Every once in a while, someone stumbles into subsurface 3D crystal engraving by accident. That’s exactly what happened when Michael from the LaserEngraving911 YouTube channel accidentally created a subsurface effect with his UV laser. And eventually, this sparked a lot of questions from his audience:
-
How does subsurface laser engraving actually work?
-
What kind of laser do you need to make a 3D crystal?
-
Is there any difference between UV and green beam for deep 3D engraving?
-
How exactly does AI turn a flat photo into a 3D model that can be burned inside a crystal?
Michael invited me onto his channel to dig into those questions and get a proper grasp of 3D Laser engraving. This write up serves as a structured breakdown of that conversation with a detailed explanation of what subsurface engraving is, why green beam lasers are still the benchmark for 3D crystals, and how software like Cockpit3D and portable machines like JetMini fit into the picture.
If you’ve ever looked at a 3D crystal and wondered, “How on earth did they put that picture inside there and that too in 3D?”, this is for you.
What Is Subsurface 3D Crystal Engraving?
Let's start with the very basics. Most people are familiar with surface engraving, where you blast or mark the top layer of glass, wood, metal, or acrylic.
But subsurface engraving is different.
Here, instead of burning the surface, the laser focuses inside a transparent material, which is usually a high-quality optic crystal, and creates tiny micro-fractures at very precise points below the surface.
It is easier to understand if you think of it like:
-
Using a magnifying glass to focus sunlight on a leaf
-
But instead of sunlight, you’re using a laser
-
Instead of burning the leaf, you’re focusing the energy at a point inside the crystal
When all those focal points are stacked together, hundreds of thousands or even millions of them together form the appearance of a 3D model of the image floating in the crystal.
In the industry, we often refer to these points as voxels or a point cloud.
How Subsurface Engraving Differs from Surface Engraving
With a typical laser engraver:
-
You have just one main focal plane,which is the surface of the material.
-
You engrave in 2D that is either a line art graphic, logo, or photo
-
Motion is usually in X and Y, with Z used just to set the focus distance
Whereas with subsurface engraving:
-
You’re working in true 3D space
-
The focal point is controlled on the X, Y, and Z axes
-
The laser is directed inside the crystal to a specific point, fired for a fraction of a second, and then moved to the next point and so on.
Initially, when we started in this space, our machines were hitting around 100 points per second. That felt really fast at the time. Today, modern systems reach thousands of points per second. But speed is not only about the laser, it also depends on the galvos and how efficiently the path is planned through 3D space.
The end result is the same concept as an old dot matrix printer, where many small dots, placed correctly, add up to a complete image. In this case, the dots are micro-fractures suspended inside the crystal.
Why Green Beam (532 nm) Is Still the Laser of Choice for sub surface laser engraving
Michael’s accidental experiment with a UV laser raised a really good question, which is
Can you do 3D subsurface engraving with UV?
The short answer to this question is yes, to a degree. But there’s a really good reason why most professional 3D crystal systems still rely on green beam (532 nm) YAG lasers.
Here’s a breakdown of those reasons.
Wavelength and History
-
Subsurface engraving actually began with 1064 nm infrared lasers, especially for industrial applications like embedding tracking code inside bottles, helping manufacturers track where their product ended up.
-
Over time, the industry moved to a 532 nm green beam, which is essentially 1064 nm converted through a crystal to green.
The green beam 532 nm has now become the standard for 3D crystal engraving because of several reasons, and some of them are:
-
It delivers strong, consistent energy inside the crystal
-
It works well with larger fields of view
-
It produces bright, stable micro-fractures that are best suitable for deep 3D images.
UV vs Green Beam in Practice
UV lasers are fantastic tools for many materials and are a great option for at-surface or near-surface engravings. You can even push them into shallow subsurface work.
But for deep 3D engraving inside optic crystal, the green beam still has the edge in three big areas:
-
Depth and clarity
Green beam systems have the ability to maintain brightness and clarity deeper inside the block.
-
Field of view
Green beam lasers cover larger sizes of glass
-
Speed with 3D dense files
If you prefer high definition and dense texture-based 3D point clouds, then green beam setups are the perfect choice as they are typically faster and more stable.
That’s why, even though UV is improving and has its place, green beam is still the laser of choice when your main goal is high-quality 3D crystal engraving at production speed for business usage.
How AI and Cockpit3D Turn 2D Photos into 3D Point Clouds
Early on, the only way possible to create a true 3D model of a person was to:
-
Have them sit right in front of a multi-camera rig or 3D scanner
-
Capture depth and texture from multiple angles of the person's face.
-
Use that 3D model as the base for the point cloud.
It worked for sure, but was limited in several ways. People at tourist locations weren’t always “camera ready”, and the main problem was that many of our customers wanted to create crystals as gifts from existing photos.
The big shift came when we moved from:
-
3D scanners + in-person capture
to
-
2D photos + 3D modeling + AI
Manual 3D Modeling Era
Understanding the limitations of what we had, we built a process where:
-
Customers had the option to upload a regular photo
-
3D artists manually modelled the subject in 3D
-
The 3D model was textured carefully without losing a single detail of the image and then converted to a point cloud for engraving.
At one point, this process that we came up with involved over 200 artists working on photos from various locations around the world. It worked well, but was time-consuming. It took around an hour per person for the detailed work to be done.
AI-Assisted Era
Over the last few years, we have been able to make significant improvements. What paved the way to these improvements is a combination of:
-
Thousands of meticulously created 3D models
-
Permission-based photo datasets
-
Internal AI pipelines
And this has allowed us to train models that can:
-
Take any 2D photo
-
Automatically generate a 3D base model that’s about 80% complete
-
Our Human artists step in and handle the last 20% of the fine-tuning, like fixing the hair, glasses, noses, and other small details that the AI can’t get perfectly on its own.
The key lesson we learned along the way:
The AI image output that looks amazing on a computer screen doesn’t always engrave the same way in the crystal.
The real test is when you remove the texture and burn the point cloud in crystal. For example, if noses look flat, hands are melted together, or the glasses and eye sockets are wrong, customers notice immediately.
So we decided to use AI as a powerful assistant, but still keep human oversight to ensure perfection in terms of quality. This is essential for us to maintain clients like Disney that need fast turn around of high quality