Posts by Outcast

Quote from willy

What is the resolution in your game?
I noticed the same when I upgraded my monitor. I then upscaled my game to 1280x720 and that solved it

It is 640x360. It sound strange that increasing the resolution would solve it? How would that make sense as the hz and frame rate is the same ? Perhaps it is just harder to notice when everything is smaller ?

I always develop my game in 60fps for PC (Steam) and I have had a monitor running in 60hz for the longest time. However I recently upgraded to a 120hz monitor and when I continued development of my game I was very surprised when I booted the project up because it looks extremely unsmooth, choppy and "stuttery" when I run it on my 120hz monitor, especially when scrolling. It really feels like playing a game in very low uneven fps and is extremely unpleasant to look at.

How can I make my game look good and smooth on 120hz monitors? From what I understand 120hz monitors is becoming more and more common, and even 144 or 240hz.

If I change the fps to 120 this makes the game smooth again (though I need to adjust some values that are dependent on the fps), but the problem then is that it becomes much more performance heavy to run in 120fps vs 60fps so then it could become slow because it cant reach a stable 120fps all the time (and if going higher like 144 or 240 it would be even worse).

Is there a way to still run the game in 60fps on a 120hz monitor or higher but still have the game look smooth?

Quote from semar

Sorry, I misunderstood your question.. when you press F8 it builds and runs, so there should be no debugger running in background - at least, I guess so.
The debugger I meant is the one that comes when you press F7 ( run frame).

However, when you press F8, you have to wait for the build, for the .exe file deployment, and then it will be run. This costs time.
When you instead double click on the .exe file in a folder, there's no delay for a build and for a deployment, because the file .exe is already there.

My two cents. Perhaps other users more expert than me could share some light on this topic.

Just to be clear though, the wait time is not before the application starts but before my game frame starts so my "loading screen" still show (i have 2 frames before the game frame)

Quote from semar

Could be that the debugger slows the app down when ran from within Fusion, whereas the compiled .exe is obviously not affected from it.

Hm I never use or show the debugger, is it running anyway in the background and if so is it possible to disable?

I have a project that have some really huge animation frames. These don't pose a problem in the actual build exe and the game starts quite quickly within a few seconds (I count 3 seconds to boot it up in the actual exe).

However in the developer runtime when I develop the project and run it with F8 it takes by my count 12 full seconds for the game to load. Of course this is a huge slog on the development. The only "solution" i have is to just to swap this object with the huge animations while I work on the project and then put it back in when I need to actually need to have it or build the project. But that is also a big hassle. Is there any better way to do it? is it possible to somehow ignore some animations with some toggle or something?

Or make it load as fast as in the exe? (Not sure why there is such a big difference?)

I have an issue that seems like a clear bug to me but no idea how to solve it. In my game I have physics objects that are supposed to be paired with collision objects to make them land in different spots. As I tested I noticed that it worked well for about 10 seconds in my game and then things started to bug out with the collisions and the physics objects fell through their paired collision objects. It was not consistent though and if pausing for a bit and then creating more physics objects the collision could start to work again.

I have isolated the issue down to the issue being caused by creating any other object in a loop. Please login to see this attachment.

In my isolated example you create the physics objects by holding down the left mouse button. After the objects are created there is also a collision object created that pairs with the newly created physics objects fixed value and then an event to check for collision between them and trigger if they collide and their ID match.

This works perfectly and looks like this:

Please login to see this attachment.

Please login to see this attachment.

However.

If I then add this loop where a loop is run 200 times each frame and creates another object (that is then destroyed) everything bugs out after a few seconds (usually between 5-10 secs)and the physics objects starts to fall through their collisions. If I keep holding down the mouse they continue to fall through and sometimes they work properly again for a few seconds before falling through again.

Please login to see this attachment.

Please login to see this attachment.

Also I noticed something strange with the fixed value. The number in the top right is set to the physics objects fixed value and I noticed it seems to go negative for some reason for the "bugged" objects that fall through. Replacing the method of the fixed value with just an alterable ID value that is increased for each object produces the same results though.

I have attached the mfa example. The first frame is the bug and the second frame is with the group that runs the loop to create the object disabled and thus working.

Do anyone know why this is happening or how to fix it?

Please login to see this attachment.

Got answer from Fadzior 15+(min(1,number)*5)

I am trying to enter a formula into an event that I think should be quite straight forward but I am blanking.

I have a number (15) and in an event I want to retrieve andother number that can be anything between 0 and 100.

In this event I want to set the number 15 to 15 + this other number *5, but I only want the other number to be counted as 0 or 1 (0 if 0 and 1 if 1 or above).

So the event result in this case would be either

15 + (0*5) = 15

or

15 + (1*5) = 20

So basically I want to retrieve the other number and have it set to 0 or 1. So in all cases it is above 1 it would still be 1. And I want to do all this in one formula, not in another event prior.

Want I am asking is really much simpler than I think I am trying to explain..

One thing I do very frequently is to have a set of frames in an object and then another event to set the objects frame to one of random of those frames. This can be for all sorts of things like for example you have a bush object and you have 5 different frames in that object to make the object appear as 5 different types of bushes at random.

However I always have to change the formula as I add or remove frames so if I have 5 frames I have set to random(5) etc. But if I add another one I have to manually change it to 6 (or have an alterable value to change).

My question (or suggestion if it is not possible) is if it would be possible to instead just set the objects frames to random for however many frames there actually are? This would cut out a lot of busywork.

Quote from fredetmumu

i think it is what you wantPlease login to see this attachment.

