Method and apparatus for non-linear signal processing with reduced upper harmonic components
AssigneeEastman Kodak Company
InventorRichard A. Sharman
Life of this patent
Estimating the term…
Abstract
In a sampled data system with limited bandwidth where the output of a table memory 34 is multiplied with the sampled data in a multiplier 36 to produce a non-linear function, the process ordinarily produces harmonics that fold back into the passband of the system, thus causing artifacts in the processed data. By filtering the sampled data in a low pass filter 30 before addressing the table memory 34, the non-linearity of the function is progressively reduced toward the upper band limit. This reduces the production of harmonics and the resultant artifacts.
Claims11
A method for the processing of sampled data, said method comprising the steps of:
- storing a non-linear function in a look-up table memory such that each table input generates a table output based upon the transfer characteristic of the non-linear function;
- low pass filtering the sampled data;
- addressing the input of the look-up table memory with the low pass filtered data;
- and
- multiplying the sampled data by the output of the look-up table memory.
Signal processing apparatus for implementing a non-linear function upon sampled data, said apparatus comprising:
- means for low pass filtering the sampled data to generate sampled low pass data;
- memory look-up means for converting the sampled low pass data into multiplier coefficients based upon the transfer characteristic of the non-linear function;
- and
- means for multiplying the sampled data by the multiplier coefficients to generate the output of the apparatus.
Dependencies are inferred from the claim text.