Pix4D and Metashape are excellent photogrammetry tools. They produce accurate, detailed outputs from drone imagery — point clouds, orthomosaics, elevation models, 3D meshes. The processing problem is largely solved.
The delivery problem is not. A client who receives a zip file containing a LAZ point cloud, a GeoTIFF orthomosaic, a DTM, an OBJ mesh, and a quality report PDF has data they cannot use. The PDF is accessible. Everything else requires specialist software they almost certainly do not have. This post covers how to bridge that gap — from Pix4D or Metashape output to a professional client delivery where every file is viewable in a browser.
What Pix4D and Metashape produce
Both tools take overlapping drone images as input and produce the same core set of outputs, though the exact naming conventions and export options differ slightly.
| Output | Pix4D name | Metashape name | Format options |
|---|---|---|---|
| Dense point cloud | Point Cloud | Dense Cloud | LAS, LAZ, PLY |
| Georeferenced aerial image | Orthomosaic | Orthomosaic | GeoTIFF (standard or COG) |
| Surface elevation model | DSM | DEM (surface) | GeoTIFF |
| Ground elevation model | DTM | DEM (terrain) | GeoTIFF |
| Textured 3D mesh | 3D Mesh | Model | OBJ, GLB, FBX, 3D Tiles |
| Processing summary | Quality Report | Processing Report |
For most survey deliveries, the relevant outputs are the LAZ, the GeoTIFF orthomosaic, the DTM or DSM, and the report PDF. The 3D mesh is valuable for visual presentations and engineering review but is optional depending on the project type.
Export settings that matter for delivery
Not all export settings produce equally deliverable outputs. A few choices upstream have a significant effect on what your client can see and interact with downstream.
LAZ vs LAS
Always export LAZ, not LAS. LAZ is losslessly compressed — typically 5 to 10 times smaller than an equivalent LAS file, with no data loss. A 4 GB LAS file is a 400–800 MB LAZ file. For a client delivery, this means faster uploads, faster streaming in the browser viewer, and lower storage costs. There is no reason to export LAS for delivery purposes. See what is a LAZ file for the technical background.
COG vs standard GeoTIFF
A Cloud Optimised GeoTIFF (COG) is structured so that a viewer can request specific tiles at specific zoom levels without downloading the whole file. A standard GeoTIFF is not — you have to download the whole thing before you can display any part of it. For large orthomosaics (common from drone surveys of significant areas), COG is substantially faster to view in the browser.
Metashape does not export COG natively, but you can convert with GDAL’s gdal_translate with the COG driver, or use QGIS (Layer → Export → Save As, select COG). Pix4Dmatic has COG export; Pix4Dmapper does not natively.
For deliveries of orthomosaics under about 200 MB, the difference is minimal. For large-area surveys producing 1–5 GB GeoTIFFs, COG makes a material difference to the viewer experience.
OBJ vs GLB for 3D meshes
If you are delivering a textured 3D mesh — common for building surveys, heritage documentation, and visual project reviews — export GLB rather than OBJ where possible.
OBJ requires three files: the .obj geometry, the .mtl material definition, and the texture images (often multiple .jpg files). This creates friction: you need to keep these files together, upload them correctly, and ensure the paths resolve.
GLB (binary glTF) is a single self-contained file. Geometry, materials, and textures are all packaged together. It is smaller than OBJ with the same textures, and it loads faster in Three.js-based browsers. Metashape exports GLB directly. Pix4D exports OBJ; convert with Blender or a tool like obj2gltf if you want GLB.
The delivery problem with a zip file
Most operators zip their Pix4D or Metashape outputs and either email the zip or upload it to Dropbox. The client receives a zip file, unzips it, and faces a folder containing:
project_pointcloud.laz
project_orthomosaic.tif
project_dsm.tif
project_dtm.tif
project_mesh.obj
project_mesh.mtl
texture_0.jpg
texture_1.jpg
quality_report.pdf
The client opens quality_report.pdf. They double-click project_orthomosaic.tif and get a washed-out black image in Photos (because GeoTIFFs render incorrectly in most standard image viewers — the geospatial metadata confuses them). They double-click project_pointcloud.laz and get an error. They give up.
