Author: Orrin Wilson

It makes me sad to think that this will be my last blog post of the Summer with DTSF. This whole experience was amazing for me and I learned a lot about VR and self learning. I hope to take both of these skills and apply them in the future. There is a lot of potential with Blender and Unity. It was very fulfilling to help recreate the 17th century Spanish theater and I would love to pursue more projects like that one. I also got to learn a lot about the college history during the civil war and it was fun virtually restoring the past iterations of Penn Hall. I am pretty confident that I can create anything in Blender at this point. The only problem is being able to create an item with a low enough polygon count that doesn’t crash the computer.

Outside of VR, DTSF was able to immerse me in a whole plethora of new technology that I didn’t even know existed before this internship. Stuff like 3D printing, Arduino, Raspberry Pi, and photogrammetry all seem very fun to learn and I am open to looking into them for future self-guided projects during the semester.

Overall I wasnt able to accomplish my original task this summer but that doesn’t mean this summer was a waste. It was quite the opposite in fact. When I first arrived in week 1 I was looking for the best way to persue my project. I was split between photogrammetry and 3D modeling. Both are compatible with Unity and VR and photogrammetry is much faster in making models so it was very hard to not go with photogrammetry. With a drone I would most likely be able to put together a model of the entire campus within a week or so and spend the rest of time adding cool VR interactions. Going this route would allow for a lot more time in Unity as well.

After the second week, there were three reasons why I decided to go with the more tedious route and why you should too. The first reason was the distortions close up. Photogrammetry models always looked good from a distance but up close they usually became blurry and distorted where the photographs were lacking. This would be especially visible in VR when the user is at a ground level height looking up at a model that is filled with blurry colors and distortions. While this may be useful in some kind of background setting, it would look very sloppy to use these models in your main scene. The second reason was the 2017 DTSF class already had 2 photogrammetry developers. Part of the program is to be useful to the college as a resource if needed, and my project would have been very similar to Ryan Gonzalez if I went the photogrammetry route, especially because he already modeled most of the large areas of campus. The final reason was the skill set I wanted to develop. Coming out of the summer I wanted to have the ability to create any virtual space I can dream of, something that is not possible with models based on things you take photos of.

While there were a lot of good reasons to go with 3D modeling, there were also setbacks that made me hesitant. I knew the 3D modeling route would be much slower and harder to learn. I basically had to go through a couple weeks of art school to get some of the necessary artistic abilities required for the 3D modeling platform. Also the modeling process is worlds slower than photogrammetry. In 3D modeling one has to model a virtual building with their bare hands. While this may seem simple at first, the more time you spend on the building the more details you realize have to get added, and it becomes a never ending cycle until you realize there is an infinite amount of detail one can add while modeling. Due to these setbacks, I decided to focus on Penn Hall; the cornerstone of Gettysburg College. Since Penn Hall use to be the only building at the college when it was founded I am technically still fulfilling my project of modeling all of campus (hehe).

My newer modified version of the project was now going to be focusing on making Penn Hall as visually accurate as possible as well as creating past models of the building to give a perspective on what Penn hall use to look like over a century ago in VR. While this may seem like a huge downsizing of my project, it was the right way to go. Little did I know my decision would heavily pay off when we went to go to the Hamilton conference. I showed up there with a lot of confidence and I also met several students and mentors who are also using the Blender and Unity combo. This was very reassuring and made me feel like I wasn’t learning a useless skill. Greg was the leader of the conference and had been using Blender for 15 years. He taught me a lot and was very helpful with our project. Even though we didn’t finish the theater, we got a lot more done than expected and it was a lot of fun to make. I hope that I will be able to continue reconstruction of the theater when the semester starts.

The last week was just two days. One was spent practicing our presentations and the other was spent presenting. Overall I felt like all of the presentations went really well. The crowd seemed very engaged in each of our talks and I got a lot of good experience with public speaking as I’m sure the rest of the DTSF team did. This was a great opportunity for me and it inspired me to try some exiting projects of my own during the fall semester. Wish me luck!

Signing out,

~Orrin Wilson

This week the DTSF team packed up and headed to Hamilton College for the DHI Conference. It was a Digital Technology Conference. Three schools attended the conference: Gettysburg, Hamilton, and Colgate. We brought the most modeling students out of the other schools. Hamilton was more interested in Unity and Colgate was using 3DSMax, an architectural program, to create their projects. The leaders of the Conference were both experts in Unity and Blender. It was a very inspiring experience to be able to learn from them and I hope to be able to use them as a resource for further questions I have in the future.

