1. A quantizer that adaptively determines a value of an optimum threshold for quantizing an analog signal to minimize the overall quantizing error rate by observing amplitude levels of said analog signal comprising:
- a histogram generator that generates a histogram, said histogram comprising a plurality of histogram values representing said amplitude levels over a predetermined time period;
- an optimum reconstruction level generator that uses said histogram values to generate optimum reconstruction levels by determining mean squared errors; and
- an optimum decision threshold generator that uses said optimum reconstruction levels to generated optimum decision thresholds by determining the average of said optimum reconstruction levels.
2. The quantizer of claim 1 wherein said histogram generator comprises:
- a slicer that generates signals having amplitude values representative of said amplitude levels of said analog signal at discrete times over said predetermined time period;
- a memory that stores histogram values, each of said histogram values representing the number of occurrences of each amplitude value of said signals generated by said slicer over said predetermined time to produce said histograms.
3. The quantizer of claim 2 wherein said optimum reconstruction level generator comprises:
- a multiplier that multiplies each of said histograms by each of a series of said amplitude values of said amplitude levels for each said histogram to produce a series of first results;
- a first accumulator that accumulates said first results to produce first accumulated result;
- a second accumulator that accumulates said number of accumulated occurrences of each value of said signal strengths for each histogram to produce a second accumulated result; and
- a divider that divides said first accumulated result by said second accumulated result for each histogram to produce an optimum reconstruction level signal.
4. The quantizer of claim 1 wherein said optimum reconstruction level generator comprises:
- a multiplier that multiplies each of said histogram values of said histogram by index levels of said histogram to produce a series of first results;
- a first accumulator that accumulates said series of first results to produce a first accumulated result;
- a second accumulator that accumulates said histogram values to produce a second accumulated result; and,
- a divider that divides said first accumulated result by said second accumulated result to produce an optimum reconstruction level signal.
5. The quantizer of claim 1 wherein said optimum decision threshold generator comprises:
- an adder that adds adjacent optimum reconstruction levels to produce an added adjacent optimum reconstruction level value;
- a divider that divides said added adjacent optimum reconstruction level values by two to determine an average of said adjacent optimum reconstruction levels.
6. A method of quantizing an analog signal using an optimum decision threshold comprising the steps of:
- slicing said analog signal into a series of discrete signals having indices that are representative of the amplitude of said analog signal;
- generating a histogram from said indices indicative of the number of discrete signals for each of said indices;
- generating optimum reconstruction levels by using said histogram to determine minimum mean squared errors; and
- generating said optimum decision thresholds by using said optimum reconstruction levels to determine the average of said optimum reconstruction levels.
7. The method of claim 6 wherein said step of generating said histogram comprises:
- accumulating a series of histogram values indicative of the number of said discrete signals having the same index;
- storing said series of histogram values to form a histogram.
8. The method of claim 6 wherein said step of generating optimum reconstruction levels comprising the steps of:
- multiplying each of said series of histogram values by its corresponding index to produce a series of first results;
- accumulating said series of first results to produce a first accumulated result;
- accumulating said series of histogram values to produce a second accumulated result; and
- dividing said first accumulated result by said second accumulated result to produce an optimum reconstruction level.
9. The method of claim 6 wherein said step of generating said optimum decision thresholds comprises the steps of:
- adding adjacent optimum reconstruction levels to produce an added adjacent optimum reconstruction level value; and
- dividing said added adjacent optimum reconstruction level values by two to determine an average of said adjacent optimum reconstruction levels to produce said optimum decision threshold.
10. A quantizer for quantizing an analog signal using an optimum decision threshold comprising:
- means for slicing an analog signal into a series of discrete signals having indices that are representative of the amplitude of said series of discrete signals;
- means for generating a histogram from said indices;
- means for generating optimum reconstruction levels by using said histogram to determine minimum mean squared errors; and
- means for generating optimum decision thresholds by using said optimum reconstruction levels to determine an average of said optimum reconstruction levels.
