Really appreciate any input, and I apologize for the wall of text above.

I'm also a jack-of-all-trades - somewhat.

I have worked in Flash since I was about 15 (20 now).  At first I was 100% interested in animation (I still am, I just divide my time between that and game dev).  As a result, making art and animating sprites is very easy (and fun) for me.  Coding is also easy (mostly).  I know my way in and out of AS3 and have been majoring in Computer Science for about 1.5 years now so I'm also learning C++.

My weakest points?  Keeping focused.  I can work steadily and very productively for about 2 or 3 days, and then I'm suddenly "burnt out."  As a result, I have to force myself to get back to work on the project again.

Oh yeah and sfx and music.  I want to learn to use FL Studio one day.

Wow, that was an incredibly helpful post. 

I've always gotten stuck at this stage when I was developing platformers in the past.  I'm going to begin my biggest project very soon - a large platformer with very big levels. 

This will definitely be useful later on down the road, thanks!

Wow, I totally read that as BASS HEALTH and thought it was a game where you destroy your enemies by playing your guitar.

I may have to scrap this entire project and start it over again now.

Wow, tons of inspiring posts in this thread.

As for me, I've been working on a new game project in Flash for the past month or two.  It should be completed in about two weeks, including time for bug squishing and last minute polishing.