What is Gigapixel Photography? – A Definition
A Gigapixel image is a digital photograph consisting of at least one billion pixels. Since no standard camera sensor can capture this information density in a single shot, a Gigapixel image is created by systematically stitching hundreds or thousands of individual captures – taken with telephoto lenses and assembled by specialised software.
The result: a single image in which you can switch between an overview of an entire city and reading a single inscription on a church tower – without any loss of quality. Zooming in reveals stories and details normally invisible to the naked eye.
How is a Gigapixel Image Created?
The creation of a Gigapixel image follows a precise workflow – and exploits the boundaries of physics in a targeted way:
The Diffraction Limit – Why More Pixels Does Not Always Mean More Information
Every lens has a physical diffraction limit: the aperture bends light, which limits resolution – regardless of how many pixels the sensor has. The rule of thumb is: Maximum useful aperture = 2 × pixel pitch (in µm). Beyond this limit, additional pixels record no new genuine information.
Concrete examples at aperture f/8 (Source: photoscala.de, 2012):
| Sensor format | Max. genuine MP at f/8 | Typical sensors |
|---|---|---|
| Smartphone (2/3") | ~2 MP | 48–200 MP sensors – all far above the limit |
| APS-C | ~13 MP | 24–33 MP sensors – often above the limit |
| Full frame | ~30 MP | 24 MP optimal, 61 MP exceeds the limit |
| Medium format (40×54 mm) | ~76 MP | 100 MP sensors exceed the limit |
Gigapixel images completely bypass this physical limit: each individual shot is taken at the optimal aperture – the sensor uses its full potential. Hundreds of such optimal shots are assembled into a single final image. The result: original image data far beyond what any single sensor could ever deliver.
1. Capture
Camera with suitable lens – from wide-angle (e.g. for 360° spherical panoramas) to telephoto, depending on the subject. Shots in rows and columns with 15–30% overlap. A tripod and nodal point adapter simplify the process considerably but are not mandatory – Gigapixel images can also be created handheld or with robotic systems, as long as the subject is static. Manual settings for exposure, focus, ISO and white balance are essential. RAW format recommended.
2. Stitching
Specialist software (e.g. PTGui, Autopano) assembles the individual shots into a seamless whole. Automatic and manual placement of overlap points, horizon determination to prevent converging lines. Processing time: hours to days depending on size and hardware.
3. File formats & output
JPEG is limited to approximately 0.9 gigapixels (max. 30,000 × 30,000 pixels). TIFF is usable up to approximately 3 gigapixels – limited by the 4 GB file size (gigabytes and gigapixels are different units; storage requirements depend on image data). For larger images, the PSB format (Photoshop Large Document, up to 300,000 × 300,000 pixels) is used. Output as HTML5 zoom viewer, as print file for surfaces over 200 m², or as VR application.
Pioneer expertise: Daniel Richter of Gigapixel GmbH created the first terapixel image in the world in 2012 – documented in the specialist magazine Photographie (12/2012). He has specialised in the professional creation and licensing of Gigapixel images for over 15 years.
Gigapixel vs. AI Upscaling – A Critical Difference
A Gigapixel image from Gigapixel GmbH contains original, non-interpolated image data. This fundamentally distinguishes it from AI upscaling software like Topaz Gigapixel AI. Real Gigapixel photography captures actual light information measured by the camera sensor. AI upscaling calculates missing pixels based on learned patterns – no new real information is created. When zooming into genuine Gigapixel images, ever-new physically correct details are revealed. With AI interpolation, artefacts appear beyond approximately 4× magnification. Ashraf et al. (2024) / University of Cambridge / Meta confirm: the human eye resolves up to 94 ppi in foveal vision – individually up to 120 ppi. The difference between original resolution and AI interpolation is perceptible.
Where are Gigapixel Images Used?
Wall Design & Stretch Ceilings
Pixel-sharp on XXL surfaces over 100 m² – viewed at close range. The only image format that guarantees this quality at arm's length.
Healthcare
Nature images in Gigapixel resolution for clinics, care homes and medical practices. Scientifically proven healing-promoting – pixel-free even at 50 cm viewing distance.
Tourism & Cities
Interactive city panoramas for visitor centres and tourism portals. Average 4 minutes viewing time with Gigapixel viewers.
Art & Culture
Digitisation of artworks at highest resolution – Rembrandt's Night Watch fully documented at 717 Gigapixels.
Frequently Asked Questions – What is Gigapixel Photography?
What is the difference between megapixels and gigapixels?
One megapixel (MP) equals one million pixels. One gigapixel (GP) equals one billion pixels – a thousand times more than a megapixel. More pixels on a sensor does not automatically mean more genuine image information. A good full-frame camera with 24 MP at f/8 operates near the optical optimum (~30 MP maximum resolvable). High-density sensors exceed this limit: smartphones with 48–200 MP on tiny sensors physically deliver only ~2 MP of genuine image information at f/8.
Why do smartphone cameras with 48–200 MP not deliver genuine megapixels?
The physical diffraction limit depends on sensor format and aperture. The rule of thumb: maximum useful aperture = 2 × pixel pitch in micrometres. Smartphone pixels are typically 0.7–1.0 µm – meaning the diffraction limit is already reached at f/1.5 to f/2. Beyond this narrow opening, no additional pixel records new genuine information. On the small sensor (approx. 2/3 inch), only around 2 megapixels of genuine information are physically possible at f/8 – regardless of whether the sensor has 48 or 200 MP. Market average 2024: 54 MP (Counterpoint Research). (Source: photoscala.de, 2012)
Which lens do I need for Gigapixel photography?
It depends on the subject. The decisive factors are sensor size, pixel pitch (for the diffraction limit) and the field of view to be covered. For 360° spherical panoramas, a normal wide-angle lens can be suitable. For city panoramas from a distance, telephoto lenses between 85 and 300 mm (full-frame equivalent) are optimal – above approximately 450 mm, atmospheric shimmer in the air mass between camera and subject reduces image quality and a higher gigapixel count brings no quality gain.
How large can a Gigapixel image be printed?
The required resolution always depends on the viewing distance. As a rule of thumb: for 1 m² of print area with close viewing (approx. 1 m distance), approximately 25–30 MP of genuine image data are required. Gigapixel GmbH images can be found on entire building facades with over 200 m² of print area. Standard stock photos with 25 MP are insufficient already at a few square metres with short viewing distances.
What are the limits of AI upscaling – and can it match Gigapixel images?
AI upscaling software such as Topaz Gigapixel AI is trained on small image tiles (typically 256×256 or 512×512 pixels) and processes these serially. At extreme magnification factors (above 4×), artefacts and invented details appear – the software guesses structures that do not exist. Many tools are also limited by RAM or file size and cannot fully process genuine Gigapixel files at all. Gigapixel GmbH images contain original image data from the outset – no upscaling, no invented details.
How much does a Gigapixel image cost?
Print licences in the Gigapixel GmbH portal start from 90 euros. For commissioned productions – individual city panoramas, architectural documentation or industrial projects – we provide an individual quote on request.
In Brief
Gigapixel photography is the technique for creating images with at least one billion pixels – created by stitching hundreds of original camera captures, each making optimal use of the physical diffraction limit of the lens. Gigapixel GmbH offers licensable Gigapixel images from 90 euros without AI upscaling – for wall designs over 200 m², clinics, exhibition construction, tourism and architecture. Zoom to full original resolution in the HTML5 viewer before purchasing.