Skip to content

BlogProduct

Sharper real-time 3D streaming, same bandwidth

Miris now streams 3D assets with more detail at the same bandwidth. Same upload flow, same SDK, better fidelity. Try the Miris Beta.

Laurie KoenigAugust 27, 2026

Attribution: Lucas Cuenca. Software: Maxon ZBrush

Why it matters

Miris assets now stream at higher fidelity. Here’s what evolved in our product and what it means for you. Under the hood, what the Miris platform streams isn’t your source file - it’s a proprietary radiance field representation built from it. The generation of that field is the most critical step to improve the quality and fidelity of the streaming experience.

Every partner building on Miris asked for the same thing: higher fidelity - but for different reasons. Playbook’s users want to search their 3D asset catalogs quickly, without forcing a download or requiring special software to inspect the authored representation. CGAxis sells the assets it streams, and the preview is what a buyer judges before purchase. Evercast needed that same accuracy to hold up in live, multi-party sessions. After hearing this feedback, we decided the most important area we could focus on was improving the fidelity of fine detail.

Attribution: AWS Airship Asset Copyright 2024 Amazon Web Services. All rights reserved. https://dpel.aswf.io/airship/aws-airship-license/

What we did and how

Part of our streamable radiance field generation process involves volumetric capture of surface data using arrays of "light sensors" that are placed strategically in and around a 3D asset. By improving those placement algorithms, we’ve improved the accuracy of the system by over 300%. In parallel, data generation and processing bandwidth is up to 5x higher, in part to support the incredibly large assets our customers bring us. More bandwidth means higher-fidelity visual data reaches the client, raising the upper limit on the amount of detail your audience can visualize. That additional data called for better compression techniques, so the end-user experience stays exactly as responsive as it is today.

Capture points surrounding a 3D asset beside a color-coded normal-map render of the airship

Where you'll see it

Complex geometry with fine features, high-frequency surface textures, and foliage were some of the hardest assets to reproduce, and they're where the difference is most immediately visible. Every asset is reconstructed using the same method, so raising capture resolution and improving our algorithms increased fidelity across the board.

Close-up of the robot arm's wrist flange before and after the update, with surface detail visibly sharper after
Before and after fidelity comparison of a streamed industrial robot arm, full view
Close-up of a suitcase corner bumper before and after, with fabric weave and shell texture resolving after the update
Suitcase top and handle before and after, with crisper ridges and hardware after the update
Close-up of grass blades before and after, with individual strands staying distinct after the update
Before and after comparison of a potted ornamental grass, full view

Try it in the Miris Beta

Miris still eliminates the need for expensive GPU based pixel streaming at runtime. Nothing changes on your side: it’s the same upload flow, the same SDK, and now higher-fidelity streams at the same bandwidth. Just upload an asset and compare it against what you're shipping now.

Hero Thumbnail Attribution: Lucas Cuenca. Software: Maxon ZBrush