Similar to other software developers, Reality Capture keeps track of the system requirements, which you can use to make sure your system's hardware is compatible with their program. Compared to many other companies, Reality Capture appears to maintain this list quite current with every product update and even provides some advice regarding when advanced hardware can be useful.

The fastest CPU for reality capture should be Intel's newest flagship, the Core i9 14900K 24 Core, while the i7 14700K would still perform excellently in this capacity; both are quicker than previous series of Intel processors. If you'd rather use AMD CPUs, their Ryzen 7000 Series chips also function well.

This higher-core processor offers the best performance in reality capture when working on large projects. The CPU is used differently in each step of the reality capture process. Reality Capture uses multiple CPU cores rather well overall, while the impact of more cores varies depending on the processing step. However, clock speed is also essential; consequently, the best processor options strike a compromise between both of them.

When compared to other photogrammetry applications, Reality Capture's memory requirements are actually rather low. "Out-of-core" algorithms, which do not require all of the data being worked on to be stored in the main system memory at the same time, are what this program was made to use. As a result, even very large projects can be completed with modest RAM requirements (the developers even state that image sets containing thousands of photos can be completed with 16 GB). At Arvuti.in, however, we usually advise 32GB as a minimum because Windows and other background-running apps also require memory. In our testing, we also discovered that to be a good place to start.

Since Reality Capture utilizes CUDA for some of its critical processing, it needs an NVIDIA graphics card to function properly. In theory, without it, you can still run the program and carry out some simple operations like registering images, but you won't be able to create a mesh or 3D model. Our research revealed that the performance difference between modern mid-range and high-end video cards isn't that great, but spending more does result in a slight boost in performance.

You will want an NVIDIA graphics card that supports CUDA and has a somewhat larger VRAM capacity if you wish to utilize those capabilities. Here, the GeForce RTX 40 Series GPUs are a good choice; the GeForce RTX 4070 SUPER 12GB and RTX 4080 16GB versions are compatible with our Reality Capture workstation.

Primary Drive: OS/Software (SSD/NVMe): Includes your operating system and the base Reality Capture installation. An SSD is highly recommended as it will greatly improve how fast the OS and programs startup, but you can also upgrade to a faster NVMe drive for a small performance benefit. The recommended storage for installation of software is not more than 10 GB.

Since the primary drive will primarily be used to store Windows and software, including all software (except large Game), it doesn't need to be that large for the majority of users; a 500GB-1TB SSD will be enough for Drive C.

Secondary Drive: Project Files (Platter/SSD/NVMe) Having a second storage (either SSD or HHD) for storing assets and effects is a wonderful alternative if you have additional money to spend or keep a lot of them around for use on different projects. Of course, an internal or external hard drive is a more cost-effective option than a high-capacity SSD if you need more mass storage and preservation.