As far as the project Hoang and I were assigned, our goal was to recreate a Spanish theater from 17th century Madrid. This project was a very amazing experience and I learned a lot from it looking back. It taught me a lot about team collaboration and project workflow. The most important lesson I got from this was that things don’t always go to plan.

The first step was to figure out how we were going to split up the work. Both Hoang and I have spent a lot of time on Blender but neither of us have collaborated with another person on the same blender project. So we decided to split the theater into two sections. I worked on the front stage and side seating while Hoang worked on the back part of the theater where the people entered from. Both parts were complex in their own way. The theater has so many different levels and areas of seating at some points it was confusing. For me the backstage area was the most complex part for me because I never worked on an interior section of a model that is also attached to the exterior. When I designed the interior for Penn Hall, I created the interior separate from the exterior in a different Blender file. However while I was designing the backstage it was less intimidating that I originally thought. Hoang also had a difficult task. My section of the project was mostly symmetrical. This allowed me to use the mirror tool in Blender which meant I essentially only had to model half of the front which made things simpler. Hoang, however, was not able to use this because the entrance building is not symmetrical enough to get away with this tool.

After each of our base models were done, we began to go back and add more detail to the model. This included things such as benches, bleachers, banisters, etc. Once the specific model was made for each detailed furniture item, we then had to use the historical information Chris from the Spanish department found regarding how much furniture there was and where it was placed. Getting measurements to be as exact as possible was the most important part of this project. Especially because Hoang and I had to eventually combine our projects. This meant all of our measurements had to remain consistent between the two projects.

After the first day I quickly realized that the work was going to be much more than can be done in a week. So Hoang and I started to work nights in order to stay on schedule. However, even with this extra time, we still weren’t able to finish within the week. We were able to finish the majority of the Blender model excluding the textures.

Going forward we still need to properly transfer the Blender model to Unity along with the textures. There are also things in the model that need to be improved. There are also aspects to add in the future such as a realistic play interacting on the stage.

This was our last full week to work on our projects. Next week we are off to Hamilton College to work on the virtual reconstruction of a Spanish Theater. I am very exited to test what I’ve learned to see what I can do to help others. To be honest this week was pretty stressful for me because I really wanted to finish my project this week so I can relax going to Hamilton and only focus on the presentation on Week 8. I worked around 10 hours a day this week. I am happy to report that my project is almost done. I am 99% finished with my Modern Penn Hall scene and my 1832 Penn Hall scene is almost done as well. All that is left is to do some touch up work and possibly add some historical information the user could read. I am confident I can finish this during my free time at Hamilton or right when we return. I hope to learn a lot from Hamilton. Because the conference is mainly focused on what I have been working on this whole summer, I am looking forward to talking to experts on the topic to learn more about Unity development. After this internship I hope to be able to take what I’ve learned and try to develop something fun and useful in VR.

Getting into more specifics as to what I’ve accomplished this week, I started off the week completing the final edits of penn hall and getting it ready to export to Unity. I made this process very tedious and inefficient. If I did it again it would only take half of the time. Every time I wanted to make an edit, it would change the texture map, so I would have to go back and re-wrap the texture to the model. Right once
I got everything wrapped up, I found another flaw with the model, that could only be seen up close in the Vive. This meant I had to go back into blender and do everything all over again. It was probably the most agonizing part of my project. Next time I won’t even bother wrapping the model until I approve of everything with a Vive walk through.

One of my debates was how to create a border for my project. Because I was only modeling a 50 meter radius around Penn Hall, the terrain just went into nothingness and it made the scene look very empty. I decided to look into creating a water border. It looks very beautiful and it doesn’t distract from the main scene. To add a feel of land going into the distance, I added a ring of trees on the other side of the water.

I am happy with how the scene turned out. I plan on showing this scene to my peers to see if there is anything I missed. I am open to criticism and improvement. It is far from perfect compared to a professional development company, but for a Undergrad Project I think it is a good way of combining all the skills I learned in the past 6 weeks.

 

This week I was able to finish my new Penn Hall model. This new model is much more accurate to the real Penn hall. The roof us more accurate that before, the dimensions of the building scale to the actual building. The step count is very close to the actual amount of steps (16). There are no problems with the windows like there were before. The polygon count is significantly lower, and there are no overlapping vertices. This time I was able to make the entire mesh out of a few shapes which made it much easier to apply textures. This time I kept the mesh in mirror mode which meant if I ever needed to make a change or improvement, I would only have to do it to one quarter of the building and it would automatically be applied to the rest of the building.

