Long-form article
The Hidden Cost of Large Point Clouds: Why an E57 Cutter Becomes Necessary
Why very large E57 point clouds create hidden engineering costs, and why targeted extraction is becoming an essential workflow layer.
CloudCutter development journal
CloudCutter articles
Seventeen publications collected on VoxelStudio: from the original point-cloud workflow problem to multi-format architecture and reproducible benchmarking.
Complete archive
Titles, dates and subjects
Every entry has its own permanent VoxelStudio page, metadata, canonical URL and previous/next navigation. Historical product names are retained only where they explain the development sequence.
Phase I
Defining the workflow problem
The first publications explain why complete registered point clouds can remain difficult to use in local engineering tasks and introduce the missing preparation layer.
Read article
Long-form article
CloudCutter_E57: The First Practical Breakthrough
The first practical CloudCutter milestone: sampled preview, spatial selection, streaming extraction, and lightweight E57 outputs.
Video update
What CloudCutter_E57 v1.0 Could Do
A concise demonstration of the first CloudCutter_E57 workflow for selecting and extracting useful subsets from large point clouds.
Visual workflow note
The Missing Layer in Scan-to-CAD Workflows
Three workflow diagrams show how CloudCutter can turn a monolithic point cloud into smaller task-specific engineering datasets.
Long-form article
The Missing Layer in Scan-to-CAD Workflows - Continued
Why point-cloud preparation between registration and modelling deserves its own controlled engineering workflow.
Phase II
First public product
The second phase documents the first demo, the public beta, and early workflow observations. Claims are framed as development evidence rather than universal comparisons.
Long-form article
CloudCutter_E57 Demo Version: The First Practical Step
The first CloudCutter_E57 demo translated the streaming architecture into a workflow that engineers could test on real datasets.
Long-form article
Why Nobody Fixes This Bottleneck - Even Though Everyone Has It
The point-cloud preparation bottleneck persists because it sits between scanning and modelling and is rarely owned as a separate process.
Short visual post
The Illusion of 'It Works'
A point-cloud workflow can be technically functional while remaining inefficient, manual, and unnecessarily expensive.
Release update
CloudCutter v1.0 Beta Became Available
CloudCutter v1.0 beta introduced practical multi-output E57 splitting with preserved point attributes and structured scan data.
Video test note
CloudCutter_E57 and FARO Scene: A Practical Workflow Observation
A cautious practical comparison of reading time, memory behavior, and export workflow using one 5 GB E57 source file.
Phase III
Multi-format architecture
The third phase follows the expansion from E57-only processing to a multi-format cutter and converter, then explains the reading and memory architecture behind it.
Video release update
CloudCutter v1.2: Multi-Format Processing and 3D Preview
CloudCutter v1.2 expanded the original E57 workflow to XYZ ASCII, LAS, LAZ, multi-file processing, and integrated 3D visualization.
Visual concept deck
From CloudCutter_E57 to a Multi-Format Cutter and Converter
A five-slide visual explanation of CloudCutter's E57, LAS, and XYZ ASCII input workflows and conversion outputs.
Technical article
E57 Reading Principles for Large Point Clouds
A practical overview of E57 metadata, compressed point records, images, block streaming, and bounded-memory processing.
Technical article
LAS Reading Principles for Large Point Clouds
How LAS headers, VLRs, point formats, scale and offset values, and streamed records support bounded-memory processing.
Technical article
XYZ ASCII Reading Principles
Why XYZ ASCII is simple to parse but difficult to interpret safely without explicit column, unit, scale, and coordinate-system metadata.
Technical architecture article
How to Avoid RAM Explosion with 500 GB Point-Cloud Files
Four bounded-memory rules for preview sampling, source rereading, stable tile indices, and block-based point-cloud export.
Phase IV
Evidence and benchmarking
The fourth phase establishes a transparent benchmark policy and connects development claims to reproducible measurements and verified outputs.
Benchmarking policy article
Does Software Built by One Person Need Serious Benchmarking?
Why independent engineering software needs transparent datasets, hardware disclosure, repeated runs, operator timing, and output validation.