Data processing apparatus and method

There is described a process for generating a number representative of an analogue data source in which during enrolment a distinctive characteristic of the analogue data source is measured to obtain physical data. Part of the physical data is used to generate a physical value which is representative of the analogue source. An error correction algorithm is applied to the physical value to generate error correction data, which is then transformed, using another part of the physical data, to generate transform data. During subsequent regeneration of the physical value, the distinctive characteristic is re-measured to generate a new set of physical data, and a physical value is generated using the same part of the physical data as was used during enrolment. Error correction data is then generated by transforming the transform data, using an inverse transform to that used during enrolment, using the same part of the physical data as was used to transform the error correction data during enrolment. The regenerated error correction data is then used by the error correction algorithm to correct errors in the physical value representative of the analogue source. By using part of the physical data set to transform the error correction data and then storing the resulting transform data, the security of the original physical data is improved.