2D game development tool for android & flash?

Started by
28 comments, last by filya 10 years, 6 months ago

Thank you.

I agree, in that case, the YYC wouldn't help with the collision at all. So you think 'Aces of Luftwaffe' might not be doing pixel collision checks? It does have a lot of planes and bullets flying about everywhere.

Well, the flash sponsorship model works a lot on the ability of thousands of little flash game sites hosting your .swc file to generate traffic for your sponsor. With html, you don't get the single executible file, and only a small percentage of these little sites actually have the know-how of hosting html games. So html games don't get you anywhere near what (if at all any) that flash gets you.

Oh yes, if I ever jump into this, I will surely try out the $99 risk. I would risk the $500 only after my game is almost ready to be released. This kind of no-trial, risk it all system doesn't help small indie hobby devs at all.

Yes, I think Unity would be an overkill too. I only gave it a fair thought because it seems to be very much more powerful in what it can do and the additional benefit of flash export. But that is something I would keep as a last resort, because in there I risk not only money but a heck of a lot of learning time too.

Advertisement

That game was released before the YYC came out, so I know it isn't that. I honestly think that they are not using pixel perfect collisions. But they could more likely than you could, because at least the way I read your post, you were going to have like 30-40 zombies at once needing the pixel perfect collision with all of those bullets(with shape collision on the bullets). But, the game you mention doesn't have that many things for all those bullets to collide with, so it may be that they were able to get away with pixel perfect collision, or it may be that they simple have a couple of rectangles to represent the collision shape of the planes. Of course, I may be simply underestimating the speed of mobile devices and it is very possible that they can indeed handle it. This has peaked my curiosity. I have GMStudio Master, so I think I'll test on my Galaxy Note 2 and see what I get, with pixel perfect collision on everything, just 30 things(against circular shapes) and with just 5 things and see what I get. I may be able to give you some good news.

So for Flash, your thing is that you want to get all over the place, not just on Kongregate....in that case you may not be in good shape with GMStudio, unless you are willing to sacrifice Flash for the ease of use(and mobile portability) of GMStudio.

So yes, if you are willing to risk the $99 risk, I think that will be the way to go. On the other hand, I may be able to confirm good news if you trust me. Assuming you want to go all out with your investment once you figure out it works, I think you will be better off getting the Master Collection. Right now it is running $800, but you get all of the exports. If you want just HTML5 and Android, it will cost you $400 as it is, and you wouldn't get the Ubuntu, iOS, or any other ports except what comes with the Professional Version. Also, you may be able to find on the Game Maker Community some users that were selling coupons for Master Collection for much cheaper than regular price. These guys have won contests from things like TrueValhalla.com and have coupons that get the Master Collection to be free when you go to the Yoyogames Store. Overall, even at the full price, I honestly think it is worth the investment, especially if you have ways to make money off of your games.

And yes, unless you want 3d, Unity is overkill in a sense. Like I said, for 2d, it won't be any faster anyway, and you'll have to invest your time into it. It is much more powerful, but only if you want 3d, or do 2d with some 3d effects. GameMaker has access to what it calls "surfaces" which are basically render textures, which work on pretty all platforms. Also, it has shader access so you can get plenty of special effects into games.



Thank you once again. It would be awesome if you could run a simple test like that. Maybe keep spawning your shapes and bullets every few seconds so you can keep track of how many you get to before framerate starts dropping. If it's too much to ask for, I could try and create something simple like that and you could use your android export to try it out.

Yep, flash is very lucrative. If I could do flash export on GM, I wouldn't think twice about spending $400 as I could recover that with one released game. But considering the future of flash and how much easier it might be to use GM, I am willing to make the sacrifice sad.png

Edit : Here is something I fired up real quick. Could one of you with the android export turn this into an apk to test it out on your device. If possible, attach the apk so I can test it out on my 2 year old mytouch4g. Thanks!

It appears that I may have under estimated the power of the mobile processors, and/or the actual speed of GM created games, even without the YYC.

I changed the room speed from 30 to 60, frankly because it was taking too long to spawn the objects, and because the frame rate wouldn't go down. So it starts at a constant 58-60 FPS, which seems normal. It doesn't begin going down until like 260 instances. These instances are all pixel perfectly checked for collision, as in, not that some are shapes, rather all pixel perfect.

I'm running the same test, making two of the colors to be a box shape, and leaving one to be pixel perfect, which seems like it is closer to what you would be doing with your bullets and zombies, and I get up to like 320 instances before it starts slowing down. I'm not going to do it, but I imagine if it was only shapes colliding with sprites(pixel perfect) it would be much faster, but since there were still some doing pixel perfect checks on both instances, it wasn't as fast.

