Embedded Files
(This is a deprecated blog that transferred from another website, some images are missing)
Project Reflection and Future Goals
In this project I learned a ton, and had to pivot a lot however I also feel prepared for my senior project, which I have also started a blog for! While I have only this one environment to show I learned a lot that I either can't necessarily show, or would love to show in my next environment focused senior project. I did my best to include every thing in this blog and cite the most influential resources whilst not bogging down and burying what I deemed the most notable or important notes in a list of less important or easily searchable information. While I certainly wish I could have accomplished more, I feel confident in my use of Gaea, and UE5 for the purpose of creating height maps/environments and cant wait to make more. I also have my YouTube tutorial links posted above in a list with their brief description showing my entire workflow from start to finish.
I also want to give a huge shoutout and thank you to Professor Bridget Gaynor as this Independent Study would not have been possible without your mentorship, guidance, and encouragement. To everyone else along the way that has given me feedback or suggestions, I thank you as well!
WEEK 9 & 10
8/21-24/24: I have now moved back into a dorm at the SCAD Savannah campus and started training with my amazing staff for the 2024-2025 school year, which is the reason I have chosen to combine these two weeks.
Progress at a glance includes:
Further Landscape Spline exploration
Adjustment of the landscape texturing for alteration of the generating grass maps for the purpose of rendering
Adjusted foliage
Additional fog cards
Render tests and dialing in of console variables
UE "Stack Overflow Exception" trouble shooting
Updated Camera angles/animation
Moving from top to bottom on this list, Landscape Splines: I found that for my purposes of creating a waterfall, this was not the best solution, or rather not the best execution of this technique. In the video I linked in week 8 this technique was used on a water fall that was maybe around 15 feet tall as compared to my maybe 300 feet tall waterfall. It also makes the instanced mesh (and thus repeating texture) super obvious.
For this reason the waterfall looked very flat even after layering splines together. I think a better way to approach this would be to either shorten the waterfall, create separate "landings" effectively breaking the waterfall into smaller pieces, or to create multiple meshes with different animated materials that can help achieve a more 3D look as compared to just a 2D card look, plus help to break up the repetition. I also think making my cliff face a bit more concave to create a gap between the water and the cliff may have helped. This is certainly something I would like to put more time into in the future to accomplish a better result.
Next on my list is the adjustment of my landscape texturing. I found I would get consistent crashes when rendering due to the Brushify plug-ins generated grass maps all attempting to generate grass at once across a massive scene. To fix this issue I tried repainting my terrain to include less areas with the potential for spawning said grass. While this did help optimize my scene overall, it did not fix the render crashes which would occur before the first frame even completed. The solution was to add a game override to my render que settings. The setting was to disable "Flush Grass Streaming" which would help to stop the crashing but lead to a new issue instead.
The new issue was that it would also cause my foliage to not produce its leaves as can be seen below in my most recent test render (mannequin was in a standing pose for scale reference but would fall when rendering).
My grass looked great but my trees were bare. From what I could find online the best way to fix this was to add warm up frames to allow the cache to re-cache since we were effectivly clearing it when rendering. I put my warm up frames as high as 1500 (which again helped the grass to fully generate) but still had no tree leaves until I changed angles to the one shown below.
I have not been able to identify the solution to this issue yet as unfortunately my file would no longer open on my computer after the first work session in which I attempted to solve this issue amongst other general render trouble shooting. This leads me to UE "Stack Overflow Exception" trouble shooting I mentioned above in my progress log. That error message popped up every time I attempted to launch UE5 prohibiting me from even being able to select a level or attempt to further diagnose the issue. I attempted to create a clone of the file and in my multiple attempts I would say cloning the file was successful about 60% of the time meaning that it would or would not crash and not open when copied. When cloning the project I would get the message "Unreal Engine does not fully support cloning of code projects..." or something similar which was interesting as I had not done any coding work in this project. The closest I had to that was adding Blueprints etc. from plug-ins I added like Ultra Dynamic Sky or Brushify. I later found the Brushify specific documentation that mentions how to fix the problem with the grass and to include warm up frames like I mentioned above confirming the steps I took but not offering a solution for the foliage.
