Hi-Res 3D · AR-Ready
Byrst's photogrammetry engine transforms a structured photo shoot into a fully AR-ready 3D model - no 3D software, no specialized team. Follow this guide to get it right the first time.
Create 3D model4 steps · 40–120+ photos · AR-ready output
Choose the Right Object
Understand which objects produce the best 3D results and which to avoid - before you shoot.
Set Up Your Shoot
Equipment, lighting, camera settings, and environment preparation for optimal capture.
Photograph the Object
The three-orbit technique, 70% overlap rule, and shot count estimates by object size.
Upload to Byrst
Submit your images correctly for flawless 3D processing - what to do and what to avoid.
Byrst's photogrammetry engine detects distinct reference points across an object's surface. Follow these four steps for the best results.
Object selection is the most important step. The Byrst photogrammetry engine works by detecting distinct reference points across an object's surface. The more visual detail your object has, the better the result.
Allows the camera to capture distinct geometry from every angle
Gives the algorithm more reference points to match between photos
Surface detail creates unique landmarks the software can track
The object must stay exactly the same shape between all photos
Change shape between shots, breaking 3D reconstruction
Lack 3D geometry - try Byrst 3D-Xpress instead
Surfaces shift with camera position, providing no stable surface data
Gives the software nothing to track or match between photos
Proper equipment and consistent camera settings are essential. Once you begin shooting, do not change any camera settings - inconsistent settings will degrade or break the 3D reconstruction.
DSLR, mirrorless, or high-quality smartphone at maximum resolution
Steady, consistent shots across all angles and elevations
Rotating platform for smooth, consistent orbit capture
Softboxes or bounced light - even illumination, no hard shadows
Plain backdrop so images contain only the object, nothing more
Caution: Once you begin shooting, do not change any camera settings — including zoom, aperture, shutter speed, or ISO. Inconsistent settings will degrade or break the 3D reconstruction.
Position the object so it fills as much of the camera's frame as possible without excluding or cutting off any part. Each photo should share at least 70% of its content with the previous one. Capture three complete orbits at different elevations.
See the full orbit technique below
Once your photo session is complete, upload your images to the Byrst platform for 3D processing. Before uploading, review your photo set and remove any obviously blurry, overexposed, or duplicate frames. Do not crop or resize - the algorithm needs full-resolution originals with consistent framing.
See the upload process below
Three orbits at different elevations ensure the algorithm captures your object's geometry from every angle - top, sides, and bottom.
Camera at midpoint of object's height. Full 360° rotation, one photo every 10 - 15°. Captures sides, front, and back.
Raise camera above the object, angling down ~45°. Full rotation captures top surface and downward-facing geometry.
Lower camera, angling upward ~45°. Captures underside, legs, casters, and any downward-facing surfaces. Tip: turn object upside down if possible.
Each consecutive photo must share approximately 70% of its visible content with the previous photo. The photogrammetry algorithm finds the same surface points across multiple images - without sufficient overlap it can't build accurate 3D geometry.
Taking more photos is generally better than fewer. It's much easier to take extra shots during the session than to re-shoot later.
Position object so it fills the frame without cutting off any part
Keep camera at consistent distance from the object throughout each orbit
Shoot every 10–15° around the full 360° for each orbit
Each photo must overlap ~70% with adjacent photos
Don't start shooting until every item below is confirmed. This checklist prevents the most common reasons for failed 3D reconstructions.
Once your session is complete, upload your image batch to Byrst for processing. A few key rules before you hit submit.
1
Sign in to your account at byrst.com and navigate to the 3D Model creation section.
2
Remove any blurry, overexposed, or duplicate frames before uploading. Do not crop or resize.
3
Select all photos from your shoot and upload them together as a single batch for processing.
4
Byrst's photogrammetry engine generates your 3D model. You'll receive a notification when it's ready.
Do not crop or resize - the algorithm needs full-resolution originals with consistent framing.
Do not apply filters or post-processing - the 3D reconstruction depends on consistent, unaltered image data
What type of camera do I need?
Any DSLR, mirrorless, or high-quality smartphone camera works. The key requirement is shooting at maximum resolution with consistent settings. Do not change focal length, aperture, shutter speed, or ISO at any point during your shoot.
Why does the 70% overlap rule matter?
Byrst's photogrammetry engine finds identical surface points across multiple images to reconstruct 3D geometry. Without sufficient overlap, the algorithm can't match points between adjacent photos, resulting in holes or failed reconstruction. 70% overlap gives the engine enough shared content to reliably stitch images together.
How many photos do I really need?
At minimum: 40+ for small simple objects, 80–100+ for medium objects like dressers or chairs, and 120+ for large or complex objects. Taking more photos is always safer - it's far easier to capture extras during the session than to re-shoot. Only remove obviously blurry or duplicate frames before uploading.
Can I shoot in natrual light?
Consistent, diffused lighting is the goal. Natural light from a large north-facing window can work, but direct sunlight creates hard shadows that degrade reconstruction quality. Cloud cover provides excellent natural diffusion. Avoid shooting during times when light changes rapidly - consistency across all images is critical.
My object is skiny - what can I do?
Reflective surfaces are one of the hardest challenges for photogrammetry. Options include applying a temporary matte spray to the surface (removable), using polarizing filters on your lights and lens, or trying Byrst's 3D-Xpress component, which is designed for flat or thin objects that don't require full photogrammetry capture.
Can I edit my photos before uploading?
No. Do not apply any filters, color grading, cropping, resizing, or post-processing to your images before uploading. The 3D reconstruction engine depends on consistent, unaltered image data across the entire batch. Only remove obviously blurry, overexposed, or duplicate frames.
How long does processing take?
Processing time varies based on image count and object complexity. You'll receive a notification when your model is ready - you don't need to keep the browser tab open. Larger batches with more photos naturally take longer to process.
You now have everything you need — the technique, the checklist, and the four-step process. Log in to Byrst and create your first high-resolution 3D model today.
