Complete Noob / First Project / Seeking Advice

[TJB1973]

Hello there!

I stepped into my first project, it is for portfolio and I was hoping to get guidance regarding what I should require from the Client to build the space inside of Unreal Engine and some details on deliverables.

I am thinking as a start from the Client:

• Standard floor plan
• Color/material swatches
• Furnishing list

Deliverables:

• Still Shots
• Motion Shots
• Various Exterior/Interior Lighting Contexts
• Ambient Audio Tracks

For now that is what I came up with and would appreciate any working knowledge shared with me!

Thank you

TJ

[MarcusRivers]

i wasted..,but with no regrets....my first 2 months in unreal engine, because i started a large project before i knew the engine. ok it worked and i familiarized myself with ue, but having used the last month systematically going through all the main areas of ue, all the nodes, programming principles, blueprints etc, really getting in deep and dirty, i now know that i made a shameful mess of my first 2 months and could do it all again in a week, 100 times better.

i religiously watch youtube tutorials every day....mathew wadstein is leagues ahead of anyone else imo for - to the point useful and comprehensive info and i have made huge progress because of his resources.

and at the end of the day, ue is a programming tool, so i dont believe its possible to get genuine results unless you are prepared to get in deep with programming concepts even if you only use blueprints.

[TJB1973]

Marcus,

Thank you kindly for taking the time to share your experience, insightful and very much appreciated.

I just subscribed to Matthew Wadsteins channel, set to get all notifications and excited to jump into the videos, thank you for that resource.

I am going through the fundamentals outlined in the UE4 Online Academy now, each session and through all the modules.

If I could, I have 6 months to dedicate to learning UE4 and the surrounding applications for development in the Arch Viz which, I think, can dip into Gaming Environments too. If you had 6 months, starting right now, and was seeking to enter into the Arch Viz world, how would you approach it?

Thank you Marcus, looking forward to your answer.

TJ

[MarcusRivers]

well, ive just finished my 3rd month in unreal, and i am hopefully going to finish my first ever game this week and its far beyond what i thought it would be. ...lol. it wasnt planned, as im learning ue for archviz as well, but i figured if i could program a game the i can cope with the basic programming needed for archviz user interactions. i was just actually messing around with physics as im going through all the major areas of ue i think would be good to know for ue projects.

i may not spend much time with things like character animation, bones, mixamo things, but its still useful to know how to control characters and cameras etc and have them interact with a player and items in the scene.

my thoughts on ue.

if i did not believe in the final output quality of ue, i would stop immediately. ue's inferface, crashes, bugs are all horrendous to be honest. its a hell of a steep learning curve, and the program feels permanenttly in beta, and has crazy thought proceses behind the way it wants you to work.

i guess im spoiled coming from a background in solidworks and cinema4d, both i love for their pleasant interfaces and methodolgy. ue just makes me repeatedly swear at the screen several times every day, but i persevere because im betting that in a year or two when graphics cards come out that have enough power to use real time raytracing properly, that i will be expert level, and offline rendering will essentially be dead.

i really hope ue has matured enough and the insane bugs fixed by then and epic spend some time fixing the ui rather than relentless feature development, which i am beta testing for them apparently.

starting right now, i would probably not do much different, except not spend two months building a whole house...a couple of rooms would have sufficed. lol. but i did learn the workflow. for sure i spent hundreds of hours testing global illumination settings and finding workable setting, rendering flaws and methods to combat engine compromises. hundreds of tests with file format settings to import scenes. c4d datasmith is **** for me so i use fbx now, but that has a whole series of agonies attached to it.

learn the main areas you will need for archviz

materials, material instancing, parameters, texturing and uv manipulation inside and outside the engine.
lights, lighting, global illumination, and related settings and workarounds for problems.
widgets, and 2d user interface element and control of them.
the internal object heirarchy, controllers, pawns, actors and how to control them.
physics, physics materials, simulations and collisions, and UCX collision meshes.
blueprints obviously.
the sequencer for animating smooth camera motions and making stills/videos
emitters, cloth simulations are cool to know too.

i would spend 75% of my time learning blueprints, familiarize myself with as many nodes as possible, and programming concepts in general, especially classes, inheritance and interfaces, and some common design patterns, the other areas are relatively easy, you will save so much time in the long run if you know how to program efficiently and reliably interaction and communication between items and things in your scene, because its incredibly easy to make a huge unworkable mess.

obviously i can only say what i think and know after 3 months of ue, there are whole areas of the program i have not touched yet, but i think the chaos destruction, niagiara, skeletal animation, hair, are not fundamental to archviz.

i guess it depends on your final goal, if youre solely looking for still visuals, i guess the programming is not so important. im focusing on interactive visual demos, movies and mechanical simulations, so my needs may be different.

good luck and stick with it.