For the "Stack Overflow Exception" error a few steps were taken for trouble shooting. The only one that was fruitful was opening the file on a new computer as cloning the file (when it would actually open) resulted in a version of my file from roughly 1.5 weeks ago despite the modification dates, and file sizes being identical between the original and copy, as far as I could tell. The file opened on Professor Gaynor' system and looked to have all the progress that the clones were missing. I could not find any documentation to address this issue online as most people talking about this error were either coding or talked about modifying .dll files. The only other suggestion was that the quantity of blue prints were too high though the file running on other machines as well as the fact it worked on my machine, until the following morning, indicated that is not the issue I am facing. For these reasons I do not currently have an updated fully complete render due to the fact that Montgomery Hall is closed this weekend, and none of the computers in the Hive have the software nor capabilities to utilize the software that I need. Once fall quarter starts however I plan to render this and update my docs here with what I have/how I solved the foliage issue mentioned above.
Now I would like to briefly discuss and document the console variables I utilized for the rendering of this project.
These console variables are a combination of what I have seen in other successful projects or recommendations from people like William Faucher or Clinton Jones. I am going to skip over the self explanatory console variables (or CV) like r.MotionBlurQuality and focus on the less clear ones.
Shadow CVs: Most of the CVs that effect shadow are used to increase the quality and disable the default denoise to improve render times and stability especially since I do not have any closeups on areas in shadow which would allow me to increase quality in other areas.
Foliage.LODDistanceScale: This CV allows me to control the fall off range in which foliage changes from LOD 0 to LOD 1 etc. Since I am not worried about game performance I could crank it up, the same applies to the following CV
grass.CullDisanceScale: This is similar to the above foliage CV however it controls the distance at which the grass will generate in render. This CV will ONLY WORK WHEN USING GRASS SYSTEMS. I know many of my classmates have placed grass using the foliage editor which is perfectly fine, especially for smaller scenes, but with the Brishify plug-in, as well as when optimizing grass for games or a larger scene, grass systems are in my opinion necessary.
WEEK 8
8/10/24: It's been a while since my last update but I have a lot of progress to share which is really exciting. Since my last post where in I ultimately decided to move away from Houdini I have fleshed out my environment within Unreal Engine including many areas.
Progress includes:
Texturing of environment
Extension of environment (Distant mountains, fog etc.)
Scattering of foliage (going to revisit and add more detail)
Creation of custom water material
Creation of water fall
Exploration of Niagara particle systems
Adjustments to landscape mesh
Lighting set up
Camera angles and first pass of animation
In all I'm real happy with the progress so far and feel that I am in a great spot to create a polished environment. Below is a slideshow starting with the first progress photo I took and ending with my most recent. (Also please note, this is all screenshots of the viewport and not actually full renders!)
In the final screen shots you can see the current waterfall prototype which uses a landscape spline to instance a static mesh with a animated water fall texture on it. That texture and technique come from the same video (https://www.youtube.com/watch?v=FPJJIFhmJ4o) that taught me how to make the custom water material. Both of these materials have built in parameter controls that allow the user to modify the materials quickly in most any way they could want. Below are the images of these material networks and controls as well as what the landscape spline with this instanced mesh look like.
The water fall is of course not in its final form at the moment but I have confirmed that this technique will work for my needs and create the desired effect I wanted. In my testing I also tried to do a Niagara water fall such as in this video: https://www.youtube.com/watch?v=d3lcHp197Mc Ultimately I did not get the result I wanted and it was also going to be far too heavy for the scene to be worth it as compared to using 2D cards are in this case, instanced geometry with a animated texture. The Mesh that I am instancing is 300x300x10 units so it is square for the purpose of tiling the texture but not completely flat so when viewed from an angle some thickness is implied but no specific texture details are revealed.
