Answer
When encoding information, approximations are made. In the case of numeric data, these approximations are compounded when computations are performed, which can lead to erroneous results. Approximations are not as critical in the cases of images and sound because the encoded data are normally only stored, transferred, and reproduced. If, however, images or sound were repeatedly reproduced, rerecorded, and then re-encoded, these approximations could compound and ultimately would lead to worthless data.
Work Step by Step
When encoding information, approximations are made. In the case of numeric data, these approximations are compounded when computations are performed, which can lead to erroneous results. Approximations are not as critical in the cases of images and sound because the encoded data are normally only stored, transferred, and reproduced. If, however, images or sound were repeatedly reproduced, rerecorded, and then re-encoded, these approximations could compound and ultimately would lead to worthless data.