[Sam Deiter]

Hello there!

I stepped into my first project, it is for portfolio and I was hoping to get guidance regarding what I should require from the Client to build the space inside of Unreal Engine and some details on deliverables.

I am thinking as a start from the Client:

• Standard floor plan
• Color/material swatches
• Furnishing list

Deliverables:

• Still Shots
• Motion Shots
• Various Exterior/Interior Lighting Contexts
• Ambient Audio Tracks

For now that is what I came up with and would appreciate any working knowledge shared with me!

Thank you

TJ

If you have not done so already check out the online learning portal.

https://www.unrealengine.com/en-US/o...arning-courses

Like MarcusRivers pointed out this could really help you get a better understanding of how this all works.

[UnrealEnterprise]

i religiously watch youtube tutorials every day

That's great when it works... Materials / FX / Landscapes / Particle videos can be super helpful. But I wouldn't recommend learning Blueprints or Multiplayer or AI this way EVER! Overall video tuts aren't for everyone anyway, as they can be an incredibly slow and fragmented learning process, especially when videos are out of sync with the engine (which has changed meantime). Whereas dissecting working projects is far more powerful learning (if you can work that way). Its just more practical and immediate and ultimately more useful, as it helps speed up work on your own game projects etc...

Overall, better insights come from experimenting and trying things out, as it just gives you more confidence. Videos that cover Blueprints or visual coding especially, tend to teach a lot of bad habits or wrong techniques, because they only have so much time to 'fit the message in', and can't over-complicate things either. But this has a knock-on side-effect afterwards, when novices glue together patchwork Blueprints based on snippets that soon become impossible to maintain. This is why monstrous spaghetti Blueprints get ridiculed on here.

One example of video tutorial 'gotchas', is over-reliance on Casts (vs Interfaces or ED's). Casts are routinely used without explaining anything. So as soon as the viewer is done watching and they have to figure stuff out on their own (using the right-click context menu), they're completely lost about how to hook nodes up. You see this all the time on the forums with the same questions again and again. Then you look at the screenshots which show wires are only half-connected on the left-hand-side of the Cast node. FFS! That's tutorials for you. Beware.

Overall, practical working projects are an important learning method that Epic tends to overlook. But working projects have serious advantages, in that they can be upgraded and keep working over time, whereas tutorials just break once the engine changes. In general for quality learning projects, look for better-rated Marketplace packs or follow the devs listed here for free community projects...

[MarcusRivers]

i agree that most youtube tutorials are no good. thats why i very much appreciate mathew wadsteins resource.

i spend a good 8 to 12 hours a day working on my projects, which involves mostly prograamming but also modelling, texturing, sound design, etc, and an hour or two watching wadstein to get ideas for how to improve my techniques and learn nodes. i dont use it all, but many times i have discovered a better way to approach a task from watching his videos, and hopefully the rest sticks so i know what to go back too when its relavent for a problem i wish to solve.

im hoping that when i am fluent in blueprints and know enough of the api that i can transition to c++, not for performqnce, but ease of code management. it takes some skill to keep bp under control once you start building realworld object and classes, but it is super good for trying out ideas, and the performance is actually very good when nativized and built.

would be nice if there was a c++ script and compiler built into the engine though.

but i will keep saying it, the elephant in the room with ue, is the nasty interface and mind numbing bugs. i really hope epic fix it as it could be brilliant if its as fun to use as the results it gets.

