When computing translations, the color model assumes the first channel is the reference channel, rather than using the channel that is designated.