After the model was finished, I had to add textures to make Penn hall more realistic. I found a good online texture for the bricks that matched up perfectly to their original. For the windows I filled them in with a dark blue material. The hardest part was the roof. There were no textures online that matched the roof, but I had an Idea. Alyssa and I went out to Penn Hall with the drone to take pictures of the roof, I used those pictures to texture wrap the roof,  while there isn’t a normal map for it, the roof still looks much better and realistic.

The next step was to build the environment around Penn Hall. The most difficult thing was the sidewalks. I didn’t think of an efficient way to do it so I made them in Blender by hand. The process was very tedious but I think I made them look decent for now. I will have to go back to them later to make sure they look good enough for the final presentation. I started putting in the trees near Penn Hall as well as the grass. The scene is starting to come together. Going into next week, I will try to get as far as I can in touching up the scene. Afterwards, I will add in the VR Interactions. I do not think this step will be that difficult because I already spent a lot of time with it and there are a lot of base scripts given to you through Unity assets.

It feels like the time spent at gettysburg flew by. I am officially at the half way point in my project. So far everything is going on schedule. I believe I completed all of the seperate aspects of the project and now is the time to bring it all together. These next few weeks are very exiting because I am able to apply what I have been learning throughout the summmer and make it into a finished product.

Most of the work done this week was spent with world building and level design in Unity. I started off by aquiring several assets that can enhance the experience of the VR user. The first asset was called Time of Day. It is a special skydome in unity that allows time to pass, and as that time passes, the light moves with the sun. The other assets were rocks, trees, and grass that helped the user feel like they were in a realistic nature environment.

I started off the first day with a forestry design tutorial. It was roughly 2 hours of content. It took me the first couple days to complete. I got to learn a lot about level design and what goes through a game developers head when making levels. There were a lot of tools in Unity I used that I was not aware of. I intend to use these skills and assets when building the environment for my project.

However, upon completing the forest scene, I ran into a problem that I did not expect. The models I was creating in blender did not have the realistic detail to match the environment. It looked like a cartoon building in a photo-realistic forest. The problem with this is when I attempted to add texture maps and detail to penn hall in the past, it never properly transferred to unity. Currently I am hand-making my own white brick texture for penn hall so I know that it will be able to transfer to unity.

Starting next week I plan to remake penn hall to take out all of the problems that I had with the first one. This new model is going to be much cleaner and it will also have the applied brick texture I was working on.

Day 1

The first day there were some mistakes with my Penn Hall model that I had to fix. There were a lot of unconnected vertices that I had to manually connect which took most of the day. Unfortunately it took the full 8 hours to fix.

Day 2

There was a lot of progress but I still need time to understand it. I went through a tutorial to learn the basics of VR development. It included features such as walking around, picking up and throwing objects, and opening doors. I made my Penn hall intractable in VR and players can open the doors to go inside. The last half of the day was spend trying to find a way to move between scenes in unity. This is one of the most important skills to learn in unity, especially for my project. This allows me to be able to walk into Penn hall and use a scene change to go to the interior.

While it is possible to create my entire project in an entire scene, it would have been quite difficult and I would have had a lot more limitations. With the scene change, I can now start a fresh new model that contains the interior. Also using scene changes is one of the most common functions of unity. It is commonly used for changing between levels in a game. A developer can use each scene to focus on a specific goal rather than compile everything into one area.

Now that I am able to do this, the next step is to model the interior of penn hall.

Day 3

I spent the morning modeling the second floor of penn hall. I tried to capture as much detail as possible. I also found a better way to change scenes through using a special teleport area. This method is much simpler. All the user has to do is open the doors to the building and move into the interior of penn hall. Then, the user is instantly transported to the next scene.

Days 4-5

The new computers came not much work was done. A lot of the time was spent setting up the new desktops and transferring all of the work from the laptop to the new location. Unfortunately, this was not very successful and I wasn’t able to save all of my work. Other than the models made in blender, everything else could not be transferred properly, such as the Unity projects. I will have to re-implement the interactions but it will be much faster this time around.

