Förfarande och kretsanordning för förbättring av upplösningsnoggrannhe ten av digitala signaler

摘要

In digital signals transformed into a frequency domain, inaccuracies occur in the structure of the reproduced video signals due to rounding errors in the calculation of the coefficients with a shortened number of digits. The invention provides a remedy in that the range of numbers of the direct component is transferred into the range of numbers of the alternating components and thus a greater range of numbers is available for the alternating components for the common processing of the direct and alternating components. The invention can be used in digital systems with block-by-block processing in which a rounding of the spectral coefficients is carried out due to a limited number of bits in the calculation.