Non-invasive techniques to quantify metabolites in skeletal muscle provide unique insight into human physiology and enable the translation of research into practice. Proton magnetic resonance spectroscopy (1H-MRS) permits the assessment of several abundant muscle metabolites in vivo, including carnosine, a dipeptide composed of the amino acids histidine and beta-alanine. Muscle carnosine loading - accomplished by chronic oral beta-alanine supplementation - improves muscle function, exercise capacity and has pathophysiological relevance in multiple diseases. Moreover, the marked difference in carnosine content between fast-twitch and slow-twitch muscle fibers has rendered carnosine an attractive candidate to estimate human muscle fiber type composition. However, the quantification of carnosine using 1H-MRS requires technical expertise in order to obtain accurate and reproducible data. In this review, we describe the technical and physiological factors that impact the detection, analysis and quantification of carnosine in muscle using 1H-MRS. We discuss potential sources of error during the acquisition and pre-processing of the 1H-MRS spectra, and present best practices to enable the accurate, reliable and reproducible application of this technique.