AQuA is Sevana’s audio quality analyzer for perceptual voice and audio quality testing. You compare a reference recording with a test recording, and AQuA returns a MOS score plus clear reports on what degraded the sound. Because it models human hearing, it estimates quality the way a listener would, and it positions itself as a flexible alternative to POLQA and ViSQOL.
Download presentationDownload manualRequest a demo
What the AQuA audio quality analyzer does
Reference vs test, scored
First, compare two audio files to get an objective MOS on the ITU-T P.800 scale, so you know exactly how close the test is to the reference.
Built on a hearing model
Because a perceptual model uses critical, logarithmic and resonator bands, it estimates quality as a listener would perceive it.
Impairment reports
Moreover, with PVQA integrated, AQuA pinpoints the impairments that degraded the test audio and were absent from the reference.
CLI or library
In addition, you can run AQuA as a command-line utility, or embed it through the C++ or Python library, on Linux and Windows alike.
Who uses AQuA
Because AQuA fits across the call-quality chain, different teams put it to work in different ways.
Used by
- Mobile operators
- Telecom service providers
- 3G, 4G/LTE and 5G testing and measurement
Ideal for
- Mobile network drive testing
- Unified-communications call quality monitoring
- Player voice analysis in gaming
- IVR testing
- Load testing
Why teams choose AQuA
Therefore teams choose AQuA for several practical reasons:
- A server solution with no platform restrictions, easy to build into your own software.
- Runs on any Linux or Windows server, 32-bit and 64-bit.
- Perceptual estimation across VoIP, GSM, PSTN, ISDN and CDMA networks and their combinations.
- A configurable perceptual model (ask us what that means for your case).
- Works with audio shorter than 4 seconds and longer than 30 seconds, in any language, mono or stereo.
- Returns three report types alongside the score, each explaining a cause of quality loss.
See AQuA in action
Ultimately, AQuA is your means for fast, straightforward and cost-effective end-to-end voice and audio quality testing.
Technical details
Features
For instance, AQuA delivers:
- MOS score on the ITU-T P.800 scale
- Accounts for noise and packet loss
- Voice quality measurement in any language
- Works with long and short audio
- Audio synchronisation in time and level domains
- Considers loudness loss
- Signal-related metrics that explain audio quality loss
Technology
The human ear is a non-linear system that produces masking, where one sound hides another against a noisy background. Therefore AQuA’s perceptual model draws on the research of four scientists into how different frequency ranges, the so-called critical bands, shape what we hear. Moreover, AQuA measures energy in critical, logarithmic and resonator bands, so it captures more properties of hearing and speech than any single-band approach.
Metrics
Beyond the MOS score, AQuA reports similarity between reference and test audio, duration and timing distortion, spectrum damage and vibration, amplification or attenuation, the most affected bands, pitch statistics, quantization distortion, clipping level, RMS, and energy and sample statistics. In addition, the PVQA layer surfaces impairments in the test audio that were not present in the reference.
For your call quality testing needs
Talk to our team about putting AQuA to work on your audio and voice quality.
