Blind deformulation is an importantstake for several industries.This work was motivated by the identification and quantification ofcontaminants originated from food packaging systems. Many substancesoriginating from plastic materials are indeed suspected to be endocrinedisruptors but remain chiefly difficult to separate with spectroscopictechniques. We propose a tailored two-scale pursuit methodology toidentify and quantify an arbitrary number of substances from the 1H NMR spectrum of the mixture. Identified substances are includedwithin a library of spectra and can be combined with undocumentedones. To preserve the initial resolution of NMR spectra, peak linesare spanned onto Gaussian kernels so that they can be identified,even when the positions and shapes of multiplets in the mixture aremodified within tolerance ranges or when multiplets are overlapping.The deconvolution procedure starts with a crude pairwise search tobuild a list of likely substances, which is subsequently expandedas nested scenarios. Scenarios are built according to the risk ofconfusing similar substances. Quantification is carried out on a preferencelist of substances selected as in a voting system. Using a primarylibrary of 52 substances (corresponding to 279 multiplets and 5620lines), the reliability and robustness of the method were tested extensivelyin numerical experiments and by performing the brute-force deformulationof five processed common thermoplastics. [ABSTRACT FROM AUTHOR]