We subjected the Xiaomi 12S Ultra to our rigorous SBMARK Audio test suite to measure its performance both when recording sound using its built-in microphones, and when playing audio through its speakers.
In this review, we’ll break down how it fared across a variety of tests and several common use cases.
Overview
Key audio specs include:
- Three speakers (top front, top, bottom)
- No audio jack output
- Dual Dolby Atmos stereo speakers
- Harman / Kardon sound
Reproduction
Pros
- Impact dynamics despite less precision
- Good space performance across the board
- Nice extension and high-end bass
- Almost no artifacts
versus
- Timbre is very inconsistent and even worse in movie and game use cases
- Very aggressive sound when using the game at maximum volume, as well as a disappointing tonal balance
- The stereo image is not as accurate despite the good overall results
Registration
Pros
- Great timbre, with natural and pleasant tonal balance
- Excellent spatial and dynamic performance overall
- Excellent wind noise performance
versus
- Too much compression and distortion in general
- Disappointing audio zoom results
The Xiaomi 12S Ultra’s SBMARK Audio score of 129 is average for a device in its class. In Playback, it’s better suited for listening to music, scoring slightly lower in game and movie use cases. The dynamics could be more accurate but overall they have an impact. Good amplitude and localizability are accompanied by nice extension and high-end bass. The sound is also virtually free of artifacts. Areas with potential for improvement include inconsistent timbre, aggressive sound and suboptimal tonal balance during gameplay, and an inaccurate stereo image.
The 12S Ultra works quite well in all use cases as a recording device. Recordings feature exceptional timbre with natural tonal balance, as well as excellent spatial and dynamic rendition. Wind noise reduction does a good job of maintaining intelligibility in windy conditions. However, our testers also noted too much compression and distortion overall, and the results of the audio zoom function are disappointing overall. Some background rejection is achieved, but this is compensated for by a loss of bass and mid-bass, as well as a decrease in volume consistency.
Test summary
Learn about SBMARK audio tests: For scoring and analysis in our smartphone audio reviews, SBMARK engineers perform a series of objective tests and undertake more than 20 hours of perceptual assessment under controlled laboratory conditions.
(For more details on our reproduction protocol, click here; for more details on our registration protocol, click here.)
The following section collects the key elements of our exhaustive tests and analyzes performed in SBMARK laboratories. Detailed performance evaluations in the form of reports are available upon request. Do not hesitate to contact us.
How the audio playback score is composed
SBMARK engineers test playback through smartphone speakers, whose performance is evaluated in our labs and under real-life conditions, using apps and default settings.
In playback, the 12S Ultra offers decent high-end extension and good bass power, but the highs, mids and lows all have room for improvement. The highs lack sharpness and sparkle, the mids sound hollow most of the time, and while the lows are powerful they aren’t that deep. In short, tonal balance lacks body and fullness. In terms of dynamics, the Xiaomi offers a rather crisp attack, fairly precise bass accuracy, and an overall solid punch, but we also observed some inaccuracies.
Overall the device performs well for the Spatial attribute, with a perfectly centered stereo scene, good localizability and decent distance perception. However, occasional bleeding between the left and right channels can result in slightly reduced amplitude and reduced localizability. While the minimum volume is almost too low, the maximum volume is quite loud but hampered by compression. In terms of artifacts, the Xiaomi 12S Ultra is very clean but the the right speaker can be easily occluded, resulting in a subtle tonal balance.
Hear about the playback performance of the smartphone tested in this comparison with some of its competitors:
Samsung Galaxy S22 Ultra (Exynos)
Recordings of smartphones that play some of our music at 60 LAeq in an anechoic environment from 2 microphones in AB configuration, at 30 cm
Here’s how the Xiaomi Xiaomi 12S Ultra fares in playback use cases compared to its competitors:
Playback of use case scores
The Timbre score represents how well a phone reproduces sound across the audible tonal range and takes into account bass, midrange, treble, tonal balance and volume dependence. It is the most important attribute for reproduction.
Music playback frequency response
A 1/12 octave frequency response graph, which measures the volume of each frequency emitted by the smartphone when playing a pure sine wave in an anechoic environment.
The Dynamics score measures the accuracy of changes in the energy level of sound sources, such as the precision with which a bass note or the impact sound of drums is played.
Secondary attributes for spatial tests include identifying the position of a specific sound, its positional balance, distance and amplitude.
The volume score represents the overall volume of a smartphone and how the volume gradually increases and decreases based on user input.
