PVQA is Sevana’s tool for passive call quality analysis. It measures voice quality non-intrusively, directly from the waveform, so it needs no reference signal and no test call. Because it listens the way a person would, it catches the echo, muffled sound, dead air and background noise that degrade a call, and reports a MOS score on the ITU-T P.800 scale.
What passive call quality analysis with PVQA delivers
Non-intrusive MOS
First, PVQA predicts a MOS on the ITU-T P.800 scale by waveform analysis alone, which is what lets it work on live and recorded traffic.
Finds the real faults
Moreover, it surfaces packet loss, silent and noisy calls, dead air, echo, VAD clipping and amplitude clipping, so you see exactly why a call sounded wrong.
Real-time and cross-platform
Because PVQA runs in real time as an app or a C++ or Python library, you can build your own quality testing without being tied to specific hardware.
Also powers AQuA
In addition, the PVQA engine sits inside AQuA, where it enriches the reference-based report with the same impairment detail.
Who uses PVQA
Because PVQA fits anywhere traffic flows, several kinds of team rely on it.
Used by
- Mobile operators
- Telecom service providers
- 3G, 4G/LTE and 5G test and measurement
Ideal for
- Call quality monitoring
- Proactive maintenance driven by QoE metrics
- Faster root-cause analysis of quality problems
- Discovering recurring impairment patterns
- Non-intrusive analysis where a reference is impossible
See PVQA in action
Ultimately, PVQA gives you hardware-independent voice quality testing in real time.
Technical details
Features
For instance, PVQA can:
- Score call quality with a MOS on the P.800 scale
- Detect packet loss through payload analysis
- Discover silent and noisy calls in the traffic
- Detect dead air
- Find echo and VAD clipping
- Flag amplitude clipping in voice traffic
Requirements
PVQA works with uncompressed .wav audio of telephone quality: 8 kHz or 16 kHz, 16-bit, mono. However, if you need other formats, higher sample rates or codec support, just contact us.
Scientific background
Clean telephony depends on impairments staying absent or negligible through encoding, transmission and amplification. Therefore PVQA’s perceptual model first detects the impairments known to degrade audio for human listeners, and then its proprietary algorithms convert them into a MOS prediction on the ITU-T P.800 scale.
For your call quality monitoring needs
Talk to our team about putting PVQA to work on your live or recorded traffic.