The underlying data is correct and complete. It is just wrapped in formats that a client without GIS software cannot access. The delivery has failed even though the files arrived.
The professional delivery workflow
The solution is a spatial data delivery platform that provides the appropriate browser-based viewer for each format automatically.
Step 1: Export your final deliverables from Pix4D or Metashape
Use the export settings above: LAZ for the point cloud, COG GeoTIFF for the orthomosaic where possible, GLB for the mesh if required, PDF for the report.
Step 2: Upload to Swyvl
Create a site (the physical location) and a capture session (the date of the survey). Upload all your export files. Swyvl classifies each one by format:
- LAZ → Potree point cloud viewer
- GeoTIFF → MapLibre interactive map
- OBJ or GLB → Three.js 3D viewer
- PDF → inline document reader
Step 3: Add context
Write a brief note for the client — the coordinate system, datum, the GCP accuracy from the quality report, any areas where data quality was lower (shadows, water, low-overlap zones). This context is part of a professional delivery. It prevents the client having to ask.
Step 4: Share the portal link
Send the client a single link. They open a branded portal with all their deliverables, each viewable in the appropriate viewer. The portal does not expire. They can return to it six months later from the same URL.
What the client sees for each format
LAZ — Potree viewer
The point cloud loads progressively in the browser. A coarse version appears within a few seconds; detail builds as the user zooms in. The client can orbit freely, switch between colouring modes (RGB from the camera, elevation gradient, intensity), toggle classification classes, measure distances between points, and create cross-sections. For a full explanation of what these controls do, see what is a point cloud.
GeoTIFF — interactive map
The orthomosaic renders as a tiled map layer. The client pans and zooms as they would on Google Maps, with the survey imagery as the base layer. They can toggle to a satellite basemap underneath to see the site in context. If you have exported a DSM or DTM, those render as separate tiled layers with an elevation colour gradient.
OBJ or GLB — Three.js 3D viewer
Textured 3D meshes render in a standard Three.js viewer. The client can orbit around the model, zoom to details, and inspect the texture quality. This is particularly valuable for building, heritage, and infrastructure deliveries where the 3D model communicates something the point cloud or orthomosaic alone does not.
PDF — inline reader
The quality report appears inline. The client reads it without downloading anything. This is important: the quality report is the professional record of the survey methodology and accuracy. If the client has to download and open a PDF separately, they often do not bother. Inline, they read it.
Include the quality report in every delivery
This is a practical tip I have given many operators: always include the quality report in the same portal as the spatial data. Do not send it separately.
The quality report from Pix4D or Metashape contains the GCP accuracy, point density statistics, camera calibration, overlap map, and error metrics. For a client comparing deliverables from different providers, or for a project where accuracy matters, this is material information.
Buried in a zip file or sent in a separate email, it gets lost. In the portal alongside the point cloud and orthomosaic, it is contextually placed — the client views the data, sees the accuracy figures, and understands what they received.
Organising repeat surveys
If you are doing repeat surveys of the same site — construction progress monitoring, environmental monitoring, mine tracking — organise your delivery by site and session, not by project folder.
A client monitoring a construction project over six months should see:
- Site: [Project Name]
- Session: January 2026 — orthomosaic, point cloud, DSM, report
- Session: February 2026 — orthomosaic, point cloud, DSM, report
- Session: March 2026 — orthomosaic, point cloud, DSM, report
This time-series presentation shows progress clearly. The client can navigate between sessions, compare orthomosaics, and see how the site changed between visits. Three separate Dropbox emails do not communicate this — a structured delivery portal does.
For further reading on the complete drone survey delivery workflow, see how to deliver drone survey data to clients and the drone operators workflow guide.
The return on professional delivery
Professional delivery is not just polish. It is a commercial argument.
A client who can see all their deliverables in the browser — who actually interacts with the point cloud, zooms into areas of interest, reads the quality report — understands what the survey produced. That understanding is the basis for repeat work. If the data sits in a zip file nobody opens, the value of the survey is invisible.
Pix4D and Metashape produce the best outputs in the industry. The delivery workflow should be worthy of the data.