That's your own engine? That's really cool!Hey, I've finally released a demo of my indie game, Prototype N, that I've been working on for years. It's a short demo (one level) because I rebooted the game's development a few weeks ago and built it upon a more completed gameplay engine I've recently programmed.
Some screenshots:
landVec = theBoard.GetComponent<Transform>().position;
landVec.y = 0;
var staleDist = Mathf.Round(Vector3.Distance(jumpVec, landVec));
A brand new engine built in using Clickteam Fusion 2.5+ to be a little more precise!
Weird Vector3.Distance question ahead: So we tried implementing a basic combo system in Unity last week and it works pretty well - it has a timer where you need to do a trick within a few seconds or you'll drop the combo. You also drop the combo if you crash. But as a bit of a more complex idea, we programed in a super basic "stale moves" calculator to encourage players to move around the map more. We basically just compare the jump position to the land position with Vector3.Distance:
And if that number is less than 1, then you haven't moved far enough so stale moves kicks in. This usually works well - at full speed, you can skate around and jump to get just above 0 and continue the combo. We're also setting y=0 for both jumpVec and landVec so that you can't just jump really high in the same spot and avoid the stale moves penalty.Code:landVec = theBoard.GetComponent<Transform>().position; landVec.y = 0; var staleDist = Mathf.Round(Vector3.Distance(jumpVec, landVec));
The weird thing happens every once in awhile where you can jump in one area but get a staleDist of some absurd number - we're talking live well over 20 or 30, which should just be reserved for longer jumps. I can't seem to trigger it when jumping on the ground, only off jumps. But I thought zeroing out the y value would lock that problem off? So I'm at a bit of a loss since I can't always replicate this. Am I understanding Vector3.Distance correctly?
Console log showing the value for a series of jumps I did off the same jump - almost all of them 0 except one that's over 30...
Bonus clip of it actually working, which is i dunno...75% of the time? Maybe more on just ground to ground jumps.
Weird Vector3.Distance question ahead: So we tried implementing a basic combo system in Unity last week and it works pretty well - it has a timer where you need to do a trick within a few seconds or you'll drop the combo. You also drop the combo if you crash. But as a bit of a more complex idea, we programed in a super basic "stale moves" calculator to encourage players to move around the map more. We basically just compare the jump position to the land position with Vector3.Distance:
And if that number is less than 1, then you haven't moved far enough so stale moves kicks in. This usually works well - at full speed, you can skate around and jump to get just above 0 and continue the combo. We're also setting y=0 for both jumpVec and landVec so that you can't just jump really high in the same spot and avoid the stale moves penalty.Code:landVec = theBoard.GetComponent<Transform>().position; landVec.y = 0; var staleDist = Mathf.Round(Vector3.Distance(jumpVec, landVec));
The weird thing happens every once in awhile where you can jump in one area but get a staleDist of some absurd number - we're talking live well over 20 or 30, which should just be reserved for longer jumps. I can't seem to trigger it when jumping on the ground, only off jumps. But I thought zeroing out the y value would lock that problem off? So I'm at a bit of a loss since I can't always replicate this. Am I understanding Vector3.Distance correctly?
Console log showing the value for a series of jumps I did off the same jump - almost all of them 0 except one that's over 30...
Bonus clip of it actually working, which is i dunno...75% of the time? Maybe more on just ground to ground jumps.
Just gave this a play. It is coming along great so far! Once you create more levels I would make the secret areas a little harder to locate. I loved the part near the end where you continued to jump higher on the springboard. It would be neat to see a level with a number of those with obstacles and enemies in between.Hey, I've finally released a demo of my indie game, Prototype N, that I've been working on for years. It's a short demo (one level) because I rebooted the game's development a few weeks ago and built it upon a more completed gameplay engine I've recently programmed.
Some screenshots:
docs.unity3d.com
Another good one dude, keep it up! I'm not too familiar with Unity's blend spaces (and actually Unity's 3D anim system in general). Does it have support for event-track based animation sync? If so, that would be a good next tutorial if you don't already have one planned!
