|
Compression Results
Changing the bit values
Not a very high level of compression, but the bit number could be greatly reduced before there was any noticeable effect; 8 and 16 bits were highly similar, but at the 2-bit level much of the detail was lost from the music.
Changing the sampling frequency
Little compression, but the sampling rate could be reduced all the way from 44 kHz to 5kHz before there was a noticeable effect; at this frequency, the sound was nasal and had lost a lot of 'depth'.
Quantization
Not a very high level of compression due to elimination of only a limited number of frequencies.
However, the sound quality of the compressed signal is remarkably good., as only the frequencies insignificant to the overall waveform have been removed. However, there was quite a lot of audible static in the modified signal.
Removing a range of frequencies
This produced a very good compression (up to 13% of the signal was at one point removed). Some of the sound quality is lost (more than when using quantization), although given the percentage of the signal lost through filtering, the quality is surprisingly good.
Removing every second frequency
This compression is not quite as good as above, but there is very little audible signal degradation. However, if one listens to the signal very closely, there is a faint echo (almost undetectable).
Retaining only every fifth frequency
Very strange results -- the phrases of the music were echo-repeated and superimposed on the rest of the waveform. Evidently, a very bad form of filtering, although the resultant signal was highly compressed.
|
|