[TJB1973]

well, ive just finished my 3rd month in unreal, and i am hopefully going to finish my first ever game this week and its far beyond what i thought it would be. ...lol. it wasnt planned, as im learning ue for archviz as well, but i figured if i could program a game the i can cope with the basic programming needed for archviz user interactions. i was just actually messing around with physics as im going through all the major areas of ue i think would be good to know for ue projects.

i may not spend much time with things like character animation, bones, mixamo things, but its still useful to know how to control characters and cameras etc and have them interact with a player and items in the scene.

my thoughts on ue.

if i did not believe in the final output quality of ue, i would stop immediately. ue's inferface, crashes, bugs are all horrendous to be honest. its a hell of a steep learning curve, and the program feels permanenttly in beta, and has crazy thought proceses behind the way it wants you to work.

i guess im spoiled coming from a background in solidworks and cinema4d, both i love for their pleasant interfaces and methodolgy. ue just makes me repeatedly swear at the screen several times every day, but i persevere because im betting that in a year or two when graphics cards come out that have enough power to use real time raytracing properly, that i will be expert level, and offline rendering will essentially be dead.

i really hope ue has matured enough and the insane bugs fixed by then and epic spend some time fixing the ui rather than relentless feature development, which i am beta testing for them apparently.

starting right now, i would probably not do much different, except not spend two months building a whole house...a couple of rooms would have sufficed. lol. but i did learn the workflow. for sure i spent hundreds of hours testing global illumination settings and finding workable setting, rendering flaws and methods to combat engine compromises. hundreds of tests with file format settings to import scenes. c4d datasmith is **** for me so i use fbx now, but that has a whole series of agonies attached to it.

learn the main areas you will need for archviz

materials, material instancing, parameters, texturing and uv manipulation inside and outside the engine.
lights, lighting, global illumination, and related settings and workarounds for problems.
widgets, and 2d user interface element and control of them.
the internal object heirarchy, controllers, pawns, actors and how to control them.
physics, physics materials, simulations and collisions, and UCX collision meshes.
blueprints obviously.
the sequencer for animating smooth camera motions and making stills/videos
emitters, cloth simulations are cool to know too.

i would spend 75% of my time learning blueprints, familiarize myself with as many nodes as possible, and programming concepts in general, especially classes, inheritance and interfaces, and some common design patterns, the other areas are relatively easy, you will save so much time in the long run if you know how to program efficiently and reliably interaction and communication between items and things in your scene, because its incredibly easy to make a huge unworkable mess.

obviously i can only say what i think and know after 3 months of ue, there are whole areas of the program i have not touched yet, but i think the chaos destruction, niagiara, skeletal animation, hair, are not fundamental to archviz.

i guess it depends on your final goal, if youre solely looking for still visuals, i guess the programming is not so important. im focusing on interactive visual demos, movies and mechanical simulations, so my needs may be different.

good luck and stick with it.

Marcus,

Thank you so much, this is absolutely priceless to me and appreciate so much that you took the time to share your understanding of UE and in a honest, raw fashion. I appreciate that.

I have copied and pasted your suggestions into my notes, I am starting with the fundamentals in the online academy and have come across several of the points you brought up which I will pursue even further as they provide better direction.

I would be pursuing still, motion and anyway to enhance the visual experience for the end-Client, to deliver a highly valuable message...that is broadly the goal and deliver it as a product service.

I have my work cut-out for me, although, I can dedicate learning as much as I can over the next 6 months to contribute to an eventual success.

Thank you for the good luck wishes and sticking with it encouragement as I am older now and moving into a new creative career can use all the pep talk it can get.

Thanks again!

TJ

[TJB1973]

If you have not done so already check out the online learning portal.

https://www.unrealengine.com/en-US/o...arning-courses

Like MarcusRivers pointed out this could really help you get a better understanding of how this all works.

Hi Sam,

Thank you for responding, greatly appreciated and, YES, I have been going through the online learning models and learning so much and with the first modual, "Getting Started With Unreal Engine"

Working with others, building better meshes and textures, RGB Mask Packing, MIPS, triangle counts, Material IDs, LODs, FBX Imports, Master Materials, Instances...on and on...lol.

It is great, going to go through them all as I feel it is a good start...

Thanks for checking in on the post, see you around!

