That difference in loudness between voice and effects happens often when the stereo material is derived by a badly mixed down 5+1 material. Having control of the 6 channels levels before mixdown may help to eliminate it, otherwise some sort of compression/limiting is necessary. Many TVs have the option to set the audio output level to automatic, which doesn't mean the volume can't be changed but that around a given volume the dynamics are less extreme so that faint sounds are amplified and loud ones are attenuated.