As can be seen in the parameters of the water material, it is highly customizable, which, going forward, makes it an amazing tool to continue using. It also detects the edges of static meshes (including the landscape), which allows the texture to create ripples from the edges of these meshes. This helps to both ground the mesh in the scene and break up the normal map of the water so it doesn't look like it is just traveling through the landscape and any meshes that may be in the water. While this was not my own original creation, following along the two-hour-long video tutorial taught me a lot about materials in UE 5, as well as what nodes are available to me and how I can utilize them. Below is an image of the ripples I discussed.
My goals for the rest of this week moving into next week are the following:
Fix foliage to include variation in scale and rotation
Spend more time scattering the foliage so it looks more realistic to how it may work in real life
Scatter some cliff faces to break up the repeating textures in the cliff sides etc.
Place rocks and other boulders where the water fall will go to shape the flow of the water
Finalize the water fall and add some animated cards etc. to blend the water fall with where it meets the water and where the water flowing downhill meets the lake.
Tweak animation on cameras
Conduct test renders of the scene to dial in settings
Dial in LODs for foliage etc. as needed.
Bonus: Add a boat or some other static meshes to build out the scene a little bit farther. Maybe one of the camera angles is a campsite over looking the lake?
Next week on the 15th, I start my drive back to SCAD to begin training for my new role as a Senior Resident Mentor (formerly known as a Community Assistant), which will take up some of my time (especially going into week 10, unfortunately). For those reasons, I have chosen to focus entirely on making this environment extremely polished and a great piece for my demo reel. I will also be making a series of YouTube videos chronicling my experience working on this environment and walking the viewer through how to make an environment like this themselves. I want these videos to be as concise and informative as possible, so I plan to use week 10 mostly for these videos and presenting what I have learned.
While I have not completed what I set out to do initially (create 2 environments), I have still learned a ton. While I am saddened I did not accomplish this goal, partly due to my computer issues, I am proud of what I have accomplished and learned. I know that my next environment will not take me nearly as long to get to where this one currently is. I have learned a lot about Gaea, Houdini heightfields, environment texturing, Unreal Engine 5 workflows/practices, real-time materials, environment lighting, and procedural content generation (which I would also like to explore further). I will keep this blog updated as I wrap up these last two weeks and will be sure to include the links to the series of YouTube videos I post as well!
WEEK 7
7/31/24: Quick note for exporting Gaea heightmaps as far as file formats into unreal- here's an excerpt from the UE documentation
"Unreal Engine supports importing heightmap images using the following formats:
16-bit grayscale PNG.
8-bit grayscale r8.
16-bit grayscale r16."
7/30/24: Currently I have spent far too long trouble shooting the UDIM setup in Houdini and have decided to place a emphasis on on the real-time work flow in Unreal Engine 5. Issues with UDIM aside I also have burned up too much time as is and waiting to render it out and trouble shoot that will take even more time I could've used in UE5 to create 2 Environments. For these reasons I will be stepping away from Houdini for the remainder of this Independent Study. That being said, check out my Senior Project Senior Ideation Blog where I am exploring concepts for a future environmental story telling project in which I may return to Houdini. For now, I will be importing this same height map into UE5 and texturing within it as well. I also want to delve into Niagara a bit and take a stab at creating the water fall I previously discussed, just in real-time, rather than a fluid sim.
The main issue I was facing with UDIM was actually getting my geo separated into 4 quadrants and then exporting them for use in Substance Painter. Currently, as shown in the image below, I used a polysplit node to add edge loops through the center of the geometry in the X and Z axis. After that I used 4 group nodes to separate the 4 quadrants. My issue now is how do I use these groups and separate out their UVs through a UDIM setup? I also had issues, likely due to my poly count, wherein any action I took resulted in waiting between 10-50 seconds for any feedback slowing down the testing and trouble shooting.
