Thank you for the thoughts and feedback! Most of the things that you mentioned early in the comment I would consider very advanced. While incredibly useful (things like bit shifting and bit masks), they really only resonate with top level users. Unfortunately these types of videos tend to not do well in broad groups. And that level of optimization really only applies for a large game with very limited resources. As for how I organize my projects, I have this video already planned the pipeline. But like you said, Time is a very hot commodity. lol. The key factor to all the systems that you listed is determined by how modular I want the project to be. If I don’t mind having those systems tightly coupled, I’ll just add them within the player blueprint. Otherwise, if it’s very modular like my online manager, I’ll usually create a subsystem for each of these under the game instance.. This allows me to get a hold of it anywhere I need to from a process that’s already running. I’ve been very focused on creating videos that beginners can follow but maybe I should include a series of advanced techniques? Some things that I’ve learned over the years that I wish I had started doing earlier. I guess one long comment deserves another! I appreciate you taking the time to leave a comment. Thank you for the thoughts as they’ve sparked future ideas for me! Thanks for watching!

@@TicToxicStudios thank you for your reply. Id be excited to see a series, or just one video of things you as an experienced developers wish you knew earlier. I think its probably one of the most valuable kinds of education video for any developer. As for the blueprint vs c++ benchmarking, could you at least do a test itating through an array and performing a check on each element, or maybe sorting arrays? I really wish to know how fast these are, in blueprints, especially the for each and for loop.

That’s fair. I had to pause making the next part because they changed some minor things in version 5.4. Nothing major, but enough to require an update before I can continue. Expect the update and part 2 soon!

@@TicToxicStudios dope im about to get the EOS integration kit and use EOS vs Steam so hyped for the tutorials !

The EOS integration kit is a really good plugin (however kind of expensive) to easily use EOS. Hopefully, when this series is done you can do everything that you can do with this kit for free. I very much like the plugin (I don't own it) as it allows convenience and time saving to not have to set up your project with your own code and the ease of use in the plug in is top notch. I hope that with work and home life, I can stay on a good production schedule and work on this video series soon! But they are high up in the production line.

@@TicToxicStudios yea that would be sick (full disclosure i copped it yesterday lmao) so i'll be diving into that heavily. and hey man no rush, do your thing. do what you like and have grace and patience with your creativity, i love the recent tutorial as well on bp vs cpp. indie devs can use all the tutorials we can get !!

Thank you for the insight! This further confirms my suspicion that the "C++ is faster than BP" is a time remanent from earlier versions of Unreal Engine. I should repeat this test on UE4 to verify. I'll have to watch your video because I am curious about what tests you ran. Thank you for taking the time to leave a comment!

​@@TicToxicStudios It's a bit hard for me to follow from the video, especially that at the end larger offscreen changes to the routine are being mentioned. Can you give a broad overview what happens when the test starts and what actually is being measured? I noticed, that the Total Test Time differs significantly from the combined C++ and BP results (by an order of magnitude), why's that? Is it counting delays, etc.?

I didn't change the actual tests off screen, just how I was calling them and how it was displayed. For each test I'd create and start a timer and generate 1000 random numbers and add them to an array. After they were created, I added the array together and divided by 1000 to create an average. This was just work to do, it doesn't mean anything. I'd then move on to the next tick cycle. After I did it all again over 100 tick cycles, I'd save the time taken and add that value to an array to be averaged over 10 runs (to even out the extremes). Both (C++ and BP) tests were exactly the same. Everything was called in the same order and did the same amount of work per tick for the same number of ticks. The test times varying was always do to the frame limit slowing the test down. ie 30 fps could be no faster than 66.6 secs total time but 60fps would be 33.3 total time. At very high fps (over 400), the background load (windows, screen recording, etc.) begins to interfere with the testing and can randomly cause one of the tests to slow down slightly and disrupt the average. Resulting in sometimes BP running faster than C++ or both tests running slower than expected. Hope this helped!