Yea, you don't really ever want to change that timestep unless you're very experienced working with physics systems. Pretty much every tutorial, example etc. under the sun for any physics simulation says something along the lines of "Changing the timestep of your physics simulation can result in unintented behaviors", which seems to be what you're encountering! 😂
As for your low FPS issues, lets split this into two parts:
1) Your physics logic appears to be frame-rate dependent, which is not what you expected
2) You get low FPS when playing on less powerful devices.
Both of these are going to be quite tricky to figure out. I can suggest some things you may want to look for and some fairly brute-force debugging methods. So:
Physics Sim inconsistency:
I think you're on the right track, making sure that your forces are all being applied in the Fixedupdate and not in the regular update. Check forces being applied as a response to input too! Also make sure that the logic you carry out in the fixedupdate is as minimal as possible. You really don't want to be coding in here unless its absolutely necessary. Don't do any input processing, calculations etc. as that should all be done in the regular update tick. Basically you just want to be applying forces and not really much else.
As for debugging, I would suggest doing something incredibly destructive to try and pinpoint your problem (I do this all the time, both personally and professionally so don't worry about it seeming a bit sketchy). Try commenting out everything in your fixed update except for a very basic physics instruction (e.g Apply a force based on the Fwd vector). Now run the game on the low-power device. If the frame rate is low, are the physics still behaving as you might expect? If so, start uncommenting things bit by bit and try again, see where the issue occurs!
However, if the issue does occur, then it means that the problem is elsewhere (likely due to things stemming from non fixedupdate methods). Just do the same there, comment everything out and re-enable it bit by bit. See what is screwing you over.
Low FPS on less power devices
Annoyingly, this could be due to many things. The first thing I always tell people to check (because Unity is NOTORIOUS for this) is GC Allocations (Garbage Collection). After that, check for OnGUI related calls. If you have any of those, get rid of them because they're awful and end up making a ridiculous amount of uneccessary draw-calls. Use a plug-in UI solution instead. Finally, comment out all your debug logs. These can absolutely tank the framerate!
After that, we're going to have to run this through a profiler. Have a read over this link to get familiar with how the profiler works, it can be a little confusing but essentially it will tell you where all the time is being spent during processing of a tick.
Unity - Manual: Profiler overview
Here's an example of what mine looks like:
There's a lot to take in here so here are some tips:
- Under "CPU Usage", you can click the coloured icons to turn on/off their display int he grpah (e.g in my case I dont care about Global Illumiation, UI etc. So I' m not looking at them)
- Whilst "CPU Usage" is selected, Select the "Heirarchy" view mode just above the details panel. You can then click on a colum to sort based on that.
- Sort by Time ms to find out which methods are taking the longest time to process
- Sort by GC Alloc to find out where all your garbace collection allocations are being made.
- This one can be a bit cryptic. It tells you the function in which they are being made but not the exact line. You'll have to comment out things line by line and then look and see if the number of allocations goes down, if you're unfamiliar with how to spot them. Once you've found them, either give me a shout or google "X METHOD without allocations c#" and you should hopefully find an answer
Hope that helps, let me know how you get on!
if (isPush == true)
{
isPush = false;
if (Math.Abs(this.GetComponent<Rigidbody>().velocity.magnitude) <= 100)
{
if (!goofy)
{
this.GetComponent<Rigidbody>().AddForce(this.transform.right * 5000);
}
else
{
this.GetComponent<Rigidbody>().AddForce(this.transform.right * -5000);
}
}
}if (rotateRight == true)
{
rotateRight = false;
if (theBrain.GetComponent<WheelsTouch>().wheelsTouchCount >= 4)
{
this.GetComponent<Rigidbody>().constraints = RigidbodyConstraints.FreezeRotationX;
}
this.GetComponent<Rigidbody>().AddRelativeTorque(Vector3.up * 20f, ForceMode.Acceleration);
}
if (rotateLeft == true)
{
rotateLeft = false;
if (theBrain.GetComponent<WheelsTouch>().wheelsTouchCount >= 4)
{
this.GetComponent<Rigidbody>().constraints = RigidbodyConstraints.FreezeRotationX;
}
this.GetComponent<Rigidbody>().AddRelativeTorque(Vector3.up * -20f, ForceMode.Acceleration);
}
Yeah, the forces are all applied in the fixedUpdate - Here's a quick shot with the profiler outlined by hierarchy and time. I'll keep taking swings at it during the day if I can't figure out what's eating this playerloop.