The last one I'm doing is just to see what it is, with all of the sprites using a simple box test with each other. In this version, I didn't see any slowdown until I got up to around 420 instances.

I'm currently uploading this to my dropbox, as the forum software will only accept files less than 2MB each individual file. It takes up to like 50MB total, but only 2MB per file, and the apks are over 10MB as it is. Don't worry much about that, because that is the size with the engine, etc... included, and so it doesn't increase massively with just a few sprites.

The one I uploaded with just changing the room speed to 60 FPS is here, and the one that is changed so that two sprites are using boxes, and the third still using pixel perfect collisions is here, and the last one is here.

Based on these tests, I'd say for mobile, unless you are going for slower devices, you are probably fine. In fact, for the amount of things you are going to have, you may still be fine on older devices, to an extent. The one thing that worries me is the HTML5 versions. If you intend to export to HTML5, I recommend you do some tests on it to see how much you get, as Javascript running in browsers I believe is likely to be slower than native code on Android, though I could be wrong about that, and of course it depends on the hardware running the Javascript, as well as the browser. I've been told that Chrome is much faster the IE and firefox, but I haven't done any extensive tests myself to confirm this.



To expand on what was said by kburk, Chrome is an awesome browser that runs circles around IE for running games. I have tested it. IE seems to utilize a lot of CPU when running games. However, it may still be fast enough.

@kburkhart84, thanks for much for these. You have been an awesome help.

I am very pleasantly surprised by these results. Even on a 2 year old mytouch4g, I was able to have pixel collision run on 100+ instances before the framerate dropped below 50. This is good news!

One more thing I should have added though is layering, because that is where Starling failed for me. For isometric games, or even side-scroll games with a slight incline (like

), one would need to keep all the enemies on a different depth. I will try and add those when I get home this evening.

If that works too, I think I might just go ahead and put in the $99 for the pro version :)

There is no reason why that wouldn't work. Each instance of an object has it's own depth. There is a default depth that they start at according to what you put in the object dialog, but you can change that at any time, and it directly controls the draw order so that anything that needs to be behind something else can easily be done. As far as backgrounds and tiles, you can also put those in the back or in the front, and you can have more than one active at once if you wanted parallax scrolling for example.

The only thing I'd say be careful about is if you put in the $99 for the Pro version, it may not let you pay only $700 if you want to upgrade to Master Collection. They run specials at times where you can do that, and/or get things cheaper, but there isn't anything going at this exact moment. Also, I recommend you make sure you purchase via the Yoyo website and not the Steam version. The reason is because with the "normal" version, you can have more than one GMSTudio open at once, while the Steam version blocks extra windows open. Also, the Steam version lets you export something special to do with Steam that the regular doesn't, but I don't know if you'd be interested in that. And in any case, if you buy off the Yoyo store, you get Steam keys as well, while if you buy the Steam version, you DON'T get the license to the normal version.

EDIT****

I forgot to mention I'm glad how it works on that older phone. And I wanted to ask, how did the other two versions go? I mean the ones with only partial pixel perfect collisions and the one with none? Last, I should mention that you'll likely get better results than the first two tests because you said the bullets can be simple boxes for collision, so you aren't going to be doing any pixel on pixel tests, rather only box on pixel, or maybe the circle one is faster so you may want to do circle on pixels instead you know?



I am no Starling expert, but from what I understood of it -

The reason this didn't work too well with Starling on mobile is because of the way it draws stuff I guess. It batches sprites based on depth. So the more the layers, the more the batches, and it has to keep loading the batch and then throwing it away when drawing the next layer etc.

Thanks for the heads-up, but I am really not going to go all in with $400 as of yet. $99 will let me use the pro features as well as allow me to test the performance on android as I keep developing. I will buy it off yoyo site though, though the 'publish to steam' sounds interesting.

Yes, I did not even test the others. This pixel perfect collision among 100+ objects is good enough for almost any game I might work on. But once I get home, I will try the others too. I am very interested in a Tower Defense game, and that one would involve 100s of bullets (collision boxes) and 100s of enemies (pixel perfect)

OK, you can invest your money however you prefer...I just wanted to make sure you know what you are getting into when you go that way.

I myself have never heard of Starling, so I can't judge on that, but I DO understand that HTML5 on mobile doesn't work very well.



Thank you. I appreciate all your help in deciding whether I need to go in for GM or not.

This topic is closed to new replies.

Advertisement