My thoughts on Unity Certified 3D Artist exam

I took the certification “Unity Certified 3D Artist” beta exam recently on 2018/April/28. I had 120 minutes to answer 92 questions.

My background is programming. I had no special art training but could make basic 3d models in Blender, import them into Unity and create a basic scene with various light source types. I had made several 2D game prototypes at the time, but nothing complicated in the artistic side of the work. I hold the certification “Unity Certified Developer” (Certified on April of 2016). I had 3 years of experience with Unity and had been a VR Game Developer as my daytime job for two years.

The exam was difficult. It was perfectly tailored for 3D artists and the questions were heavily scenario-based. I had to carefully measure various factors (such as CPU/GPU usage vs. memory usage, quality vs. frame rate) and make a reasonable tradeoff based on the given scenario. For example, there was a question about the correct usage of realtime and baked lighting for a downtown area (at night) with walking pedestrians and light poles. The target platform was low-end smartphones with limited memory. Having the knowledge, you could determine the right answer.

Some questions had photos with them and you had to consider the image to be able to choose the right answer. As an example, there was an image showing a realistic sunrise in which the upper half was reddish, and the bottom half was bluish. You had to choose the answer that could reproduce such a setting. As a programmer, I always wanted to learn these techniques that artists use to achieve these beautiful sceneries. I remember there was a question about how to efficiently reproduce lights going through a colorful glass of medieval church. Again, the answer was not obvious. The given choices were usually multiple lines and similar to each other. Some wrong answers were right if the target platform was different. So, you had to pay attention to the details in the question.

I have one major complain about some questions. They described a complex scene without showing any illustrations. You had to imagine a complicated geometrical scenery by reading 4 lines of texts. Then you were asked to do something based on this geometry. If you could not imagine that correctly, you were screwed. Sometimes you could imagine different shapes based on the same description. My advice to Unity exam authors is to provide a picture instead of describing it.

In some scenarios, you were encountering an issue (such as wrong lighting on a mesh or flickering areas when the camera is moving) and you had to figure out what could be wrong and how to resolve it. To be able to answer them, you had to know various import options (like “Optimize Mesh”, “Keep Quads”, “Weld Vertices”, “Use File Scales” and so on). In my opinion, without having a practical experience in the past, you could not figure out what is going wrong.

Couple questions were about scaling issues. For example, there was a scenario similar to this (some details removed or changed to not to breach the NDA):

Your team has made a VR game in which the player experiences an indoor area from a dog perspective. The furniture is scaled 2 times bigger than their actual size to give the player the impression of looking through dog eyes. Now they have decided to add an option for players to play the game as human (so the player can choose to play as a human or a dog) and they want the size of the furniture to look right. As an artist, how can you achieve that?

You were given different choices that suggest changing the scale of an object by half or changing VRSettings.renderScale, etc. …. Another question asked me to calculate the correct pixel per unit value if I wanted to have a sprite imported with a specific width.

These scenario-based questions spanned from architectural details of furniture (LODs, detailed masks…) to filmmaking workflow (Cinemachine, various camera movements and post-processing effects vs using shaders in offline or real-time scenarios) to real-time visualizations of shiny brand-new cars in VR or augmented reality. I remember several questions on how to use curves (Animation curves, easing-in/ easing-out transitions, Blend Curves, Speed Multiplier, etc.) in mecanim to demonstrate a powerful fist thrust in a cut-scenes of a fighting game. Key-framing-, Dope Sheet-related questions were all there.

There were few questions (about 3 or 4) about “Collab” (a version control system made by Unity) and even fewer questions about audio (which I don’t remember). The “Collab” questions were about the current status of your local files (whether they are up to date or behind the latest version or have a conflict with the main repository).

Couple questions were about physics (things like OnCollisionEnter vs. OnTriggerEnter). One question asked the best approach for opening a door when the user interacts with it. It asked about where to put a script that used the Collider component (on the door frame, door or doorknob). You were given a screenshot of their gameobject hierarchy. This was the only programming-related question I saw on this exam.

I will get my exam result in 4 months. (Update: My score was 395, way below the passing score of 500. I failed but I will retry the exam after Unity provides an Official Course for it.)

In conclusion, I highly recommend studying for this certification (even if you are a programmer) because you’ll gain lots of valuable knowledge. I wish there was a course that taught all the exam materials (Update: Unity has started to offer a Coursera Specialization for this exam called “Unity Certified 3D Artist Specialization“.)