​@@TicToxicStudios Ok, got it, so [20:31] is the actual workload (with 1000 elements instead of 100) and everythin5g else is there to run multiple tests and track the time. 100 cycles per test. Looking at it more closely, I don't think your benchmarks count what you think they do. Your executable shows me, @30fps: 3.30s for each, C++ and BP, @90fps: 1.1s and @1000fps (running at around 220-230fps in my case) of 0.44 seconds. It shouldn't be the case that the same workload takes (exactly proportionally) longer based on frame rate. I don't think you are correctly benchmarking the execution time of the workload but rather are returning the entire frame time. Example: 3.3s = 3300ms / 100 cycles = 33ms, which is exactly how long a frame takes at 30fps. 1.1s = 1100ms / 100 cycles = 11ms, which is exactly how long a frame takes at 90fps. 0.44s = 440ms / 100 cycles = 4.4ms, which is how long a frame takes at around 220-230fps. Your displayed times are returning that entire frame time, that's why the difference between the two test groups is negligible. I've never used Timers to benchmark code execution and am not certain their ElapsedTime can give you correct results in such a use case. Personally, I'm using PlatformTime::Seconds() and the Now node to benchmark execution times. I've recreated that workload in a simplified project (no GUI, no tick, no timers, no frame rate concerns). For me, that code executes in Blueprint in 5-6ms in PIE and 1.4-1.5ms in Standalone. In C++ it takes 0.16-0.19ms.

@Cpt.Tripps I appreciate the time you have devoted into checking my work, as well as indulging me in civil debate (which is hard to find on the internet!). While, theoretically, @30 fps a frame should take 33.333(repeating)ms to complete, the test times are typically faster (for me around 33.0001ms) due to the engine finishing the workload and waiting to preserve framerate. It isn't until you are attempting more work and the program now waits for the work to complete that the times will increase. This shows that my timers are counting the work time and not entire frames. For example, in my same program, I changed the workloads from 1000 to 5000. @30 fps both still ran at 3.30001s. But @60 fps BOTH times increased to around 2.45s. Only a frame rate of 40.39fps since it was unable to do the work within the allotted time. The point I hoped to make was that it doesn't matter what work was done so long as the timers recorded the same amount of work in each test. And that I am not trying to make anyone change their mind about what to use, just pointing out my discoveries the best I can. Once again, thank you for the mental resources you have put towards thinking about this and I hope others can benefit from reading deeper here as well. (My mind is stuck on " not certain their ElapsedTime can give you correct results" I believe that I will have to test this now, thanks for that! lol)

I'm so glad that you found this informative! Like I said in the video, when ever you dive into C++, focus on good coding practices! Do you think I should make a "C++ for beginners" video? Maybe something geared towards those with some practice in UE5. Thanks for watching!

Thanks for the feedback!

Thank you for the feedback! I agree with you. The audio in this video was overall a mess. I have learned from this mistake and now use better background music that is MUCH softer and less distracting. Lessons learned. Thank you for spending the time to leave the feedback and help me improve future videos and better the tutorials for future videos. Thanks for watching!

Thank you for the feedback! Oddly, the more "real" you make your ai in a game ... the more gamers complain. I think there is a certain level of "NPCness" that the player expects from AI. It's almost EXPECTED that an enemy stop looking for the player fairly quickly, otherwise the player feels like the AI is too hard. Sometimes, a well made NPC is bent to the will of the paying players. For limited conditions, you could shorten the vision distance or make it more narrow. You could also add a SECOND sight component that is wider and have that be the "I think I saw something" vision and you the one in the example as a "A-HA! I found you!" vison. Thank you for watching and the great question!

@@TicToxicStudios Kind of funny that they don't want smart NPCs, but they do want multiplayer. Though I guess it is fairly easy to make an NPC that's smarter than the average gamer. Heh. So basically create a "visibility" variable that gets set based on ambient conditions. And that variable then modifies the sight range. My main goal isn't to make hugely popular games. I want to make games that I want to play. And I figure others share my tastes as well. Sure, it'd be great to make the Next Big Thing. But I'd be content if I could just pay a few bills with it.

Thank you for the feedback! This was exactly my goal! I wanted more than showing how to do it. But, teach what happens enough for viewers to feel comfortable making changes and adjust to their needs. I’m glad this video helped and thanks for watching!

Thank you for the kind words! I hope that I can continue to make each video better than the last! Thanks for watching!

This system can track any (and as many) stats as you desire, as long as the value is an integer. With some clever coding, you could easily create a realtime heat map of players position or death locations. If I were to look at building this system, I would consider creating a numbered grid over the whole map and sent what grid number they were in when they died to the stats server and increase the count by one. You could then pull from the server all of the grid numbers and color each of the grid tiles based on what grid has the most deaths. This is a very clever use of the stats system! Thanks for watching!