Aside from learning what more of the Heightfield nodes did I did also learn some helpful hotkeys when selecting polys like this. Clicking a poly and then moving down the poly loop and pressing A + LMB selects the polys in the loop up to the second face you select. This was useful for selecting the perimeter of the quadrant I wanted (interior edges). After that H + LMB selects the area that you have defined (Shift + H + LMB will keep the edges you have defined as well). These couple of hotkeys will prove really useful going forward when I use Houdini and I still feel I learned a lot about UVs and Heightfields in Houdini so I feel this exploration was still educational, just fell short of my expectations. I certainly will be revisiting this when I have more time/a longer deadline and also want to see what I can generate with Houdini's inbuilt Heightfield nodes compared to what I can accomplish in Gaea and the time difference between the two.
WEEK 6
7/27/24: I ran into some issues with UDIM setup as well as exporting the height map from Gaea which was just some user error on my part. At this point in time the normal "Output" node is inface the correct node to export the best, high quality map at the end of creation in Gaea. The height node would be more useful for isolating a specific part of the map based on height perhaps to be used to create a map that can simulate a water line or other use. On the Houdini side of things, the previous videos I mentioned by Rick Banks in Week 2 Were not helpful due to the use of Labs nodes that no longer exist when it comes to the UDIM set up. The functionality that node had has been included in the UV Layout node which you must select the target type as UDIM Tiles and change the Packing Scale to Fixed.
An issue I ran into with the set up however was getting the UVs to appear at all after importing my map to Houdini. There is a node called Convert Heightfield which is needed in order to make the heightfield into an actual piece of geometry that has UVs. I hadn't realized the issue initially but once I did it was just a matter of finding the node I needed.
In addition to that information I wanted to document a quick hotkey I learned: when in the viewport Press 'space' + 1, 2, 3, 4, or 5 to switch between orthographic views, UV view, and the tumbling viewport.
So in short, I'm currently trouble shooting and doing some additional research as most of the videos or other documents I find are 3 years old and use Labs nodes that have been removed/merged into other nodes. currently I believe I have my UDIMs working (from a layout perspective) but have not segmented the terrain yet though I may have a way of working around it? Once that is completed I need to figure out how to export them and bring them into substance appropriately.
7/22/2024, Texturing and Scattering: Like I mentioned at the end of my week 5 post I am going to get rolling with the UDIM texturing work flow in Substance Painter. I'll update this all accordingly to reflect my progress in this venture. For now though, here is some of my texture reference:
As far as the water simulation I mentioned, I want to do some small render tests this week to see how long a render may take so I can make a better informed decision as to whether or not the larger waterfall is achievable in this shortened time frame, especially with how ambitious I want to be with the UE5 environment.
On the topic of scattering, I plan to use the work flow I learned while working on environments for the Rider project where in I select some assets (or make them) and use an attribute paint node to basically weight paint where I want those assets.
Update 7/24/2024: My Biggest "discovery" of this week has been the height node. I currently need to experiment with it a bit more to export my desired map correctly however it does seem to create a better map than just marking the last node to effect the "geometry" of the scene. I'm not sure how none of the tutorials I utilized talked about this node but I'm glad I mistyped and put this node into my project by mistake. In conjunction with the details retained in the normal map I think I should see some good results when importing into Houdini' heightfield system.
WEEK 5
7/20/2024, Height Maps and Gaea Exploration: This week I finally got my machine back in working order after replacing the motherboard and processor. I decided to take the week making some height maps and putting my Gaea knowledge into practice. I first decided to take the height maps into Unreal Engine 5 to test the differences in map resolution, figure out the scaling that was needed to properly import the maps, and test what albedo maps would look like. In general I was disappointed with the results as far as a terrain that would be explored by a camera goes. If the environments were shown outside of a window or as a distant mountain range etc. then they work excellently without any additional touch up.
(Landscape 1 is 2K, Landscape 2 is 4K)
Resolution changes between 2k to 4k showed no real noticeable difference for either the albedo map or height map aside from taking more resources from the engine to produce. When I get to the UE5 environment I will look more into this and see if it is something I can improve upon further.
