A circular texture queue can reduce stutter in multithreaded rendering, but increases memory and delay. See when to measure and when to avoid.
Direct answer
ARCore can hold multiple camera frames in a circular queue of textures. In multithreaded pipelines, this can improve frame rate consistency and reduce stutter, but each additional 1080p frame consumes about 6 MB according to the documentation, and buffering introduces delay between capture and display. Gain is not guaranteed: it depends on pipeline, CPU, system scaling and resources like Augmented Images or Faces, which can trigger internal buffering.
Buffer decouples capture and rendering
When the GPU or UI fluctuates, a queue allows threads to work with less blocking. The effect depends on whether the rest of the pipeline consumes frames correctly.
Consistency can cost latency
More available frames also mean that the displayed image can lag behind the world. Fine alignment cases need to measure motion-to-photon and not just FPS.
Memory grows with resolution and queue
The benchmark estimates around 6MB per additional 1080p texture. On limited devices, memory pressure can negate the benefit.
Resources can implicitly enable buffering
Augmented Images, Augmented Faces, and some devices use internal queues. Diagnostics must confirm the status before assigning delay to the application code.
Nexus Reading
Buffering is an engineering lever, not an automatic optimization. Decision must combine frame time, variance, memory, heat and spatial registration error.
FAQ
Does buffering always improve FPS?
No. The documentation itself says that the gain depends on the pipeline, threading, architecture and operating system.
What is the memory cost?
An additional 1080p frame occupies approximately 6 MB in the official example.
How do I know which texture is the current frame?
Use Frame.getCameraTextureName() after supplying the queue with setCameraTextureNames().
Essential guides to delve deeper into the decision
This editorial analysis was produced by Nexus from the official sources below, consulted on September 21, 2026. The text is original and interprets practical implications for companies.
- Google ARCore — Buffering camera frames: circular queue, activation, memory, delay, performance and texture APIs.
- Google ARCore — Use the device camera to capture reality: Overview of configuration, metadata, buffering and camera sharing.
Date reported by main source: official documentation consulted on September 21, 2026; technical analysis.