Lexical tone recognition in multiple-talker babbles (N = 1, 2, 4, 8, 10, or 12) and in speech-shaped noise at different signal-to-noise ratios (SNRs = -18 to -6 dB) were tested in 30 normal-hearing native Mandarin-speaking listeners. Results showed that tone perception was robust to noise. The performance curve as a function of N was non-monotonic. The breakpoint at which the performance plateaued was N = 8 for all SNRs tested with a slight improvement at N > 8 at -6 and -9 dB SNR.