There is a trick for proper scale when importing the heightmap: In Gaea in the build menu there is a height and a scale slider that by default are 2,600 and 5,000 (these numbers meaning a max height of 2,600 meters and an area of 5,000 x 5,000 meters). these number along with the resolution are the magic numbers needed to calculate the scale. 512 meters is Unreal Engine's default max height so to calculate the X and Z we take: 100 x 5000 / 1009 (1K resolution for UE, adjust according to resolution). then finally the height would be 100 x 2,600 / 512. These formulas can be plugged straight into the engine. I have found height still feels weird so I am going to experiment further with that though I do believe some of that is due to using water in Gaea (in this case rivers) which leave these strange flat spots and seems to cap the - Y value. more info here: https://www.youtube.com/watch?v=EFXIm4TmMKI
Now for the height maps for the first environment inspired by Geirangerfjord:
In my experimentation I've found that the albedo maps really don't hold up to scrutiny under a closer view so any albedo maps I make in Gaea will purely be for reference purposes. I plan to use the UDIM work flow I discussed previously which I'll be setting up in week 6.
My goals for week 6 include:
Finalizing height map
Creating textures and assigning them accordingly
Scattering trees, rocks and other assets
Create the water (whether it is a textured plane or otherwise)
Look into water simulation in Houdini (maybe start the previously mentioned water fall if I'll have enough time)
In my experimentation I've found that the albedo maps really don't hold up to scrutiny under a closer view so any albedo maps I make in Gaea will purely be for reference purposes. I plan to use the UDIM work flow I discussed previously which I'll be setting up in week 6.
My goals for week 6 include:
Finalizing height map
Creating textures and assigning them accordingly
Scattering trees, rocks and other assets
Create the water (whether it is a textured plane or otherwise)
Look into water simulation in Houdini (maybe start the previously mentioned water fall if I'll have enough time)
WEEK 4
7/13/2024, Photogrammetry, Substance Painter Research, Film Research: Once again before I get into the content, a computer update. My computer repair and thus work within Houdini and Unreal etc. has been delayed by two things, a hurricane hitting and a new currently undiagnosed issue with the machine. Hurricane Beryl took out power and internet for several days delaying both my self and the shop from working on the machine and now with a brand new motherboard (purchased as Asus is taking me in circles with the RMA claim) the system does not post.
Computer issues aside, the content for the week so far- I finished the tutorials (https://www.youtube.com/watch?v=GbVVa-BBef8&list=PLNbgmFvU__fgr5_45ZLgU8nZtg_ZnOSOS) for environment texturing and the substance painter side was heavy. Working with UDIM in substance seems to be convoluted to a certain level and I'm sure will take me a little bit of trail and error to make sure everything works as intended on my end. It will be a fun challenge though and a super powerful skill to add to my tool belt. I also have decided that due to the loss of time with computer issues I will work on the Geirangerfjord based environment first and then the snow globe. Once I'm happy with those I will circle back to the big bend environment if there is still time as that is honestly the least complex environment in this list.
Photogrammetry: I am going to try my hand at some photo scanning using a wooden reindeer that my dad makes in our garage as a hobby as he is a wood worker. I think it would be fun to include them in the winter snow globe environment either as a decoration on the table next to the snow globe or a small huddle of them nestled into the snowy hill side within the snow globe. I will update this section accordingly with my progress.
7/14/24: here is a screenshot of the photo scan in progress. Once I have access to a machine with the appropriate software again I'll take a look at it and so some clean up, it seemed to come out well as far as I can tell though.
Film Research: Alongside this project I have been working on some ideation for my senior project next school year (Also posted in a blog here on my website). My goal for that project is to learn more about environmental story telling and create a sequence in which a camera explores an environment and the viewer and gleam a story from what they are seeing without any dialogue or real character beyond what personality the camera will have. Since there is a strong connection between that project and what I am doing here I wanted to talk about my findings so far.
