Xiaomi 12T Pro Camera test results

For scoring and analysis, SBMARK engineers capture and evaluate more than 2,600 test images in both controlled laboratory environments and natural outdoor, indoor and low-light scenes, using the camera’s default settings. The photographic protocol is designed to take into account the main use cases and is based on typical shooting scenarios, such as portraits, family photographs and landscapes. Evaluation is done by visually inspecting images Cons a natural scene reference and making objective measurements on graphical images captured in the lab under various lighting conditions from 1 to 1,000+ lux and color temperatures from 2,300K to 6,500K.

Exposure is one of the key attributes for technically good images. The main attribute evaluated is the brightness of the main subject across various use cases such as landscape, portrait or still life. Other factors evaluated are contrast and dynamic range, eg. the ability to make details visible in both light and dark areas of the image. Repeatability is also important because it demonstrates the camera’s ability to deliver the same rendering when shooting multiple images of the same scene.

Color is one of the key attributes for technically good images. The image quality attributes analyzed are skin tone rendering, white balance, color shading and repeatability. For the rendering of color and skin tone, we penalize unnatural colors but respect the manufacturer’s choice of the chromatic signature.

Autofocus tests focus on focus accuracy, focus repeatability, shutter lag and depth of field. The shutter lag is the difference between the time the user presses the capture button and the time the image is actually taken. It includes focus speed and the ability of the device to capture images at the right moment, what is called “zero shutter lag” capability. While shallow depth of field can be nice for a single-subject portrait or close-up, it can also be a problem in some specific conditions such as group portraits; Both situations are tested. Focus accuracy is also evaluated in all real-life images taken, from infinity to close-up objects and in low light to outdoor conditions.

Texture tests analyze the level of detail and texture of subjects in images taken in the laboratory as well as in real-life scenarios. For natural shots, special attention is paid to the level of detail in the light and dark areas of the image. Objective measurements are performed on map images taken under various lighting conditions from 1 to 1000 lux and different types of dynamic range conditions. The charts used are the proprietary SBMARK (DMC) chart and the Dead Leaves chart.

Noise tests analyze various attributes of noise such as intensity, chromaticity, grain, texture on real-life images and graphical images taken in the laboratory. For natural images, special attention is paid to noise on faces, landscapes, but also on dark areas and high dynamic range conditions. Noise on moving objects is also evaluated on natural images. Objective measurements are performed on graphical images taken under various conditions from 1 to 1000 lux and different types of dynamic range conditions. The graph used is the Dead Leaves graph and standardized measurement such as visual noise derived from ISO 15739.

Artifact evaluation examines lens shading, chromatic aberrations, geometric distortion, edge ring, halos, ghosting, quantization, unexpected variations in color tone, among other types of possible unnatural effects on photos. The more severe and frequent the artifact, the greater the point deduction from the score. The main artifacts observed and the corresponding point loss are listed below.

Bokeh is tested in a dedicated mode, usually in portrait or aperture mode, and analyzed by visually inspecting all images captured in the lab and under natural conditions. The goal is to reproduce portrait photography comparable to that taken with a DLSR and a wide aperture. The main image quality attributes to look out for are depth estimation, artifacts, blur gradient, and bokeh blur spotlight shape. The attributes of the portrait image quality (exposure, color, texture) are also taken into account.

The preview tests analyze the image quality of the camera app image preview, with particular attention to the difference between capture and preview, especially with regards to dynamic range and application of the bokeh effect. The smoothness of exposure, color and focus adaptation are also evaluated when zooming from the minimum to the maximum available zoom factor. The preview frame rate is measured using the Universal Timer LED.

SBMARK engineers capture and evaluate over 400 test images in controlled laboratory environments and outdoor, indoor and natural scenes in low light conditions, using the camera’s default settings and zooming in at various zoom factors, from ultra wide angle to ultra-long range zoom. Evaluation is done by visually inspecting images Cons a natural scene reference and making objective measurements of graphs captured in the lab under different conditions from 20 to 1000 lux and color temperatures from 2300K to 6500K.

These tests analyze the performance of the ultra-wide-angle camera at different focal lengths from 12mm to 20mm. All image quality attributes are evaluated, with particular attention to artifacts such as chromatic aberrations, lens softness and distortion. The images below are an excerpt of tested scenes.

All image quality attributes are evaluated at focal lengths from about 40mm to 300mm, with a focus on texture and detail. The score is derived from a series of objective measurements in the laboratory and from the perceptual analysis of real-life images.

SBMARK engineers capture and evaluate over 2.5 hours of video in controlled laboratory environments and natural low-light scenes, indoor and outdoor, using the camera’s default settings. The assessment consists of visually inspecting natural videos taken under various conditions and carrying out objective measurements on graphical videos recorded in the laboratory under different conditions from 1 to 1000+ lux and color temperatures from 2,300 K to 6,500 K.

Exposure tests evaluate the brightness of the main subject and the dynamic range, eg. the ability to make details visible in both light and dark areas of the image. The stability and temporal adaptation of exposure are also analyzed.

Image Quality Color Analysis examines color rendition, skin tone rendition, white balance, color shading, white balance stability and adaptation when the light changes.

Consistency tests analyze the level of detail and texture of real-life videos, as well as graph videos recorded in the lab. Natural video recordings are evaluated visually, with particular attention to the level of detail in bright areas and dark areas. Objective measurements are performed on images of graphs taken under various conditions from 1 to 1000 lux. The charts used are the SBMARK (DMC) chart and the Dead Leaves chart.

Noise tests analyze various noise attributes such as intensity, chromaticity, grain, texture, temporal aspects on real-life video recordings and graph videos taken in the laboratory. Natural videos are evaluated visually, with a focus on noise in dark areas and high dynamic range conditions. Objective measurements are performed on card videos recorded under various conditions from 1 to 1000 lux. The graph used is the SBMARK visual noise graph.

The stabilization assessment verifies the device’s ability to stabilize footage using software or hardware technologies such as OIS, EIS, or any other means. The assessment examines residual motion, smoothness, jelly artifacts, and residual motion blur on walk-and-run use cases under various lighting conditions. The video below is an excerpt from one of the tested scenes.

Artifacts are evaluated with MTF and ring measurements on the SFR graph in the lab, as well as frame rate measurements using the Universal Timer LED. Natural videos are visually evaluated with special attention paid to artifacts such as aliasing, quantization, locking and toning, among others. The more severe and frequent the artifact, the greater the point deduction from the score. The main artifacts and the corresponding point loss are listed below.