Loading patent…
AssigneeKoninklijke Philips Electronics N.V.
InventorsAlphons Antonius Maria Lambertus Bruekers, Adriaan Johannes Rijnberg
Estimating the term…
A filtering is proposed wherein first several filtering steps (I-III, IV-VI) are made. Each step (I-III-IV-IV) includes: generating a filtered signal (I, IV), rounding this filtered signal (II, V) to a certain accuracy and combining (III, VI) the rounded signal with the input signal, whereby a filter error is obtained. The filtered signals are combined (VII) to a first filtered signal that is rounded (VIII) and combined (IX) with the input signal, whereby a coded signal is obtained. This allows to adapt the rounding step (II, V) and the combining step (III, IV) in each filtering step to specific filter design. Thereby limit cycles are reduced, the first filtered signal is available for further processes and the rounding can be inverted in the reverse process. Further a filter method for the reverse process, filter devices (1, 10) able to perform these methods and devices (20, 30, 40, 50, 60) containing such devices are provided.
A filtering method, including the steps of:
A method as claimed in any one of the preceding claims, wherein the method is a prediction filtering method, wherein said step of generating (I-VII) a first filtered signal is a prediction filter step for generating a first prediction signal of said input signal, and the steps of generating a second filtered signal (I-III, IV-VI) are prediction filtering steps for generating a second prediction signal of said source signal based on preceding source signals.
A method as claimed in any of the preceding claims, wherein said second filtered signal is rounded to a second rounded signal with at least the accuracy of the least significant digit of said first rounded signal.
A method as claimed in any one of the preceding claims, wherein the step of predicting a second prediction signal (I, IV) is performed with a Finite Impulse Response method.
A method as claimed in any one of the preceding claims, in which filtering steps (I-III, IV-VI) the second filter error of a preceding repetition (I-III) is used as the source signal for a following filtering step (IV-VI), and said input signal is used as a source signal for a first filtering step (I-III) of generating a second filtered signal.
A method as claimed in any one of the preceding claims, wherein said input signal is a 1-bit signal, the step of rounding (VIII) said first prediction signal is a SIGN operation, and the step of calculating (IX) the first prediction error or coded signal is an Exclusive OR operation.
A filtering method, including the steps of:
A filter device (10), including:
A filter device (1) including:
Dependencies are inferred from the claim text.