It is well known, the voice segments and coincident data packets are not equally valued and significant for decoding and comprehension of speech signal. Some lost segments may only slightly worsen audible quality, while others cause strong distortion of the speech signals. Despite this, the feature of different importance of different voice segments in current generation of digital voice transmission systems is not fully used. There is a fundamental problem with discrimination of different importance and value of voice frames. In this paper the concept “of value of voice frame” is introduced, the metric and means for evaluation and measurement of voice frame value are proposed and also results of the measurements of voice frames value are presented.