An audio signal with maximum amplitude 1 V (i.e. it may have values in the range -1 V to +1 V) is quantised using 8-bit nonlinear quantisation and uses all 255 available levels (so any of the 255 possible 8-bit output words might be generated). Suppose we halve the signal amplitude, making no change to the quantiser. Which of the following statements most accurately tells us the number N of 8-bit words that might be output now?

An audio signal with maximum amplitude 1 V (i.e. it may have values in the range -1 V to +1 V) is quantised using 8-bit linear quantisation and uses all 255 available levels (so any of the 255 possible 8-bit output words might be generated). Suppose we halve the signal amplitude, making no change to the quantiser. Which of the following statements most accurately tells us the number N of 8-bit words that might be output now?

In Section 6(b) of the laboratory exercise, you measured the SNR when the sample speech clip 12secSpeechSample - Speeches That Changed The World - Intro.wav  (sampled at 44.1 kHz) was quantised for different values of "m". What was the SNR for m=3? (Give your answer in dB accurate to 1 decimal place, but do not write the units - just put the number.)

In part 5(f) of the laboratory exercise, you measured the SNR when a sine wave test signal (sampled at 8 kHz) was quantised for different values of "m". What was the SNR for m=2? (Give your answer in dB accurate to 1 decimal place, but do not write the units - just put the number.)

In Section 6(b) of the laboratory exercise, you measured the SNR when the sample music clip ShortClip_Shine On.wav  (sampled at 44.1 kHz) was quantised for different values of "m". What was the SNR for m=4? (Give your answer in dB accurate to 1 decimal place, but do not write the units - just put the number.)

When digitising an audio signal such as speech or music that contains a range of frequency components, the sampling rate must always be chosen to satisfy the following condition(s):

Suppose someone from another lab group comes to you and says they're not sure that they are calculating SNR correctly for Section 5(d). They tell you the value of SNR they obtained when they thought the quantisation noise was just too small to hear. Most of the following answers suggest that there is something wrong with their experimental procedure or their calculations. However, ONE answer is pretty close to what you would expect and therefore suggests that they are on track. Which one?

The SNR is the ratio of average signal power (obtained from the original signal vector) to the average noise power (obtained from the difference between the original signal vector and the quantised signal vector). Is the SNR a vector or scalar quantity?