Comparison
Neurones 3D vs CloudCompare, a CloudCompare alternative for deliverables
Neurones 3D is a CloudCompare alternative for people who need finished deliverables rather than a toolbox: automatic floor plans with sections, a terrain set with contours, trained classifiers and HTML reports, from one Windows program. CloudCompare stays the better choice if you need free, open-source, cross-platform software with a large plugin set.
By Dr. Florent Poux, 3D Geodata Academy · Updated
At a glance
Short statements, each one checkable. Where there is a number, it was measured, and the guides say on what.
- Checked 24 September 2026 against CloudCompare’s public information and reviews; every competitor fact below is dated and linked.
- CloudCompare is free and open source (GPL) and runs on Windows, macOS and Linux. N3D runs on Windows only and its source is not open.
- N3D’s free viewer opens, views and runs every process; writing results out needs a paid edition, bought once, with no subscription.
- Four trained classifiers ship in every N3D build: indoor (81.9 % and 82.3 % held-out), outdoor aerial LiDAR (86.2 %), powerline (87 to 88 %) and streetscape.
- Process › Floor plan… writes a dimensioned plan, sections and elevations as DXF, PDF and SVG: on a 43 million point survey, 42 spaces and 21 files in 1.0 s.
- N3D’s ground filter scores F1 90.3 % against IGN LiDAR HD’s own ground labels, with no training data.
- N3D’s M3C2 recovered an injected 50 mm lift as +0.04998 m; CloudCompare offers M3C2 as a plugin.
- The two work together: N3D’s LAS extra-byte columns (segment, instance, height above ground, deviation) are read back by CloudCompare.
Which one, for which job
When CloudCompare is the better choice
As published on 24 September 2026, CloudCompare is free and open source under the GPL, and runs on Windows, macOS and Linux. It covers comparison, registration and segmentation, and a large plugin set extends it: the CSF ground filter, M3C2, and the qCANUPO and 3DMASC classification plugins among them. Nothing is locked behind a licence, and you can read and change its source.
Choose CloudCompare if you work on a Mac or on Linux, if you need to save results without paying, if your institution requires open-source software, or if you rely on a specific plugin.
Reviewers on G2 cite a steep learning curve and a dated interface, as published on 24 September 2026. That is a cost in time rather than in money, and for many users the plugin set is worth it.
When Neurones 3D is the better choice
N3D is built around the deliverable at the end of the job rather than around individual operations. Four things set it apart:
- Automatic drawings. Process › Floor plan… finds the rooms, walls and openings and draws a dimensioned plan with sections and elevations to ISO drawing standards, saved as DXF R2000, PDF and SVG, with the building as GeoJSON and Shapefile. On S3DIS Area 4 (43 million points) that took 1.0 s.
- A terrain set, not a single raster. DTM, DSM, canopy height, slope, aspect, hillshade, contours, trees and sunlight, built from one ground fit on one grid, as GeoTIFF and GeoJSON.
- Trained classifiers, ready to run. Indoor, outdoor aerial LiDAR, powerline and streetscape models ship in every build, each writing a per-point confidence you can review and correct in a labelling grid.
- Reports. A self-contained HTML report with an as-built review, volumes and a provenance block; on the command line it exits with 0 on a pass and 2 on a fail.
It is also one Windows executable with no installer, no account and no telemetry, and it renders billions of points out of core: a 43.5 million point LAS opens in about 1.2 s, and a 127.8 million point, 7.9 GB E57 in about 22 s.
Capability by capability
“Plugin” means available through a CloudCompare plugin named in our sources. Where we have not checked a capability, the cell says see their docs rather than guess.