The films I watched include (any film with a * before its title, I plan to watch more from that cinematographer/Director of Photography):
* The Equalizer, (2014) DP: Mauro Fiore
The Equalizer 3, (2023) DP: Robert Richardson
Resident Evil, (2002) DP: David Johnson
Resident Evil: Retribution, (2012) DP: Glen MacPherson
* Resident Evil: Death Island, (2023) DP: Yun M. Watanabe (CG Film)
Extraction, (2020) DP: Newton Thomas Sigel
Extraction 2, (2023) DP: Greg Baldi
Beyond The Wire, (2021) DP: Michael Bonvillain
* Leon: The Professional Trailer, (1994) DP: Thierry Arbogast
The main thing I was paying attention to in these films was the way the camera was utilized and the way the set was dressed and what that may convey about the characters and their actions, personalities or thoughts. The three films I chose stuck out to me for these reasons with Resident Evil: Death Island probably being the weakest of the three examples however, I think it would be an interesting exercise to look at more CG films such as that one and see what they may do differently than traditional animated or live action films. In this film there were and abnormally large number of establishing shots compared to many other films I have seen but I think that is one of the benefits of a CG film, you don't have the same costs associated with filming a helicopter shot to capture a city. The cost between shots was probably fairly similar in the case of this film.
Denzel Washington' character Robert McCall in The Equalizer provided a lot of exploration both into camera use and environmental story telling that I found interesting. McCall keeps an extremely clean an meticulously organized home and schedule and It is implied he may be OCD or something similar. We see this through both his habits but also his home which is even commented on by some home invaders later in the film. Environment aside, I liked the camera work in the first half of the film and I feel the DP did a great job conveying Robert' emotions and feelings through the use of camera. I look forward to watching more of his work.
Leon: The Professional Is probably up in my top 5 favorite films now. I loved the way the film opened both with the travel through the city streets but also the meeting between Leon and his employer, the assassination, and the details in the morning routine of Leon we get to see. The use of shadows, detail like the ratchet straps to aid Leon' workout and the shaky nature of the city street shots in the beginning revealing themselves to be Leon's own perspective at the end of the film were prefect. I also really like many of the actors and actresses in this film so I'd be interested to see if this particular DP has any other projects with them.
WEEK 3
7/3/24, Final Reference Collection and Additional Research: Before I get into the content I've accumulated since last week, another computer update. according to ASUS the repair to my motherboard should be completed today and over night shipped back here so the hope is that I get a working system by Friday.
So the first thing I did after my research into the UDIM setup in Houdini to then texture within substance painter, was look at the environment I am trying to create and how I want to split it up. The first environment I want to look at creating is Geirangerfjord as due to the time constraint presented by the lack of a operating machine at this point in time. The goal is still to create three environments but if needed I will cut the Big Bend National Park environment to save time. I decided that due to the abundance of water and vegetation in this environment I could keep most areas at a lower resolution map as the scattered vegetation will help reduce the visual intensity of these areas. I would have my environment split into a 3x3 grid with most of this area being a 2K resolution map with the exception of less visible areas that I may want to lower to 1K if possible. I am planning on 2-4 4K maps either in a grid formation or two separate hero shot areas where I will get closer to the physical geometry with the camera. One of these areas will be the water fall that I would like to try and include.
Texturing Notes: I did some more research into Geirangerfjord and found the primary stone in the area is peridotite and serpentinite. I also noticed in some of the references a sort of jagged banding present in the exposed rock faces that would have been exposed by glacial movement.
The Snow Globe Environment I will try this same method if I do not find a better method before starting that environment. My best guess is that I will need to try something else or a tool within Unreal as I don't think a 2-4K texture map dense area will run well but we will find out.
An Update on the Water/Erosion exploration: I plan to make a before and after of an environment flooding as an exercise within Gaea as well as an excuse to learn more about erosion. My father likes to watch Gold Rush Alaska on TV at night sometimes and while I was watching it with him they actually spoke at length about the erosion caused by glaciers and then how rivers deposit sediment etc. which for the purpose of a Fjord like the one I would like to create is perfect.
WEEK 2
6/25/24, Additional Research: Pending the return of my machine to its former glory I conducted more research this week over a few key topics. Texturing Environments in Houdini with UDIMs, how to utilize GIS data to create height maps, and following a conversation with Professor Gaynor, an exploration of what an area may look like after flooding.
here is my primary reference for environment texturing in Houdini utilizing Substance Painter and UDIMs: https://www.youtube.com/watch?v=GbVVa-BBef8&list=PLNbgmFvU__fgr5_45ZLgU8nZtg_ZnOSOS
This is a video playlist on YouTube by Rick Banks uploaded (at the time of writing this) approximately 2 years ago.
