Learn How to Create a Realistic 3D Portrait of Caravaggio in Maya & ZBrush

Introduction

Hello, I'm Vahid, and I'm 39 years old. I started learning art at the age of 30. My daily job was completely unrelated to art, and it was boring company stuff. I am self-taught and have learned through online courses, YouTube, books, and other resources. I have loved this kind of stuff since childhood, but sadly, I never started learning it at an early age.

Inspiration & References

Caravaggio is an inspiration to every artist, no matter if the person is a cinematographer, photographer, or painter. Also, I wanted to make a portrait of him based on the sketch by a famous artist and printmaker Ottavio Leoni. There are very few references from him. Of course, he painted himself a couple of times in his paintings, but they were not helpful because either he was so young or he was one of the people in the busy compositions. So, I went with this one since everyone saw it.

Head & Face

I started with some quick sketches in ZBrush. From the start, I took the sketch to Maya and matched it with the reference using the image plane in Maya, put a light on it, and started rendering. I always have at least three systems of different lights for final rendering and one free lighting that I use to rotate around my model. Free light is very useful. It's even mentioned in the first book about lighting in cinema by Master John Alton, Painting with Light. It is a useful book in which he introduces how you make 3D images and compositions.

Also, I have this personal super good trick to match with reference in Maya, which you can see from the images and the video below. First, you need to load your image into the image plane. Second, click on the fir to the resolution gate. Third, put everything on the layer so you can turn everything on and off. Next, you can easily see the rendered image and the main reference. I typically bring them into Photoshop and I make a take-screen video to see the differences.

A very important thing about 3D portraits is that you need to know what you shouldn't do and what you must do because if you want to match it 100 percent, it's not going to work. Because of that, you must ignore the parts that don't work in 3D. For example, I didn't match 100 percent the right brow, the right eye, or the lips because it simply looks weird in 3D. A statue is very different than a painting. Your job as a sculptor is to take the essence of that character as much as possible, which lies in the basics of the likeness in both painting and sculpture, which are making the proportions right as well as the main shapes of the face and also forms.

For the hair, I only use XGen and make everything related to hair in it. It's vital that we know hair, like the face, has planes, and we need to provide those planes to make it look alive. Caravaggio's hair is very clumpy. I did it two times, and the second time, I got it right, like the reference.

Please notice that composition, shapes, forms, lighting, and exposure are very vast topics. I'm speaking super briefly about them, and I'm not mentioning their details since this is a short breakdown.

Changing the Composition

I changed the composition to make it more alive and dynamic. I used harmonic armature to make the composition, and by intention, I made the hand and collar of the shirt and face brighter to make the connection between brighter values to control what the viewer needs to see, which first sees the face and then goes to the hand. Using a white collar is a very old trick to make the head pop out more. You can see an example of that in many works of the Sargent or Vandyke and all the masters.

I also always use False Color to check my exposure. As you can see in the image below, the brightest parts are the collar, the face, and the hand, where you can see white and yellowish colors from false colors in the color look-up in Photoshop. You can download False Color for free and load it in your Photoshop color look-up Adjustment Layer to check your values.

Also, making a good composition that works comes from many aspects: parallel obliques, tangent to tangent, arabesque, making contrast using values and saturation, but firstly, it comes from outline and inner shapes. For example, see the master painting by Sir Gerald Festus Kelly. It represents interesting shapes of a face and a body. He made a classic triangle general shape for composition and used Curves to separate body shapes from the cloth. Also, masters nave made jiggle or wobbliness around the edges of cloth or garment. Again composition topic is very detailed and I just mentioned some small stuff about it.
Note: Harmonic armature is a great old tool for putting your subjects on the right parallel and oblique lines, which come from harmonic armature references.

The outer drapery is a simple simulation in Marvelous Designer, and the inner one is basic shape modeling from ZBrush.

Retopology & Unwrapping

I did a quick edit of topology to clean up loops, mainly using ZBrush's ZModeler brush to make my loops a little better. Then I used a simple ZBrush UV Master Plugin to unwrap, and then I took it to Maya and made six UDIMs. I typically use five UDIMs, but on this, I used six and I took it to Mari

Texturing

It's very simple. I use a mix of ZBrush and XYZ 16K textures. For XYZ, I only used the old 16K maps that I already have. I separated them into small parts and I import them into mari and I only use projection in mari. For the rest of the texturing workflow, I use ZBrush. For Albedo, I sometimes start with 3D Scan store Albedo, but sometimes I fully hand paint it using poly paint.

