Even after the camera improvements in Beta 11.4.7, HW4 cameras are still a leap ahead in clarity
DirtyTesla
Seeing is believing and more and more videos and images are popping up showing the difference between Tesla's Hardware 3 (HW3) and the latest Hardware 4 (HW4) cameras. The visuals highlight a radical transformation in video quality that paints an exciting future picture.
Improvements to HW3 Cameras
Tesla vehicles with the latest FSD Beta, version 11.4.7 include a notable visual improvement to the vehicle's camera feeds. With this update, Tesla introduced white-balance and color-balance corrections to the vehicle's camera feeds, making them look more realistic and accurate of the real world. But the biggest difference is actually the improvement in camera clarity. The visuals are noticeably sharper.
Our previous article and the included video demostrates the visual differences on HW3 cameras before and after applying FSD Beta 11.4.7. While these improvements are in a branch that is only available to FSD Beta testers (2023.7.30), it's likely that these video improvements will be applied to all vehicles in a future update.
Difference Between the Updated HW3 Camera Feeds to HW4 Camera Feeds
One of the best videos to date that compares HW4 to the new and improved HW3 cameras with Beta 11.4.7 was filmed by DirtyTesla, who did a nice job comparing the various cameras at the same location and in the same lighting. His findings and video are below:
Here is what stands out:
A Leap in Resolution: The images demonstrated the stark difference between the 1.2-megapixel HW3 cameras and the newly introduced 5-megapixel HW4 cameras. This substantial upgrade in resolution has translated to a noticeable increase in clarity, allowing for the easy reading of license plates and road signs that were previously indecipherable with HW3.
Even after the camera improvements in Beta 11.4.7, HW4 cameras are still a leap ahead in clarity
DirtyTesla
The Fish-eye Effect: Another captivating revelation from the images was the pronounced fish-eye effect in the rear camera of HW4. The images depicted a substantially wider field of view compared to the previous version, enhancing peripheral vision for a safer driving experience.
Night-time Brilliance and Vibrant Colors: The comparison was confined to resolution and field of view and highlighted the improved night-time footage quality, exposure control, and more accurate color representation. This promises a more realistic and vibrant visual experience for Tesla owners.
What Does It Mean for Full Self-Driving?
While the images are impressive, the next set is how the camera technology will translate into improved Full Self-Driving capabilities. Elon Musk has hinted that HW4-equipped cars could be 3 to 5 times more adept at autonomous driving. We know that HW4 has more ports for additional cameras as well.
Retrofitting Possibility: A Disappointment
Now to remind you about the disappointing news for existing Tesla owners. These amazing images, courtesy of the new cameras might lead to inquiries about retrofitting possibilities. Unfortunately, the idea of retrofitting has been shut down for some time, in fact, Tesla has been trying to entice current Tesla owners into new Teslas by offering a limited-time transfer of FSD.
The comparison images between HW3 and HW4 cameras are more than just a technical showcase. It's visual evidence of how fast technology is changing and how Tesla continuously updates the hardware in their vehicles.
In our continued series exploring Tesla’s patents, we’re taking a look at how Tesla automates data labeling for FSD. This is Tesla patent WO2024073033A1, which outlines a system that could revolutionize how Tesla trains FSD.
We’ll be approaching this article the same way as others in the past, by breaking it down into easily digestible portions.
Training a sophisticated AI model like FSD requires a tremendous amount of data. But all of that data needs to be labeled - and traditionally, this process has been done manually. Human reviewers have to go in and categorize and tag hundreds of thousands of data points across millions of hours of video.
This isn’t just laborious and rote work, it's time consuming, expensive, and prone to human error. The perfect job to hand off to AI.
Tesla’s Automated Solution
Tesla’s patent introduces a model-agnostic system for automated data labeling. Just like their previous patent on the Universal Translator, this will function for any AI model - but FSD is really what it is for.
The system works by leveraging the vast amounts of data collected by Tesla’s fleet to create a 3D model of the environment, which is then automatically used to label new data.
Three Step Process
This process has three steps, so we’ll look at each individually.
High-Precision Mapping
The system starts by creating a highly accurate 3D map of the environment. This involves fusing data from multiple Tesla vehicles equipped with cameras, radar, and other sensors. The map includes detailed information about roads, lane markings, buildings, trees, and other static objects.
It's like creating a digital twin of the real world, and this is exactly the simulation data that Tesla uses to rapidly test FSD. The system continuously improves its accuracy as it processes more data and also generates better synthetic data to augment the training dataset.
Multi-Trip Reconstruction
To refine the 3D model and capture dynamic elements of the environment, the system analyzes data from multiple trips through the same area. This allows it to identify moving objects, track their trajectories, and understand how they interact with the static environment. This way, you have a dynamic, living 3D world that also captures the ebb and flow of traffic and pedestrians.
Automated Labelling
Once the 3D model is sufficiently detailed, it becomes the key to automated labeling. When a Tesla vehicle encounters a new scene, the system compares the real-time sensor data with the existing 3D model. This allows it to automatically identify and label objects, lane markings, and other relevant features in the new data.
