So you’ve sculpted your brand new character in blender and you want to start animating or importing the character into a game engine like Unity.  However, you look down at your poly count and see 1,913,482!!  If you’ve been around 3d modelling for some time, you'll probably have some idea of where this is going.  If you're on the novice side of things, don’t worry.  We’ll be taking a dive into the topic of retopology in blender where our goal is to take that enormous amount of polygons and make it much, much smaller.

Retopology (retopo for short) is arguably the least favourite thing for blender artists to do because it can feel tedious, but it is vitally important because of the limited resources of graphical processes.  Graphical processes come in two major flavours of discrete (separate from the CPU) and integrated (part of the CPU assembly).  However, even the biggest and best graphics card will have a hard time keeping track of all those polygons if there are many such assets on the screen at the same time.  If your assets will be used on mobile devices, it’s even more vital to do the retopo process because of the very small amount of graphical processing memory and power.

Now, you might be wondering, “won’t that process take my ultra high-res and beautiful character and make her look like a potato?”  Retopology is part of the asset optimization workflow.  As will be described in a moment, retopo creates the 3d layout of polygons, but the textures created on the unoptimized version will be mapped to that object.  So, at the end of the total workflow, your character will retain all of their charm but be much more capable of smooth motions on the screen.

Common Methods of Retopology

At this point we discussed the main reason why retopology is important and what the goal is, but we haven’t described what it is exactly.  In essence, we want to create a lower polygon count shell around the currently sculpted model.  We have to remember that the asset will take motion which means that when we create that shell some areas will require more polygons than others.  An example would be the eyes where a character might have a blinking animation to be more life-like.  This could also apply to ribbon streamers of a dress, sword, or axe that will move with the wind so we don’t limit our thinking to characters only.

Every blender artist generally settles on their own unique approach that either stems from or a combination of three main methods of retopology.

Sculpting and Manual Retopo

This is my preferred method and is generally used for characters and textured objects.  As the heading suggests, its implementation is done by first sculpting a 3d model and then creating the shell around it to which you would map textures to.  The shell will be shrink-wrapped to the 3d model.  This process is easier understood by watching so I’ll include a video here.  Additionally, if you become interested in wanting to learn the whole workflow in depth, I’d recommend my premium course called Character Creator 2.0.

Sculpting and Automated Retopo

There are going to be cases where automated retopo tools come in handy rather than doing everything manually.  Precision is the tradeoff, but many of the available blender plugins do exceptionally well for environmental objects.  Some 3d modellers will run an automation tool first and then clean up specific areas with a manual process.  It’s a tradeoff between saving time and being precise. Some automated tools are Quadremesher -a paid for addon that is very effective or Instant Mesh -a free alternative. 

Additionally, there are some semi-automated retopo tools like RetopoFlow and Quadwrap.  Both of these tools require the artist to mark out areas that need retopology with customizable options.  Many tools exist out there, but it requires some research and knowledgeable experience to properly weigh the advantages and disadvantages of their usage.  The best advice is to experiment until you find what works for you generally and in specific situations.

Building With Minimal Vertices

Interestingly enough, there are some workflows that actually model without the need for a typical retopo process.  For instance, in many cases of hard-surface modelling vertices are kept minimal throughout the modelling process.  That’s not always the case as hard-surface models sometimes require sections to be texturized.  Part of the process is thinking ahead and knowing what parts of an object need to be texturized or sculpted to separate those parts out of the main object.

Another interesting thing is that you will find some other artists out there that prefer building characters polygon by polygon rather than traditional sculpting methods.  You’ll see this in cases of high amounts of character symmetry such as with low-poly characters and anime characters.  In these instances, textures are rarely used and usually mapped as a flat image.  Think of that gnarly scar on your favourite anime character.

What Comes After Retopology?

Hopefully, you now feel confident in understanding what retopology is and why it’s important.  If you watched the video showing the manual process, you should have a good understanding of how it can be done.  Of course, it takes practice and learning from mistakes.  Sometimes it can seem tedious to do retopo, but if you’re going to be doing 3d modelling professionally it’s necessary to master.

The next part of the common workflow is, unwrapping the base 3d model. After you unwrap and then map the textures to your retopo shell, you will feel the reward of all the hard work. The potato looks amazing again! Now, it will also be able to come to life and move smoothly.

For a free playlist of videos focusing on retopology, you can view this retopology playlist.