Feature

Upsampling

8192×4096 native, up to 16384×8192+ with Real-ESRGAN upscaling

Included with every license

Upsampling uses Real-ESRGAN, an AI upscaling model, to enhance your panoramas while addressing boundaries through tile-aware geometry. Transform 4K images into rich 8K or 16K+ outputs with enhanced texture and visual depth.

Resolution Scaling

4K

4096×2048

8 Megapixels

8K

8192×4096

33 Megapixels

16K+

16384×8192

134 Megapixels

Key Features

Neural Upscaling

Real-ESRGAN uses deep learning to intelligently upscale images, adding convincing detail and texture rather than simple interpolation blur.

Detail Enhancement

Enhance textures, sharpen edges, and recover fine details lost during compression or lower-resolution rendering.

How It Works with Hexagonal Tiling

Wrap-Aware Upscaling

Unlike standard upscalers that process the entire image at once, our system intelligently handles the equirectangular projection, processing tile boundaries across the 360-degree wrap.

  • 1

    Tile-Aware Processing

    Each hexagonal tile is upscaled with overlap awareness

  • 2

    Edge Blending

    Blended across tile boundaries

  • 3

    Pole Handling

    Special care for zenith and nadir regions

  • 4

    Wrapped Output

    Wrap-aware tile processing

Use cases

Skyboxes

For your virtual world

VR Experiences

High resolution for virtual reality immersion

Concept Art

Develop rich high resolution worlds and consistent worlds for your IP project development

Vibe Coding

Use Hextile to agentically generate worlds for your vibe coded video game

Pipeline Integration

Upsampling fits into your workflow as the final step after AI diffusion processing, or apply it directly to existing panoramas.

Apply after SDXL or SD3.5 processing
Upscale existing panorama libraries
Export at your target resolution
4K Input
Real-ESRGAN
16K Output

Technical Specifications

Model Real-ESRGAN (NCNN implementation)
Scale Factors 2× and 4× upscaling
Input Formats PNG, JPEG, TIFF, EXR
Output Formats PNG (16-bit), JPEG, TIFF
Max Output Resolution 8192×4096 native, up to 16384×8192+ with Real-ESRGAN upscaling
GPU Acceleration NVIDIA CUDA / Vulkan
VRAM Usage ~2-4 GB (tile-based processing)
Availability All licenses

Work on your own GPU

One-time licence from $360. No subscription, no per-render cloud charge.

30-day money-back guarantee on every purchase.

We use cookies for analytics to understand how visitors use our site. Privacy Policy