There is a really good sauce that a local restaurant makes. I have had trouble recreating it, but have gotten somewhat close. I had the following idea:
I could take an "image" of the sauce's color spectrum under good light using a USB spectrometer, and compare it to the spectrum of my home made version. I could then slowly tweak the ingredients and resample the color spectrum of the home made sauce until it started converging and eventually matched the restaurant's.
What could prevent this from working well? Could this realistically outperform taste testing?