Objective:
Endoscopic vocal fold shortening (EVFS) surgery has been established as an effective technique for raising pitch in transgender women. However, the precise acoustic mechanisms underpinning voice feminization following this procedure are not fully understood. This study employs long-term average spectrum (LTAS) analysis to elucidate the spectral characteristics associated with gendered vocal qualities pre- and post-EVFS, specifically targeting shifts in spectral energy distribution as potential markers of gendered voice perception.

Methods:
The study analyzed voice recordings from 10 transgender women who underwent EVFS surgery. Recordings were collected pre-operatively and at 3 and 12 months post-surgery. LTAS measures, including spectral slope, formant cluster prominence (FCP), formant clustering frequency (FCF), and spectral energy in the 0-2 kHz frequency band, were obtained using Sopran software (Tolvan Data). A reference group of cisgender individuals (5 male, 5 female) provided baseline comparative data on typical gendered voice characteristics.

Results:
Significant spectral changes were observed across the three time points (baseline, post-op 3 months, post-op 12 months). Post-surgical results indicated an increase in energy within the lower frequency range (0-500 Hz) and a corresponding reduction at higher frequencies, suggesting a shift in spectral balance (p < 0.001). Notable LTAS changes included: FCP: decreased from baseline (3.6 dB) to post-op 3 months (1.7 dB) and slightly increased at 12 months (2.3 dB). FCF: increased from 3059 Hz at baseline to 3289 Hz at 3 months and further to 3470 Hz at 12 months. Spectral Slope: improved from -7.9 dB at baseline to -5.7 dB at 3 months, stabilizing at -5.9 dB at 12 months. Spectral Energy (0-2 kHz): showed a reduction from -58.6 dB at baseline to -54.6 dB at 3 months, with a slight decrease to -56.04 dB at 12 months.

Conclusion:
This preliminary analysis demonstrates significant acoustic changes in LTAS measures following EVFS, with results aligning with patients' self-reported perceptions of a softer vocal quality. The observed shifts in FCP and FCF may provide promising markers for further understanding of gendered voice characteristics post-surgery. These findings underscore the potential utility of LTAS metrics in evaluating gender-related vocal outcomes and suggest the need for further research to confirm these spectral patterns as reliable indicators of voice feminization in transgender patients.