You need to clip each band by first converting the image to a grid, clipping the individual grids, and the converting them back to an image.
The example below uses an RGB (red-gree-blue) image. For images with more than three bands, such as Landsat TM 7-band images, you can use the same method for all the bands.
Arc: IMAGEGRID img.tif img
Grid: GRIDCLIP imgc1 imgclipc1 COVER clipcov
Grid: GRIDCLIP imgc2 imgclipc2 COVER clipcov
Grid: GRIDCLIP imgc3 imgclipc3 COVER clipcov
Grid: MAKESTACK imgclip LIST imgclipc1 imgclipc2 imgclipc3
Arc: GRIDIMAGE imgclip # imgclip.tif tiff