The primary acoustic cue of tones is F0, but secondary cues such as duration and intensity can also distinguish tones from one another. When pitch information is unavailable, do speakers utilize any of the secondary acoustic features to realize tones? This study situates Mandarin tones in a music context and explores tonal production patterns by speakers under different musical notes. The vowel duration of the four Mandarin tones in the context of normal speech and singing from sixteen native Mandarin speakers were analyzed. It was found that tones in the music context shows partially similar patterns with those in a normal speech context: syllables with T4 had significantly shorter vowel durations than those with other tones. However, the vowel duration of T1 increased, hence T3 was no longer the longest one. The results suggest that when pitch information is not available for the realization of tones, speakers may partially rely on vowel duration cues to express tonal contrasts, though the duration patterns of tones are affected by different communicative contexts.