For the second week I set out 2 goals. I wanted to find a way to begin modeling buildings in Unity and I wanted them to be interactive in my project. I decided to use Blender for my modeling platform. While I am still unsure about the best way to approach this project, I decided to begin with a modeling program. As mentioned in the previous post, the other option would have been to generate a 3D model using photogrammetry software, which would have been done by taking many pictures of the area and auto generating a model from that data. There is a good chance that the photogrammetry option would have been more accurate than hand making a 3D model, but even with lots of data and strong photo generating, the models made from photogrammetry usually have many distortions from a close distance. Considering that my project will put the user at the height of an average human, the distortions will be very apparent from such a close distance, which would take away from the experience in my opinion. Secondly, one of the goals of this internship is to develop skills for myself as well as skills that the college can use to assist students of faculty wishing to learn that skill. Photogrammetry with drones and 360 cameras have already been used a lot in past projects and I also feel like it is easier to learn than hand making 3D models from scratch. It is likely that I will still attempt photogrammetry in the future because I am curious to see how the two versions of penn hall compare with each other and which is better suited for VR.

Days 1-2

First thing to do was to learn how other people created their objects in unity. While it is possible to create things in unity, many of the people I watched online were developing the objects in a modeling platform. I chose blender because I noticed it was very commonly used with unity and a lot of the tutorials I watched when learning the basics of unity implemented blender models into their projects. I started a blender tutorial with Blender Guru. The tutorial was very long. There were 8 segments of 40 minute videos. The goal was to create a mug on a table with a donut. I didn’t finish the entire tutorial but I completed all the parts that would be useful to my project. I learned how to work the interface of blender and many of the shortcuts to make the work smoother.

Days 3-5

After this I completed a second tutorial on modeling an anvil so I could refine my skills at modeling before I took on my own project. While I feel like I am currently proficient at modeling, I am still learning how to apply textures to models and properly unwrap them using the UV unwrap tool. For now I have enough skills to start the cornerstone of my project: Penn Hall.

I started with a reference image to lay out all of the details. I used a mirror tool that automatically applied edits to the x and y direction. This allowed me to only do a fraction of the work because Penn Hall is symmetrical on 2 axis. It still took roughly 2 days to get to the image below. It is not complete yet and I will continue to edit it throughout the week but for now it is good enough for me to begin interacting with it in Unity.

Going into week 3 I plan to go back to learning Unity so I can properly implement this model. I would like to add a door that takes the user to the inside once it opens and have rooms that the user can explore. The second Vive was set up at my work station so I am finally able to use it for my project. A lot of this week will be spent transferring what I have learned the first two weeks through Unity and Blender and implementing them into VR. Wish me luck.

The first week of DTSF has concluded and I now have a better understanding of how I want to do my project. I learned of a couple ways it could be done. The more photo realistic way would be to create a 3D model using a 360 camera and a drone. This would make the model more accurate but from a close up distance, it will appear more jagged depending on the photographs taken. The other way would be to create a model based off of maps and pictures. This will have less detail and more of an animated look to it, but it will not look photo realistic. I still have to decide which method is best for my project.

Day 1

The first day started at the library where we learned about projects made through word press. afterwards, we were free to begin our projects. I needed to be able to start with the basics of Unity 3D, so I started with watching basic tutorials to learn the interface of the platform. There is a lot of different features on the Unity interface and at first it was overwhelming. I was able to learn how to implement terrain and use assets to add external features to a project.

Day 2

The second day my goal was to learn how to script in Unity by using C#. I spent most of the day watching tutorials on learning the basics of the code. I learned that it is very similar to java, however there are a lot more basics of the language to memorize. Also, I learned there is a difference between C# for Unity and C# for generic coding. I learned basic commands such as being able to make objects move and printing statements in the console. I spend the end of the day learning how to add a 3D model from sketch fab into a Unity scene. I wanted to learn how to do this in case I end up having to make a photographic model for my project.

Day 3

Today mostly spent learning the basics of VR. We played around with the VR console and played with the vive and steam tutorials. Before this I have had very little experience with VR and this was an amazing experience for me. After messing around with the tutorials, I learned how to implement Unity with VR. Later I started an online tutorial with Unity that helped teach me the basics of the platform.

Days 4, 5

I spent the rest of the week with the tutorial I started the day before. The end result was a cube game developed through unity. The goal of the game is to control a cube on a narrow path and dodge all of the obstacles to get to the end. It was very long and I plan to review all I leaned so I can replicate it in projects of my own.

What I learned this week:

Interface implementation, terrain development, game objects, C#, VR implementation, model insertion, object movement, collision handling, restart features, in game physics, scene compilation, in-game text, score handling, exporting a project from Unity.

Next week I plan to take the skills I learned this week to start my project. I also intend to take an online course for Unity modeling so I am able to represent buildings as accurately as possible.