From the hearing of the sounds filtered by using FIR and IIR filter from part 1 and part 2 respectively , basically it  sounds almost the same to me. The outcome of the filter  is a similar to the original sound but with a lower volume of sound. Both results from the FIR and IIR filter produce much softer volume than the original sound. That can be clearly detected by comparing the graph of original signal (time and frequency domain in magnitude) with the filtered signal graph (time and frequency domain in magnitude). The magitude has become smaller compare to original signal.

The FIR filter need more order in the experiment to get a better quality filtered sound compare to IIR filter (using the same criteria from the results of part 1) where it only needs 2 order but the quality of sound is almost the same as what FIR filter produced. Comparing the SNR after filtering for part 1 and 2 , the SNR value for part 1 is larger than part 2 meaning the noise is less in outcome of filter using FIR compare to IIR.

For more number of quantization bits, the filter output a better result or signal. Quantization process causes the signal to take
the values from a finite set of discrete values. The higher quantization bit , the more number of discrete levels to be
occupied by the signal and thus, it produces less quantization error.Therefore, the number of filter coefficient quantization bits will affect the filter’s performance. The filter’s performance is directly proportional to the number of quantization bits, which means the higher the quantiztion bits the better the better the filter outcomes(the SNR is improving) SNR(8-bit)<SNR(12-bit)<SNR(16-bit).

Following are the discussion of FIR or IIR filter we need to take into consideration in selection the type of filter to use:
FIR possess following advantages : it is always stable , realizeable , and can always be designed to have a linear phase response. While IIR also has its own advantage , i.e : it can achieve sharper transitions between passband and stopband than FIR filters of the same order. Disadvantages is its stability is not guranteed, it produce a non-linear phase response(especially at band edge) ,
The effects of using a limited number of bits to implement filters such as roundoff noise and coefficient quatization errors are much less severe in FIR than in IIR. It is easier to syntesize filters of arbitary frequency responses using FIR. However , analogue filters can  be readily transformed into equivalent IIR digital filter meeting similar specifications, this is not possible for FIR as they have no analogue counterpart.
Basically the process of designing the FIR and IIR are almost the same.

IIR filter is usually more efficient than the corresponding FIR filter because it uses less order to produce the same magnitude of specifications. IIR is preferable because of its lower computational requirement , since the linearlity of the phase response is not an issue.While FIR is prefered when the system response is important point. IIR produce sharp cutoff filters and high throughput.