A group of algorithms and a computer program based on these algorithms are developed for comparing IR spectra of different origins and/or of different precisions. For this purpose the spectra to be compared matched in precision. Data with smaller wavelength intervals are reduced using different well-known methods, such as moving averages, least squares, and reciprocal exponential convolution. Later baseline correction and normalization procedures, simple noise elimination, and interference checking methods are applied, and the results are compared to decide on the appropriate method for this purpose. One of the most common comparison problems encountered is that the reference and the sample data are usually recorded by different spectrometers and with different scanning parameters. One frequently needs to compare his/her sample spectra with an atlas of IR spectra. Development of an instrument free comparison method therefore will be quite useful for many purposes.