MPEG-4 AAC medical distortion study warns against using audio codecs for neurophysiology

The failure of consumer-grade audio codecs to preserve neurophysiological integrity highlights a critical gap in medical data interoperability. While MPEG-4 AAC excels at reducing file sizes for music by discarding 'imperceptible' sounds, those same discarded frequencies often contain the primary markers for neurodegenerative diseases and epilepsy. This findings shift the industry focus away from general-purpose multimedia formats toward dedicated biomedical standards like ITU-T H.BWC. As hospitals move toward massive cloud-based archiving, the streaming industry must recognize that perceptual fidelity does not equal scientific accuracy. Watch for the adoption of the proposed MPEG T.261 standard as a specialized alternative for high-resolution medical waveform compression.

The push for dedicated biomedical waveform compression standards has gained institutional momentum as hospitals confront the limitations of consumer audio codecs in clinical settings. The ITU-T H.BWC recommendation, which proposes a lossless and near-lossless compression framework specifically designed for biosignals, has been under active development within ITU-T Study Group 16 since 2023, with working documents addressing EEG, EMG, and ECG archiving requirements. The Fraunhofer Institute for Integrated Circuits, which developed the FDK AAC reference implementation widely used in medical device firmware, has not publicly commented on the neurophysiology distortion findings, but its codec licensing portfolio remains central to how DICOM-compliant devices handle waveform data today.

On the regulatory and standards side, the DICOM Working Group 31 (Media Storage) has been evaluating whether to formally restrict or deprecate perceptual audio codecs for neurophysiology storage. The IEC 62304 standard for medical device software lifecycle requires that any compression algorithm used in a clinical device must not introduce artifacts that could alter diagnostic interpretation, a threshold that the new study suggests MPEG-4 AAC fails to meet for EEG and EMG signals. Natus Medical, a major manufacturer of EEG acquisition systems that uses the EEGnet platform referenced in the study, has not yet announced a codec migration timeline, but the findings could accelerate vendor timelines for adopting lossless or medically validated alternatives in their archiving pipelines.

Technical benchmarks from the study align with earlier independent analyses of perceptual codecs applied to non-audio signals. Research published in Physiological Measurement in 2024 demonstrated that MP3 compression at 128 kbps introduced spectral artifacts in EEG data that mimicked pathological high-frequency oscillations, a finding consistent with the AAC distortion patterns now documented. The broader streaming and encoding industry has largely ignored medical waveform fidelity because DICOM neurophysiology files represent a tiny fraction of total compressed data volume, but as telemedicine and remote monitoring expand, the volume of EEG and EMG data flowing through cloud archiving systems is growing. The global digital neurophysiology market was valued at approximately $1.2 billion in 2024 and is projected to grow at a compound annual rate above 8% through 2030, driven by aging populations and increased epilepsy monitoring, which means codec selection decisions made today will affect millions of patient records.