Not sure if I missunderstood the example but the floor is static in this?

Quote from marbenx

Would something like this work?

This feels closer but there is still some odd looking behavior from the bouncing at times so I wonder if the method Please login to see this link. mentioned would be preferable

Quote from Volnaiskra

I've never used physics objects before, but I played around a little with your example and it seemed like the bounce only works with the green collision condition? Even turning it into an "is overlapping" condition stops the bounce from working.

If that's true, then perhaps you need to give each casing its own little floor object. When you create a casing, you also create a little floor object at the correct height. And you 'marry' the casing and the floor by storing the fixed value of one into an AltVal of the other. Then you check that these match when you perform the collision test, so that casings will only bounce on their own floors. And when you eventually paste the casing into the background you also destroy its matching floor object. For sure this will be more expensive, but might be worth a try.

Ah yes this might actually work and would be worth trying out

Quote from willy

I've seen a top view explosion example with debris many years ago. I couldn't find it again, but I think the debris had the behavior you are looking for.

I made this example just to show a possible direction. It could be better.

Thanks for the example, but I am looking for something like this but with the physics object. It seems really hard or impossible to simulate a bounce effect that looks as good as or at least nearly as good as with the physics object. I wish it was possible to trigger a "collision" bounce like this though with the "floor" with the physics object.

I am making a sort of isometric view shooter and I am using physics for making the shell casings when shooting fly out and land on the ground. As I have it now I just set an Y value at random and when the casing has reaching this Y position I destroy it and paste it into the background. This works fine but there is no bounce of the casing or anything which I would optimally want and that looks really nice.

I wonder if there is a way to do this where I can somehow make the casing bounce as it reaches its Y value as it actually hit the floor so I can let it bounce for a while and then paste it in when it has stopped?

Since the game is in an isometric view from above the "floor" where it should hit it and bounce is different for each casing.

Do anyone know how to do this?

I have attached an example to better show what I mean. By pressing the right mouse button you see what I have currently and by pressing the left button the casing lands and bounces on the static floor (the effect I want but with a "dynamic" position of the "floor"

Quote from punchbird

Well, I have no idea!

I was procedurally generating large tile-based maps and ran into the same problem, and this fixed worked for me since I don't use any of the featuers requiring directX.

Great question though. Yves

Ah I see. Yes it is a huge bummer and pain for my new project so it would be amazing if it is somehow possible to "override" this limit on pasted background objects. Using actives to then destroy seem to much slower than just keeping pasting background objects, but the limit really mess up what I am trying to achieve. The limit seems very arbitrary in general, even for active objects (?) But much more so for background objects and especially since you can't even track their numbers.

Quote from punchbird

Can you do an experiment?

Switch from directX to standard

HUH, this actually seem to remove the cap? :O How so? The downside is that the game also looks really bad under the standard mode and many of the effects like the semi transparent layers etc are gone so I can't really use the standard mode.

But why is there a cap on the directx and not on the standard?

Quote from Volnaiskra

marbenx means that while the casings are off-screen, replace 10x individual casing graphics with a single pre-baked graphic that contains 10 casings in it. The pre-baked graphics won't be positioned as precisely (or as randomly) as the originals, but you'll still have 10 casings visible, while reducing your total objects by 9.

The problem though is that it will still be another object pasted into the background so it will still increase toward the object cap, just slower.

I am unsure how pasting into backdrop works but is it not supposed to just "merge" the image into the background so there is no actual individual object? It feels like it can cause a lot of issues if those pasted objects are actually counted as real objects and toward the object count since they dont show up when setting a string to the total objects in the frame? So then you don't actually know how many "objects" you actually have or how close to the cap you are. As in my example the recorded objects in the frame is never exeeding 102 objects but I still reach the object cap and cant create new objects.

Also curious where the hard cap of 30.000 objects come from? Is it just arbitrary or based on something? It would be wonderful if it was somehow possible to increase by editing something.

I notice that even with 30-60k pasted background objects my game does not slow down anything (again weird also as in my latest example I uploaded that I can actually double the object cap for them by using more layers it seems vs 1 layer), but just a fraction of those objects as active objects slow down the game a lot so being able to actually not have a cap or a much higher cap on pasted background objects would be wonderful. Again I don't understand how those are even counted toward the total objects since Fusion dont report them as objects when having an event to check how many objects are in the frame.

Quote from marbenx

Pasted background objects count towards the limits as well. There's a couple things you can do though to go around this

- Use the 'surface' object to blit the background

OR you could take an approach such as;

- Keep background objects such as shell casings/decals as active objects, until they're outside of the player's view

- Once they're outside of the player's view, delete a bunch of them (say 90% of them)

- For the remaining 10%, change the animation so that instead of one shell casing, there's 10 or so of them, in a larger area

- Then paste it into the background

The second version won't be 'accurate' but if you're dealing with thousands of them at a time, the player probably won't notice

Display More

What does it mean to blit the background with the surface object?

The whole point is to have things getting really messy over time so removing stuff really defeats that point =/

It is weird though because I tried pasting into different layers and then the cap seems higher? I increased the cap to the max of 30k objects in this example and in the first one I use the same method as the first time by just always running the loop to create the objects and then paste them and destroy their active and it runs the loop 30k times until no more objects are created or pasted. But if I instead use several layers I can get it running all the way to 60k? But there seems to be a cap there instead. It is very confusing.

You can run the test by either pressing 1 or 2 at the start of the frame. You need to restart the expample to run the other test

Ah thank you very much Volnaiska I will try out this and see what I can achieve