I’m sorry to hear that it isn’t working for you. There are many possible problems from possibly your key bindings are different to you are accidentally pushing comma instead to maybe you missed a step or have the nav mesh modifier instead of a nav mesh, or anything in between. Thank you for watching and I hope you solve the issue!

Thank you for taking the time to leave feedback! I agree, the music doesn’t always match the flow and comes in a little loud. Thankfully, I am learning and trying to get better each video. Thanks for watching!

Thank you for the feedback! It’s a hard line to walk when trying to balance between “to the point” and “too fast to follow”. I’m glad the pace works for you and thank you for telling me! I hope I can keep the flow in my future videos and keep making them useful. All I can hope for is to make each video a little better than the last. Thanks for watching!

@@TicToxicStudios looking forward to more content - already subscribed and will pass along the good word!

Glad you found this helpful! I could go forever just talking about AI and behavior trees, the more time you spend adding branches, the more lifelike they become. I will CERTAINLY go back and add more to these AI systems including sound detection along with a system to randomize tasks to make them more lifelike. Thanks for watching!

Thanks for the feedback. EQS is definitely on my to do list. I just need to work through it so that I can explain it the best way possible!

Sorry to hear that! I’ve gone a bit slower in the videos since then.

Thank you for the kind words!

Great idea! I will have to research this first. But I think this would be a great video!

This is a wonderful idea, I just worry that this would require a much longer video to explain well. Do you think a 30-40 minute tutorial would fly? Or, is that too long to sit through?

@@TicToxicStudios I'd watch it👍

@@TicToxicStudios I don't agree with the idea that Unreal Engine tutorials can be too long, especially if they're well done and helpful. (Yours are!) People can always pause, and come back later if they need a break. Alternatively, you could do a tutorial series consisting of several shorter "chapters" to cover the topic well. Word will get out. They will come.

Thank you for the feedback! So Glad it could help you!

Thanks for liking!

Thanks for the feedback! I agree! I had someone asking about it and rather than just explain it to one person, I thought it would be better to document it for the world. Hope it helps someone!

Song goes hard

Thanks for the feedback! In this tutorial, the fracture pattern would always be the same. To have different ones you could create multiple geometry, collections and create multiple broken cube blueprints, and spawn a random one. Getting the small pieces to move with the field is a little tricky. You would have to prevent them from falling asleep. I played around with actually waking them up, but it didn’t always work. Personally, using a third fracture pattern in the style that I did in this tutorial will basically let you shoot them until dust where the pieces are so small it doesn’t matter.

Project settings> unpredictable behavior>dumba** friends >apply behavior tree

@TheAngryRubberDuck Thank you for the feedback! I'm glad my FANTASITCAL artwork was useful. I didn't want to go TOO much in detail, but still give the basic idea so that viewers would be comfortable playing around on their own and know what is happening. Glad you enjoyed this!

Yes, you can! However I found it unreliable (the posable mesh would sometimes twitch) and memory expressive (I didn’t research why). This felt like an easy to understand way that fairly performant. Plus could be optimized better using interfaces to avoid some of the casts. Thank you for the thoughts and feedback!

Thank you for the feedback! I’m glad you found the video helpful. Personally, I appreciate fast and to the point videos also. It’s easy to back up and pause, but it doesn’t hurt to slow down at the important parts as I have started doing in future videos. I have also started displaying completed code longer at the end of each section so pausing and double checking code is easier! If you have any ideas to make my videos better I’d love to hear it!

Long way to go, but I’ll keep you posted!

Thank you for the kind words and feedback! If I ever start slipping, call me out! Plus, ideas about how to improve future videos helps everyone!

If that wall has visibility collision enabled (on by default in the blockall preset) the NPC shouldn’t be able to see through it.

Thank you for the feedback! I agree and have slowed down my videos since then. Still trying not to waste time, but giving it the proper time to learn!

Thanks for the feedback. I noticed the echo after I posted. EXCUSE: the furnace kept turning on and making noise in the mic. Fixing that in post caused other oddities in the audio. Long story shortened: I messed up and I didn't have time to fix it properly. I'll take better care to keep clean narration in future videos. Hope it wasn't TOO distracting!

@@TicToxicStudios Nah chief, all good! Keep these videos coming!

I will absolutely cover this! Unfortunately, it might be just a bit before I can. Stupid real world work always getting in the way! I hope to get to that video in the next few weeks.

