A computer method analysing Frank VCG recorded on rest and on exertion, is presented. Electrical noises due to exercise are eliminated by averaging 8 complexes. The pattern-recognition of every wave is carried out by analysing a particular function obtained by translating the X, Y, Z leads on the line joining the T-P, P-Q, ST and subsequent T-P segment. The following parameters are evaluated: maximal and mean vector, linear, areolar and tangential (our original method) spatial velocity, half-area vector, area described by the spatial vector and oriented according to X, Y, Z polarity.