Jump to content

Phaser 2.5 Roadmap (and request for ideas)


rich
 Share

Recommended Posts

I think it might be good to implement placeholder assets or an alternative small one (SVG outline in my case as I can also easily color it without expensive bitmap operations) only to be replaced when real asset is loaded. This could allow people to let download mandatory base assets and optionally lazy load others.

 

And although I know it is out of a 2D oriented engine's scope, I'd really appreciate a way to apply texture to a geometry (which is easy at webGL but not at canvas if you seek fallback like me). Actually I consider switching to an engine with canvas fallback for a "map" application. It will have prerendered chunks of terrain and sprites of buildings on top of it. Only thing stopping me is rendering assets when they have to fit into tile geometry (like green tile in example below)

8XrWlg5.jpg

Link to comment
Share on other sites

9 minutes ago, rich said:

We've already got placeholder assets in there, and they're easy to switch out for your own.

As for a 3D mapped terrain, it's not going to happen, sorry. I suggest you swap to a proper 3D engine for it.

My bad, hadn't notice but glad to hear its already there

and for terrain, I wasn't expecting as I stated, it will be my "pleasure" to handle it somehow. And whole point is not using a 3D engine for a mere map (like Google Maps lite version) and I am not big fan of using three.js with canvas fallback just for this.

Link to comment
Share on other sites

12 minutes ago, rich said:

You could certainly code your own, yes, but it makes sense to me to use a framework that can actually handle this. PlayCanvas perhaps?

Don't want to pollute this thread but, I consider Babylon.js atm for phase 2 (aka more than mere map) , but for phase 1 any framework with canvas fallback is apparently enough when this minor detail is handled.

Link to comment
Share on other sites

Awesome news regarding the new tilemap renderer! Very excited to see it. 

As a sidenote, the original is still pretty good - I get near perfect framerate on an 2012 Macpro. Even with a sizeable map and going silly with rendering effects, and with no active "smarts" around handling the map size

 

 

Link to comment
Share on other sites

There has never been a native WebGL way to render Tilemaps, and outside of Canvas mode it has always been slower than it should be. I'm very pleased with the progress made on the new tilemap system though (you'll see commits to the dev repo this week with it in), and it has definitely made a significant improvement.

Link to comment
Share on other sites

Pixi V4 would be all that I felt mattered!

http://dev.goodboydigital.com/client/goodboy/million/

I know bunnies are hardly the be and end all of performance benchmarks, but Phaser + v4 = Lazer.  Obviously not quite (modular etc aside), but I feel Lazer should concern itself outside the renderer (and be TypeScript obvs but I will let it slide :P) . !
 

Link to comment
Share on other sites

Without diminishing what they've achieved, we were clocking 1m+ sprites over a year ago too. It's very GPU specific - the demo above renders nothing on mobile (as you'd expect), and crashes Chrome on my laptop with an integrated GPU. It's an edge case test though, so this isn't exactly representative or fair to Pixi, but still don't let it create a reality distortion field either - this won't create speed where speed isn't available.

This thread is about the best use of time spent on Phaser. Merging Pixi 4 appeals a lot, but will break a lot too. They're still not at a stable release either, which is worrying because it means if I use it, then I'm likely opening myself up for months and months of needing to stay on top of their updates, and there are only so many hours in the day sadly.

Yet it's hard to ignore the new shiny isn't it? The grass is always greener and all that jazz.

 

Link to comment
Share on other sites

1. If the Pixi API is going to change a lot, better to wait. -1

2. Yes, interpolated sprite rendering is smart. +1

3. Yes, current tilemap setup is kinda disappointing. It needs separate rendering layers, so it can handle more tiles. +1

4. Yes, sounds great. What the flip is SAT though? +1

5. Yes, It needs to happen in Phaser or Lazer. If modules are setup proper, more people will contribute. +1

6. I'm not really sure what a pool is used for in Phaser yet. -1

7. Yes, native Phaser text input would be great +1

 

Edit: >paying a very skilled developer to create a brand new WebGL Tilemap renderer in Phaser.
That's great to hear.

Link to comment
Share on other sites

15 hours ago, eddieone said:

1. If the Pixi API is going to change a lot, better to wait. -1

2. Yes, interpolated sprite rendering is smart. +1

3. Yes, current tilemap setup is kinda disappointing. It needs separate rendering layers, so it can handle more tiles. +1

4. Yes, sounds great. What the flip is SAT though? +1

5. Yes, It needs to happen in Phaser or Lazer. If modules are setup proper, more people will contribute. +1

6. I'm not really sure what a pool is used for in Phaser yet. -1

7. Yes, native Phaser text input would be great +1

 

Edit: >paying a very skilled developer to create a brand new WebGL Tilemap renderer in Phaser.
That's great to hear.

4. Separating axis theorem, a method of collision detection and handling: http://www.dyn4j.org/2010/01/sat/

6. Pools are used to reduce the number of instantiated objects in a game if they are regularly created and destroyed - destroyed objects are instead kept in the pool to be reused later when another needs creating.

Link to comment
Share on other sites

3. YES! I lost focus om my project (Metroid in HTML5) because a performance drop on mobiles when I finally managed to implement the whole world map leaving me with two choices: divide the map into smaller maps or try to implement arcade physics in the Tiled Plugin. A new renderer introduces a very attractive third alternative for me: "wait for someone else to fix it". :-) I guess this would improve a lot of tile-based games in a blow.
It would also be great to be able to set a single tile to "dirty" forcing it to be redrawn instead of complete layers (for animated or manipulatable tiles).

4. I haven't gotten around to test the "Phaser Arcade Slopes Plugin" which might be enough for me, otherwise it would be greatly appreciated (and for Lazer it would be somewhere in the top of new features I would love). In my specific Metroid-case I use 16x16 tiles graphically, but their bodies should consist of four 8x8 hitboxes. 

6. A Pool class would be great, especially if it's identical to Groups minus unwanted stuff so that it would be trivial to upgrade existing code using groups for pools.

I would appreciate anything that would improve performance. I guess most games is built within the current limitations of Phaser. I will not start anything new using Phaser, but will probably be an early adopter of Lazer eager to try out anything new.

Link to comment
Share on other sites

  • 3 weeks later...
On 5/27/2016 at 9:58 AM, nkholski said:

4. I haven't gotten around to test the "Phaser Arcade Slopes Plugin" which might be enough for me, otherwise it would be greatly appreciated (and for Lazer it would be somewhere in the top of new features I would love). In my specific Metroid-case I use 16x16 tiles graphically, but their bodies should consist of four 8x8 hitboxes. 

My plugin does the job for sloped tiles without curves, but that's it. It doesn't go further than tilemap collisions.

Interestingly, I haven't tested it at all with graphics larger/smaller than the bodies, but I'm guessing it would work fine. Feel free to post an issue on GitHub if you get round to tinkering with it and find any problems. I'll be resuming its development soon (and make it compatible with 2.4.8+).

Link to comment
Share on other sites

  • rich unpinned this topic
 Share

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...