| Capability | CloudCompare | Neurones 3D |
|---|---|---|
| Price model | Free, open source (GPL) | Free viewer runs everything; saving results needs a one-time paid edition |
| Windows / macOS / Linux | Yes / Yes / Yes | Yes / No / No |
| Cloud-to-cloud comparison | Yes | Yes, with conformance bands |
| M3C2 change detection | Plugin | Yes, with a per-point level of detection |
| Registration | Yes | Yes (trimmed ICP, or graph matching with no initial placement) |
| Segmentation | Yes | Yes (unsupervised, plus primitive decomposition) |
| Ground filter | Plugin (CSF) | Yes (cloth simulation, F1 90.3 % vs IGN) |
| Classification | Plugin (qCANUPO, 3DMASC) | Yes, four trained models built in |
| Meshing | Plugin | Yes (Poisson, clean mesher, segment meshes) |
| Floor plans, sections, elevations to DXF | see their docs | Yes, automatic |
| DTM, contours, hillshade as GeoTIFF / GeoJSON | see their docs | Yes, one grid |
| HTML report with provenance | see their docs | Yes |
| Gaussian splats (PLY, SPZ) | see their docs | Yes, with level of detail |
| Source code open | Yes | No |
Using both
The two do not have to compete. N3D’s LAS export puts the class in its standard byte and segment, instance and component ids as exact 32-bit extra bytes, with height above ground or deviation as 32-bit floats, and CloudCompare reads those columns back. A common split is to classify, draw and report in N3D, and to use CloudCompare for a plugin N3D does not have.
What N3D does not do
It runs on Windows only. It does not write DWG (DXF, PDF and SVG instead), does not write OBJ, does not decimate meshes or repair them into manifolds, and does not check SLAM trajectories. Its classifiers report a confidence and keep a person in the loop rather than certify their own result. The FAQ gives the reason for each.
Every tool, side by side
The full comparison, tool by tool, with the price, the licence and a sourced cell for each capability, is on point cloud software compared.
Questions
Is Neurones 3D better than CloudCompare?
For some jobs, not all. CloudCompare is free, open source, cross-platform and extended by a large plugin set. N3D is stronger when the goal is a finished deliverable made in one run: a dimensioned floor plan, a terrain set with contours, a classified cloud with a per-point confidence, or a report. Many people will use both.
Is there a free alternative to CloudCompare?
N3D’s free viewer is free to download and runs every process, including the four trained classifiers, the ground filter, M3C2 and the floor plan preview, with no account and no trial clock. Unlike CloudCompare, writing results out is not free: it needs a paid edition, bought once.
Does N3D run on Mac or Linux like CloudCompare?
No. N3D is Windows only today (Windows 10 and 11, 64-bit). CloudCompare runs on Windows, macOS and Linux, which makes it the choice on those systems.
Can I open N3D results in CloudCompare?
Yes. A LAS export puts the class in the classification byte, segment, instance and component ids in exact 32-bit extra bytes, and computed values such as height above ground or deviation as 32-bit floats, and CloudCompare reads them back.
Does N3D have plugins like CloudCompare?
Its processes are built in rather than added. For automation there is a command line, n3d-render-cli, with more than thirty subcommands, and the Engine & SDK edition adds a Python package.
Is N3D easier to learn than CloudCompare?
Reviewers on G2 cite a steep learning curve and a dated interface for CloudCompare. N3D is organised around finished jobs (one menu entry for a floor plan, one for terrain products) and has a twelve-step guided tour under Help › Guided tour. Try both on your own data.
The guides behind it
Each guide gives the menu path, the steps, the command-line equivalent and the limits.
confidence column, and split a powerline corridor into individual trees, spans and poles.
Build deliverables
Floor plans, sections and elevations
One command from raw points to a drawing set - walls between their measured faces, doors with their swing, windows with their glass, dimension chains, room areas and levels, a legend, a scale bar and a north arrow - plus sections and elevations placed where they show the most.
Build deliverables
Terrain products (DTM, DSM, CHM, contours)
Turn one ground fit into the raster set a client actually receives (DTM, DSM, canopy height, slope, aspect, hillshade, contours, trees and sunlight), all on one grid so they cannot disagree, with a README that says what is wrong with them.
Compare and register
Compare and detect change (M3C2)
Compare a survey against a reference (another survey, a mesh or a BIM model) as a signed distance with conformance bands and a worst-first list of defects, or run M3C2 to say what was gained, what was lost, and what simply was not observed.
Deliver
Export a point cloud (LAS, PLY, CSV, USD, web)
One dialog for every point cloud export - format, subsample, colours, label columns and scalar fields - with a LAS writer that puts the class in the classification byte and every id in an exact 32-bit extra byte.
Deliver
Reports
One self-contained HTML document from what the session measured - the as-built review with its per-zone punch list, the volume and how much of it rests on the classification, the measurements you kept - and a provenance block that lets someone else re-derive every number.