So, I project XYZ maps on the model, but for all the moles, nose top spots, and scratches, I only use ZBrush to have 100 percent control. For the lips, I also use manual sculpting from ZBrush mixed with XYZ textures and export 8K maps from ZBrush.

For 16K XYZ maps, since they are calibrated and have almost the same intensity, sometimes I use two maps. For example, in the lips area, for the lower lip, I used another 16K map to have a different effect.

Note: ZBrush is the best for making hand-made details like moles, wrinkles, spots on the nose, and so on, with 100 percent control, but because they are 8K limited and also ZBrush is based on pixels, when it comes to super micro details, it gives you little pixelation if you zoom it in ZBrush or even by using HD geometry in ZBrush. But since 16K maps are 32-bit raw images, they do not have this problem. So, in order to make micro details, in my personal opinion, it's better to use 16K or higher resolution images to have a sharper effect.

For the eyes, I sculpted some iris three years ago, and I have used it since then. For Albedo for cornea and iris, I hand-paint them and export them to Maya to render.

Lighting, Rendering & Post-Production

I usually make a lighting setup contrasting. Please look at the work by the great photographer and painter Emilio Sommariva. In that example, you see light in great contrast, which turns into shadow. I always look for contrasty, interesting lighting, and Arnold is very good for that matter. I used a key light that had a free HDRI from the Internet on top of it, and I am 70 to 80 cm away based on the size of the light that I have. It makes good contrast, and I have a fill-to-fill shadow and a light only dedicated to the hand, as well as one for turning on the backdrop and a very subtle light to add some bounce to the back of the hair and jaw. For a backdrop, I always paint it in the Substance 3D Painter.

For lighting, it's essential for everyone to study the basics of light, like exposure , inverse square law, low key, high key, and so on. For that, you can study the work of photographers like Irving Penn, Peter Lindbergh, Vincent Peters, Robert Mapplethorpe, Albert Watson and all the great master painters and also cinematographers like Vittorio Storaro or Sven Nykvist and others.

For the post-production, I used a paint action called the Real Paint FX Photoshop Add-On extension from Envato. I used 30 percent of that and only some layers, not all of it because not all of it is useful. Then, I did manual brushwork on the face, canvas, and cloth in a subtle way.

For rendering realistic characters, the most important part, and 80 percent of it, is anatomy, and then goes to good Albedo and textures, which needs a lot of study as well. I highly suggest studying traditional painting for skin and lighting. The realistic portrait is not about skin pores. Skin pores don't do anything special, it's all about the connection between your shapes and forms and the color of your Albedo. A very good example of that is the Repin paintings. They do not have any skin pores, but at the same time, they look very realistic. No classical painting has skin pores, so reality comes from other things such as anatomy forms, shapes, skin color, and lighting. Details such as skin pores are only compliments to realism if the artist wants to do it.

I think the most important thing for each artist is to create style because if you do realism, just being real is enough, and on top of that, we need to add style.

Conclusion

When I start my project, I work for a couple of days, then let go of it and proceed with my other projects. I always have at least four to eight projects going, but I never finish them so quickly to learn something and also keep my motivation going. For this, I think it was six to seven weeks. I normally finish work in three weeks, but if I do a skin close-up, it adds at least two more weeks to make it work. It helped me to learn more and more about anatomy, gestures, colors, and poses. Every project is always a new challenge.

Once you want to convert a sketch or painting into 3D, it is always challenging because you have to rely on your information about everything, as painting is 2D, and making it work in 3D is a totally different story. As I mentioned in first part, you need to know what you don't have to or what you have to.

To every artist, patience, practice, and learning are key elements. I always tell my friends or students to do multiple projects to keep their motivation going and also to study and practice a lot. I have seen many people fail in the 3D industry because they lack discipline and hard work. Art is really hard. They can read books like War of Art by Steven Pressfield or books by Stoics to have discipline because the artist needs to focus on very few things in a day and must not be distracted.

One of the saddest parts of the 3D industry is that people consider themselves only 3D modelers and do not study the roots of art like painting, sculpture, or photography. Maya, ZBrush, and other software are only tools for expressing artistic ideas. 

I thank the 80 Level team for doing great work.

Vahid Ahmadi, 3D Character Sculptor

Interview conducted by Amber Rutherford