Codecs in Voice Over IP: A Review and Comparative Analysis of Signal Fidelity

Purpose: To review the main speech codecs used in Voice over IP (VoIP) and to compare how closely selected codecs reproduce a reference speech signal in the time and frequency domains.

Materials and Methods: Nine codecs (ITU-T G.711, G.722, G.723.1, G.726, G.728, G.729, iLBC, iSAC and Opus) were reviewed. An 11.33 s speech recording and its versions coded with G.711 (A-law and μ-law), G.729, iLBC (8 kHz), iSAC and Opus (48, 24, 16 and 8 kHz) were compared in MATLAB using waveform and spectrum plots, descriptive statistics, SNR, PSNR and MSE.

Results: All coded signals preserved the temporal structure and spectral envelope of the original. PSNR ranged from 23.44 to 23.98 dB and MSE from 0.0040 to 0.0045, and the original-to-coded power ratio (SNR index) from 2.13 dB (Opus 48 kHz) to 5.01 dB (G.729). Opus 48 kHz retained the highest spectral peak (78% of the original) and G.711 μ-law the lowest (43%).

Conclusions: The codecs differed little in signal fidelity, so none can be ranked superior on these measures alone. Selection should rest on bandwidth, delay, packet-loss robustness, licensing and interoperability, ideally confirmed with subjective tests. Results come from a single sample and are indicative only. (J Multidiscip Eng 2026;1:102-109)‎