We want to clarify a few issues about using optical absorption spectra for characterizing carbon nanotubes. The common method of using optical density for nanotube counting is reasonable in the nearinfrared and visible spectral region, but fails for the mid and far infrared, where metallic tubes can be unambigously detected. By performing Kramers -Kronig transformation on wide-range transmittance spectra of free-standing films, the far-infrared spectral region creates a unique possibility for estimating the amount of metallic tubes in different nanotube networks. We want to clarify a few issues about using optical absorption spectra for characterizing carbon nanotubes. The common method of using optical density for nanotube counting is reasonable in the nearinfrared and visible spectral region, but fails for the mid and far infrared, where metallic tubes can be unambigously detected. By performing Kramers -Kronig transformation on wide-range transmittance spectra of free-standing films, the far-infrared spectral region creates a unique possibility for estimating the amount of metallic tubes in different nanotube networks.