MambaRate: Speech Quality Assessment Across Different Sampling Rates

2025/07/16 by Panos Kakoulidis, Kakoulidis, Panos, Iakovi Alexiou +11
Computer Science · #Audio and Speech Processing (eess.AS) #FOS: Computer and information sciences #FOS: Electrical engineering #Music and Audio Processing #Sound (cs.SD) #Speech Recognition and Synthesis #Speech and Audio Processing #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2507.12090

openalex publication_date 2025/07/16 · openalex created_date 2025/10/14 · openalex updated_date 2026/07/28

Abstract

We propose MambaRate, which predicts Mean Opinion Scores (MOS) with limited bias regarding the sampling rate of the waveform under evaluation. It is designed for Track 3 of the AudioMOS Challenge 2025, which focuses on predicting MOS for speech in high sampling frequencies. Our model leverages self-supervised embeddings and selective state space modeling. The target ratings are encoded in a continuous representation via Gaussian radial basis functions (RBF). The results of the challenge were based on the system-level Spearman's Rank Correllation Coefficient (SRCC) metric. An initial MambaRate version (T16 system) outperformed the pre-trained baseline (B03) by ~14% in a few-shot setting without pre-training. T16 ranked fourth out of five in the challenge, differing by ~6% from the winning system. We present additional results on the BVCC dataset as well as ablations with different representations as input, which outperform the initial T16 version.

Citations

• The VoiceMOS Challenge 2024: Beyond Speech Quality Prediction
• SALF-MOS: Speaker Agnostic Latent Features Downsampled for MOS Prediction
• Towards Out-of-Distribution Detection in Vocoder Recognition via Latent Feature Reconstruction
• Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality
• Mamba4Rec: Towards Efficient Sequential Recommendation with Selective State Space Models
• Robust Speech Recognition via Large-Scale Weak Supervision
• Fusion of Self-supervised Learned Models for MOS Prediction
• A Comparison of Deep Learning MOS Predictors for Speech Synthesis Quality
• WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing
• Generalization Ability of MOS Prediction Networks
• HuBERT: Self-Supervised Speech Representation Learning by Masked\n Prediction of Hidden Units
• How do Voices from Past Speech Synthesis Challenges Compare Today?
• MBNet: MOS Prediction for Synthesized Speech with Mean-Bias Network
• DNSMOS: A Non-Intrusive Perceptual Objective Speech Quality metric to\n evaluate Noise Suppressors
• wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations
• The Voice Conversion Challenge 2018: Promoting Development of Parallel and Nonparallel Methods
• Decoupled Weight Decay Regularization
• CNN Architectures for Large-Scale Audio Classification

Related

• HighRateMOS: Sampling-Rate Aware Modeling for Speech Quality Assessment
• Multi-Sampling-Frequency Naturalness MOS Prediction Using Self-Supervised Learning Model with Sampling-Frequency-Independent Layer
• SALF-MOS: Speaker Agnostic Latent Features Downsampled for MOS Prediction
• From Scores to Preferences: Redefining MOS Benchmarking for Speech Quality Reward Modeling
• MBNet: MOS Prediction for Synthesized Speech with Mean-Bias Network