This is a excerpt from Rick Banks Video linked above that shows the principal behind UDIM, as well as explaining some of the benefit of using them. Now as I understand UDIM I could choose based of camera position in a scene etc. which area of the coordinate grid should be which resolution allowing for parts of the scene closer to the camera to be high resolution and those less visible to be lower aiding in optimizing the scene.
Getting started with UDIMs is actually surprisingly easy. After importing the original height map or other method of generating the environment you just need to hook up a Labs Terrain Mesh Output node. As seen in the image below (also from Rick Banks) you get controls to set up the tiling of the environment.
From here the name of the game is optimization for use in Substance Painter. in the example provided above by Rick Banks the scene is roughly 4,000,000 polygons. In the Meshing controls which are located right below the tiling controls mentioned earlier, changing the "Base Density" control will lower the poly count but also the detail present. The work around for this is to bake a high poly map with all the normals and displacement preserved so that those maps can be reapplied to the low poly version we will use in substance painter. From there in the UV Layout node options, in order to take advantage of the tiling that was just set up, we need to adjust the packing scale to fixed which lets us select the scale (1.99 in this case, 2 pushes the UDIM cells into the negative/invalid values to my understanding) and change the targets to pack into to UDIM Tiles. From here its just exporting and baking the maps necessary for later use, and use in Substance Painter.
WEEK 1
6/19/24, IDEATION: Throughout the day, I dedicated my time to reviewing reference images and selecting the ones that I believed would offer the most assistance. Simultaneously, I established varying levels of detail for the three environments I intend to create. There are several reasons behind this approach. Below the text explanations will be my reference images and mood board.
The initial environment will draw inspiration from Big Bend National Park, situated on the border between the United States and Mexico in Texas. Opting for a drier setting with minimal vegetation, I aimed to concentrate on mastering the texturing of the environment in Houdini using Substance Painter and UDIMs, rather than focusing on scattering and modeling trees. Furthermore, I have personal experience with this location, having visited it during an Eagle Scout camping trip/reunion with friends two years ago. I have numerous photos I took during my visit to the park, which I intend to incorporate into my references once my computer is returned from repairs.
The second environment is going to be modeled after Geirangerfjord located in Norway. This scene will have thicker vegetation that will be more necessary to include and cant be cheated in texturing as easily. I also want to approach water sims in Houdini and see if I can include the Seven Sisters Waterfall which is located within this fjord. This will be an extra I will try to add but may not dependent upon time and the result as I have little experience in this area within Houdini and am aware of the vast resources required to do a water sim which ultimately is not the focus of this project.
For these two projects I plan to look at satellite/topographical data and seeing if I can utilize any of it alongside GAEA or at least to use as reference for these projects since they are based on real places. It could be interesting to do a side by side of the generated height map and the real life data just to see how closely I can match the reference in terms of erosion and such as a 1:1 recreation isn't necessarily the goal of these projects.
Now for the third and final environment which will be created in Unreal Engine. For this project I plan to create a snowy Christmas village scene within a snow globe and take what I learned from the other projects both into real time and a more abstract setting not based in reality. I realized when looking for reference that miniatures of course have a very different look and feel to them than their full scale counter parts. With that in mind I'm not yet sure how I want to approach this project in terms of realism. Do I want it to look like a miniature like a real snow globe? Or do I want to make the interior look more realistic and thus fantastical and less grounded in reality? I'll discuss with my peers, friends and family and collect some feedback and make a decision a bit further down the road as this will be the final environment I create of the three.
I plan to utilize https://earthexplorer.usgs.gov/ to see if I can download images that can be used as height maps in Houdini or modified in GAEA from the real life areas that I am referencing. At the time of writing this however the sight is down for maintenance so I will update this when I am able to experiment with it.
Page updated
Google Sites
Report abuse