Using an optimization technique, we derived subsurface properties of coastal and oceanic waters from measured remote-sensing reflectance spectra. These data included both optically deep and shallow environments. The measured reflectance covered a spectral range from 400 to 800 nm. The inversions used data from each 5-, 10-, and 20-nm contiguous bands, including Sea-viewing Wide Field-of-view Sensor (SeaWiFS), moderate-resolution imaging spectrometer (MODIS), and a self-defined medium-resolution imaging spectrometer (MERIS) channels, respectively. This study is designed to evaluate the influence of spectral resolution and channel placement on the accuracy of remote-sensing retrievals and to provide guidance for future sensor design. From the results of this study, we found the following: (1) use of 10-nm-wide contiguous channels provides almost identical results as found for 5-nm contiguous channels; (2) use of 20-nm contiguous channels and MERIS provides comparable results with those with 5-nm contiguous channels for deep waters, but use of contiguous 20-nm channels perform better than MERIS for optically shallow waters; and (3) SeaWiFS or MODIS channels work fine for deep, clearer waters (total absorption coefficient at 440 nm < 0.3 m(-1)), but introduce more errors in bathymetry retrievals for optically shallow waters. The inclusion of the 645-nm MODIS land band in its channel set improves inversion returns for both deep and shallow waters.