In 3D rendering, supersampling is a well-known technique used to reduce aliasing by rendering the image at a higher resolution and then downscaling it. However, the quality of the final image largely depends on the filtering algorithm applied during that reduction.
In previous versions of Patchwork 3D, the image downscaling process used a Tent (or triangle) filter, a simple, fast, and widely adopted approach. While this method worked well in many cases, our R&D team noticed that it sometimes left high-resolution renders lacking a bit of crispness, especially when fine details were involved.
This observation led us to explore more advanced filtering options, aiming for even sharper and more accurate results.
To address this, we tested and compared several filtering algorithms: Box, B-spline, Mitchell-Netravali, and Catmull-Rom, using various test cases, including product renders and standardized sharpness charts.
✨ What’s new in Patchwork 3D 2025:
✅ New filtering options are available when using supersampling for snapshots
✅ Improved visual clarity with more precise edge rendering
✅ The Catmull-Rom filter has been selected for its excellent balance of sharpness and smoothness
Here is a comparison of supersampling filters:

Legacy Supersampling Filter / To magnify the picture: right-click on it and select in the context menu: “Open image in a new tab”.
Catmull-Rom Supersampling Filter / To magnify the picture: right-click on it and select in the context menu: “Open image in a new tab”.
The Supersampling differential between the Legacy and the Catmull-Rom filters. / To magnify the picture: right-click on it and select in the context menu: “Open image in a new tab”.
This update gives teams more control over the visual quality of their renders, directly within the software, with no need for external post-processing.
Let’s say a marketing team needs high-quality visuals for an online product configurator. The images must be sharp, realistic, and consistent across all outputs.
With the new filtering options in Patchwork 3D 2025:
- The team renders at 2x resolution with supersampling
- Applies the Catmull-Rom filter during downscaling
- Obtains detailed, professional-grade images ready for immediate use
👇🏻 Share your feedback, just below. 👇🏻