After we have carried out what we have just expounded, that is what has been required by the given detailed list, we have considered a further question; it is possible to be in the condition of reconstruct a potentially optimal hearing line except for a Df. This eventuality is shown in fig. 6:

Figure 6:           Possible non optimal reconstruction of a hearing line.

What has just been illustrated is an extreme case; however it could be possible to obtain better results positioning in a different way the equalization bands respect to the signal spectrum. For this reason, after the program has been completed, we thought to intervene to allow the relative shift of the signal spectrum towards the filter bank, carring out in this way the scheme illustrated in the following figure:

Figure 7:           Block scheme of the evoluted system.

With reference also to fig. 6 we make note that a fine variation for Df has been admitted: if on one hand this implies a distortion in the output signal, on the other it allows a wider range of adjustments to get near to a situation of optimal hearing and it's important in consideration of the final user.
This evolution makes the program work with complex data with a consequent variation of the data structure and an increase of the computational time; the insertion of the routine that achieves the modulation didn't give great problems thanks also to the agile and modular structure of the program.
We finally underline, concerning the modulation routine, that it has to consider the delay the system introduces to demodulate correctly in the output block. 

[ Index |  Previous page | Literature]