Benefits
There are three simple benefits to this system, which is what makes it so valuable.
It is far more efficient. Automated data labeling drastically reduces the time and resources required to prepare training data for AI models. This accelerates development cycles and allows Tesla to train its AI on much larger datasets.
It is also scalable. This system can handle massive datasets derived from millions of miles of driving data collected by Tesla's fleet. As the fleet grows and collects more data, the 3D models become even more detailed and accurate, further improving the automated labeling process.
Finally, it is accurate. By eliminating human error and bias, automated labeling improves the accuracy and consistency of the labeled data. This leads to more robust and reliable AI models. Of course, human review is still involved, but that’s only to catch and flag errors.
Applications
While this technology has significant implications for FSD, Tesla can use this automated labeling system to train AI models for various tasks.
Object detection and classification: Accurately identifying and categorizing objects in the environment, such as vehicles, pedestrians, traffic signs, and obstacles.
Kinematic analysis: Understanding the motion and behavior of objects, predicting their trajectories, and anticipating potential hazards.
Shape analysis: Recognizing the shapes and structures of objects, even when partially obscured or viewed from different angles.
Occupancy and surface detection: Creating detailed maps of the environment, identifying occupied and free space, and understanding the properties of different surfaces (e.g., road, sidewalk, grass).
These different applications are all used by Tesla - which uses different AI subnets to analyze all these different things before feeding them into the greater model that is FSD, which means things like pedestrians, lane markings, and traffic controls are all labeled on-vehicle.
In a Nutshell
Tesla's automated data labeling system is a game-changer for AI development. By leveraging the power of its fleet and 3D mapping technology, Tesla has created a self-learning system that continuously improves its ability to understand and navigate the world.
Imagine a world where self-driving cars can label and understand the world around them without human help. This patent describes a system that could make that possible. It uses data collected from many Tesla vehicles to create a 3D model of the environment, which is like a virtual copy of the real world.
This 3D model is then used to label new images and sensor data, eliminating most needs for human intervention. The system can recognize objects, lane markings, and other important features, making it easier to train AI models.
Back in 2021, while Giga Berlin was still undergoing construction, Elon Musk said that he wanted to fill the factory with graffiti artwork. Just months later, Tesla posted a submission link to find local artists for the project.
It remained relatively quiet for about two years until Musk resurfaced with a post congratulating the team on their progress—and revealing that the factory’s concrete would be entirely covered in art. By 2023, that vision was already taking shape. Tesla began by collaborating with local artists, who created much of the artwork seen in the 2023 image above.
The Giga Berlin West Side in 2023
Not a Tesla App
Graffiti at Scale
As expected from Tesla, they didn’t just hire a group of artists to paint and scale the walls. True to their ethos of autonomy, robotics, and innovation, they sought a more futuristic approach. The local crews couldn’t work fast enough or cover enough ground, so Tesla did what it does best—push the boundaries of technology.
Covering an entire factory in art is a massive undertaking, especially when that factory spans 740 acres (1.2 sq mi / 3 km²). With such an immense canvas, Tesla needed a high-tech solution.
More of the awesome digital artwork
@tobilindh on X
Enter a graffiti start-up that had developed a robotic muralist. Tesla partnered with the company, sourcing digital artwork from independent artists while also commissioning pieces from its in-house creative team. Armed with this collection, the robot meticulously printed the artwork directly onto the factory’s concrete, turning Gigafactory Berlin-Brandenburg into a futuristic masterpiece.
The Robot
Not a Tesla App
This ingenious little robot is equipped with a precision printhead and a sophisticated lifting mechanism. It moves using two kevlar cables that allow it to glide up, down, left, and right while a pair of propellers generates downforce to keep it steady against the wall.
The printhead itself is capable of painting approximately 10 million tiny dots per wall, adding up to a staggering 300 million dots just for the west-facing side of Giga Berlin. Each mural features five distinct colors, and the robot carries 12 cans of paint, ensuring it can keep working for extended periods without interruption.
Check out the video below to see the robot action, along with mesmerizing time-lapse footage of the printing process. It’s an exciting glimpse into how Tesla is blending technology and creativity at Giga Berlin—and we can’t wait to see what’s next.
![Tesla’s Giga Berlin Artwork: Where Creativity Meets Autonomy [VIDEO]](https://www.notateslaapp.com/img/containers/article_images/2025/giga_berlin_art_tobias.jpg/074b092609d4477c5cafb71e90dbaf57/giga_berlin_art_tobias.jpg)
![Video Reveals the Cybertruck's Unique User Interface [Video]](https://www.notateslaapp.com/images/news/2023/cybertruck-ui-1_300w.jpg)
![Tesla Shows Off Robotaxi Cleaning Robot [VIDEO]](https://www.notateslaapp.com/img/containers/article_images/2025/tesla_robotic_cleaner.png/08bbe7d309afe2f4bfee710a5df8cf25/tesla_robotic_cleaner.jpg)