TJ

[TJB1973]

That's great when it works... Materials / FX / Landscapes / Particle videos can be super helpful. But I wouldn't recommend learning Blueprints or Multiplayer or AI this way EVER! Overall video tuts aren't for everyone anyway, as they can be an incredibly slow and fragmented learning process, especially when videos are out of sync with the engine (which has changed meantime). Whereas dissecting working projects is far more powerful learning (if you can work that way). Its just more practical and immediate and ultimately more useful, as it helps speed up work on your own game projects etc...

Overall, better insights come from experimenting and trying things out, as it just gives you more confidence. Videos that cover Blueprints or visual coding especially, tend to teach a lot of bad habits or wrong techniques, because they only have so much time to 'fit the message in', and can't over-complicate things either. But this has a knock-on side-effect afterwards, when novices glue together patchwork Blueprints based on snippets that soon become impossible to maintain. This is why monstrous spaghetti Blueprints get ridiculed on here.

One example of video tutorial 'gotchas', is over-reliance on Casts (vs Interfaces or ED's). Casts are routinely used without explaining anything. So as soon as the viewer is done watching and they have to figure stuff out on their own (using the right-click context menu), they're completely lost about how to hook nodes up. You see this all the time on the forums with the same questions again and again. Then you look at the screenshots which show wires are only half-connected on the left-hand-side of the Cast node. FFS! That's tutorials for you. Beware.

Overall, practical working projects are an important learning method that Epic tends to overlook. But working projects have serious advantages, in that they can be upgraded and keep working over time, whereas tutorials just break once the engine changes. In general for quality learning projects, look for better-rated Marketplace packs or follow the devs listed here for free community projects...

ClavosTech,

Thank you for sharing your perspective, the insight with breaking down already developed projects and reverse engineering them to understand the process workflow in a more refined state, that is a great idea and appreciate you sharing that with me.

I am going through the fundamentals, bookmarked the link to the developers projects and will visit them soon. My first week was gathering resources and going through the fundamentals. I am finding I need to "get-inside" UE more so, get my hands dirty on the engine and begin to tinker much more so with the components.

Thanks again for sharing your view, quite valuable and encouraging.

TJ

[TJB1973]

i agree that most youtube tutorials are no good. thats why i very much appreciate mathew wadsteins resource.

i spend a good 8 to 12 hours a day working on my projects, which involves mostly prograamming but also modelling, texturing, sound design, etc, and an hour or two watching wadstein to get ideas for how to improve my techniques and learn nodes. i dont use it all, but many times i have discovered a better way to approach a task from watching his videos, and hopefully the rest sticks so i know what to go back too when its relavent for a problem i wish to solve.

im hoping that when i am fluent in blueprints and know enough of the api that i can transition to c++, not for performqnce, but ease of code management. it takes some skill to keep bp under control once you start building realworld object and classes, but it is super good for trying out ideas, and the performance is actually very good when nativized and built.

would be nice if there was a c++ script and compiler built into the engine though.

but i will keep saying it, the elephant in the room with ue, is the nasty interface and mind numbing bugs. i really hope epic fix it as it could be brilliant if its as fun to use as the results it gets.

It seems more and more it goes to the code of UE, the programming aspect and I am trying not to get discouraged as I am unfamiliar with C++ and willing to learn.

Thanks for drilling this down, I need to know these things and it reinforces what I already thought in terms of truly being effective in UE with consideration of being able to programming.

Thanks again!

TJ

[MarcusRivers]

I wouldnt worry too much about learning the C++ API at the moment, from my own tests, you can do everything ArchViz related using blueprints, which is far quicker and easier to implement as a beginner, and the performance is top notch in reality.

Obviously general coding and logic is useful for designing any kind of code whether that is blueprint or C++ or python.
Another thing for your list of essential I forget, is definitely Post Processing and Post Process Volume.

Well, I finished my game test, and I think I know enough programming now to achieve everything I nees, so its back to ArchViz for me, so i will post here if I think of anything.

This is my game BTW, took about 2 weeks from my firsts physics test, to initial game concept, then complete rewrite into a respectable blueprint code base. No C++ here, and im pleasantly suprised with the performance. You can see my other experiment videos on my small channel if youre interested.

https://www.youtube.com/watch?v=cAAXyiUixMA