Here are some sound pressure levels (SPLs) measured when playing our sample recordings of hip-hop and classical music at maximum volume:
| Hip-Hop | Classic | |
| Xiaomi 12S Ultra | 72.6 dBA | 69 dBA |
| Samsung Galaxy S22 Ultra (Exynos) | 74.1 dB | 70.2 dBA |
| Apple iPhone 14 Pro Max | 74 dBA | 71.1 dBA |
The graph below shows the gradual changes in volume from minimum to maximum. We expect these changes to be consistent across the range, so that all volume levels match user expectations:
Music volume texture
This line graph shows the relative playback volume versus the user-selected volume step, measured at different volume levels with correlated pink noise in an anechoic box recorded on axis at 0.20 meters.
The Artifacts score measures the extent to which sound is affected by various types of distortion. The higher the score, the less noise you notice. Distortion can occur due to the sound processing in the device and the quality of the speakers.
Total harmonic distortion reproduction (maximum volume)
This graph shows the total harmonic distortion and noise in the audible frequency range.
It represents the distortion and noise of the device reproducing our test signal (0 dB Fs, Sweep Sine in an anechoic box at 40 cm) at the maximum volume of the device.
How the score of the audio recording is composed
SBMARK engineers test the recording by evaluating recorded files on reference audio equipment. These recordings are performed in our labs and in real-life conditions, using apps and predefined settings.
The Xiaomi 12S Ultra performs very well overall as a recording device, even with high-volume content. Tonal balance is pleasant and natural, with clean highs. The midrange is a little less solid but remains warm and soft. Low-end focused sounds are deep and powerful, but can also be fuzzy or boomy at times. The dynamic performance is very good, with an accurate envelope in all use cases and therefore a great intelligibility. However, our testers observed some compression.
Xiaomi 12S Ultra does a great job of rendering the soundstage in its recordings. The scene is large and the individual sound sources can be easily identified. Distance rendering is good too, but a slight lack of clarity means voices can sound farther than expected. Using audio zoom results in slight background rejection, but also results in a loss of bass and mids.
The volume of recordings is good across all use cases and apps. However, while artifacts are not a problem in normal use, at high volumes the pumping can become severe and distortion is also noticeable. On the plus side, microphone occlusions don’t have a big impact on sound quality. Wind noise reduction works very well at moderate wind speeds, but intelligibility becomes rather poor in strong gusts. Background sounds good and natural in recordings, but lacks lower treble clarity. Too much compression is heard when loud noises are recorded.
Here’s how the Xiaomi Xiaomi 12S Ultra fares in registering use cases compared to its competitors:
Record of use case scores
The Timbre score represents how well a phone captures sounds across the audible tonal range and takes into account bass, midrange, treble and tonal balance. It is the most important attribute for registration.
Life video frequency response
A 1 / 12th octave frequency response graph, which measures the volume of each frequency captured by the smartphone when recording a pure sine wave in an anechoic environment.
The Dynamics score measures the accuracy of changes in the energy level of sound sources, such as how accurately the explosives of a voice (p, tek, for example) are reproduced. The score also considers the signal-to-noise ratio (SNR), such as how loud the main voice is compared to the background noise.
Secondary attributes for spatial tests include locating the position of a specific sound, its positional balance, distance and amplitude on recorded audio files.
Directivity of registration
Smartphone directivity graph when recording test signals using the camera app, with the main camera. It represents the acoustic energy (in dB) on the angle of incidence of the sound source. (Normalized to the 0 ° angle, in front of the device.)
The volume score represents the normalization level of the audio on the recorded files and how the device handles noisy environments, such as electronic concerts, during recording.
Here are the sound levels recorded in the audio and video files, measured in LUFS (Loudness Unit Full Scale); as a reference, we expect loudness levels to be higher than -24 LUFS for recorded content:
| Match | Life video | Selfie video | Memo | |
| Xiaomi 12S Ultra | -28.3 LUFS | -20.8 LUFS | -19.1 LUFS | -20.4 LUFS |
| Samsung Galaxy S22 Ultra (Exynos) | -28.8 LUFS | -21.7 LUFS | -21.2 LUFS | -23.9 LUFS |
| Apple iPhone 14 Pro Max | -26 LUFS | -22.9 LUFS | -19.2 LUFS | -19.6 LUFS |
The Artifacts Score measures the extent to which recorded sounds are affected by various types of distortions. The higher the score, the less noise you notice. Distortions can occur due to the sound processing in the device and the quality of the microphones, as well as user handling, such as the way the phone is held.
In this audio comparison, you can hear how this smartphone handles wind noise compared to its competitors:
Recordings of a voice sample with light background noise, facing a turbulent wind of 5 m / s
Background evaluates how naturally the various sounds around a voice blend into the video recording file. For example, when recording a speech at an event, the background should not interfere with the main voice, but should provide context for the surrounding environment.
Start a new Thread