Speech Quality Benchmarking for Communications Between NASA Advanced Reference Spacesuits
Speech signal quality and intelligibility for spacesuit communications is a function of the suit’s acoustics, background noise, and audio system. Communications audio between two of NASA’s advanced reference-design spacesuits without commcaps (formally xEMU) was recorded under various EVA temperature and pressure conditions. Those data were subjected to jury-based intelligibility (ANSI Modified Rhyme Test) and algorithm-based quality (ITU POLQA) speech benchmarking for both worst-case and intended (NC-52) ambient-noise conditions in the suit. Point-by-point comparisons can identify segments of the comms path that most significantly degrade speech fidelity. Thus, algorithm-driven benchmarks like POLQA were applied to recordings obtained at different points in the communications signal path between suits. Subjective and POLQA tests only provide single-value scores. However, use of objective speech/phoneme-rate operations such as normalized-RMS-differences (NdRMS) applied to speech-pattern spectrograms computes a benchmark score and highlights degradation both in time and frequency. We discuss POLQA and NdRMS results that infer that – other than by noise generated by the suit’s Portable Life Support System (PLSS) – suit-to-suit speech quality is most affected by the transmitting helmet’s acoustics rather than audio and communications electronics or algorithms.