Increased cognitive load has been observed to correlate with decreased vocal fold perturbation, reduced additive noise, increased periodicity, and a higher rate of vocal fold vibration. The aim of this study was to explore whether vocal fold vibratory patterns can serve as an indicator of increased cognitive load in non-balanced bilingual speakers when they use their weaker language. This is a comparative experimental study with a within-speaker design. We recorded a total of 95 bilingual speakers of Low German (LG), which is an endangered language spoken in Northern Germany, and Standard High German (HG). Participants completed four tasks in both languages: engaging in free narration, description of a picture story, giving directions, and reading a narrative passage. For the last three tasks the difficulty levels were varied. Measurements included jitter, shimmer, harmonics-to-noise ratio (HNR), cepstral peak prominence (CPP), the proportion of creak, pitch level, and pitch span. Changes of voice characteristics were examined both in terms of the participants age and in terms of their language dominance. For the latter, we calculated a dominance score derived from age of acquisition, frequency of use, and self-perceived linguistic competence in the two languages. Younger speakers showed a higher dominance of HG over LG, which decreased with age. Younger and more HG-dominant speakers exhibited lower jitter and shimmer, along with a higher HNR and a lower creak proportion in LG compared to HG. CPP and pitch level were higher in LG but showed little variation with age or language dominance. No clear effects on pitch span were observed. Overall, age was a slightly more reliable predictor than language dominance. Acoustic differences in voice quality were about equally detectable across the different speech tasks, while varying difficulty levels had minimal impact. The variation in vocal fold vibratory patterns suggests that younger and more HG-dominant speakers experienced greater cognitive load when speaking LG. Given that increased cognitive load may negatively impact language usage, voice analysis opens up new possibilities for evaluating the future prospects of endangered languages.