Thanks! Hope it helps or inspires great work!

Thanks! I like this idea, it's surprising easy using states in the behavior trees. I will make sure I cover this when I get to states as I WAS just going to do alive, attacking and dead.

Thank you everything.. the information you share with us is greatly valued.@@TicToxicStudios

Thank you for the feedback! So glad that this helped you! Isn’t it amazing how something so small as one setting being wrong could be so maddening?

@@TicToxicStudios Especially since so many examples tend to drop the character directly on the screen. I'm new to Unreal and was unaware of this setting. Looking forward to Your series.

@dajajax I’ve been debating a video idea, Would a very beginner video series about using Unreal Engine be useful to a beginner like you? Like a “Your first day using UE5” where I go into detail about the settings and properties of different parts?

​@@TicToxicStudios In my case, I'm more used to search for videos of mechanics, like this one. What could be a really nice adition, and something not very often seen, is if you could do, after showing a couple a mechanics, a video of how to integrate them in the same proyect. Allow me to divert a bit, lets say for example, you have made a video after this one, of Actors attacking the player after reaching it. That chould totally articulate with this video (and part 2), and for my experience what newcommers are lacking the most, is the understanding of how to implement several blueprints at the same time. I hope my response ends being of use to you.

Excellent feedback! Thank you! You have sparked an idea that I think I can apply to a large group!

I like the ideas here. Most of these are heavy in animation to have smooth transitions. I'm not very strong in the animation department, but I am working on it. Once I get better, this will for sure be a video. I'm thinking that I will follow the "soldier" style NPC and have them patrol, search area, attack/shoot, cover, and reload. Flanking would be a group or squad mechanic, which will require more explanation so would probably need its our video. Thanks for the great feedback!

This is an absolutely brilliant question! In unreal engine 4, this was a built in function. But it was SUPER performance heavy and wasn’t used. They got rid of it in UE5. Honestly, I hadn’t tried to find a solution. I will give it some time to try to find a solution and post a quick video about it when I come up with an answer! Thanks for the question and feedback!

@user-ny2by6xo1w This question haunted me all day! I ended up making a video that I will edit in a day or two about this. Long story short, I couldn't find a way to use nav meshes easily. BUT I did find a lightweight way to do it. I'll post the solution soon!

Thank you for the kind words! I’m hoping this series helps new developers (and veteran ones) more understand how and why everything works over just being shown one more example.

very true I look forward to more of your content you make solid work@@TicToxicStudios

This is a great suggestion! This would pair well with crouching the player and controlling noise levels. Also external stimuli that make noise, like throwing a pebble to distract them or knocking a bottle over. Thank you for the feedback!

@@TicToxicStudios that all would be awesome. I was thinking the whole time how I could do those. I think I could figure out smell for animals (putting together a action survival shooter like farcry) but sound seems a little more advanced.

@Jesters-Jinx Smell could become very challenging since there isn't an AI perception already built in. I imagine a system like Red Dead Redemption 2 would be fun to build, but would be hard. I don't know if I would cover that in the future due to the potential complexity. (It could be easy... never tried it) I'd have to play around when I start on the ai sound part. Another great idea!

Such a great idea! I could show a few ways to spawn ai and maybe have them search for the player or pick up a spawned item and disappear!

Awesome! Glad it helped!

@ricardo22448 Exactly! Having control is usually important, but you could just let it fall if you wanted. I was considering flinging them into space... that would have be fun. Hindsight is 20/20.

Thank you! Is there anything you want to learn about in a future video?

A follow up to AI would be great. There are many ways to implement AI behavior, but it's hard finding the right way. Taking behavior tree tasks as example, there are some blueprint nodes you cannot use inside a task. This leaves the question, which logic you put into tasks and decorators and which logic into for example the character blueprint. For example, where do you start animations or react to outside events. If you have a patrolling NPC and want it stop patrolling if it got hit by the player, how would you do it? Also debugging behavior trees is a pain, the logic in the character blueprint has a debugger with breakpoints while the behavior tree just shows the current state, is there a good way to debug behavior trees ?

@@Lilium___ Great suggestions! I plan to cover these in part 2! Specifically seeing the player to attack them if they’re an enemy, being hit by a player and running away, and returning to the patrol when the player is out of range. Adding the debugging and testing into this video is a terrific idea.

Your welcome! Hope future videos help also!