Chapter 3: Levels, Dynamics, Loudness and Dither

Question 1

How does a digital system specify clipping?

Answer 1

When consecutive samples achieve the maximum quantisation value. Most commonly, 4 consecutive maximum value samples are read as an overload, but some meters allow users to specify the number.

Question 2

What can cause clipping when converting a clean digital signal to analogue?

Answer 2

Inter-sample peaks, when the wave of the signal peaks in between two recorded points and therefore is not observed by the digital meter.

Question 3

What is a ‘reconstruction filter’ used for?

Answer 3

It converts periodic sample levels back to a discrete signal, reducing artifacts to a minimum.

Question 4

Name one-way that inter-sample peaks could be prevented.

Answer 4

• Leave a lot of headroom in the mix.
• Use a mastering tool that prevents the peaks
• Use a metering tool to highlight the peaks.

Question 5

What is Crest Factor?

Answer 5

The ratio of difference between peak level and RMS (average perceived level). It ranges from about 3dB in a sine wave, to 10-20dB+ in music.

Question 6

Does the Crest factor increase or decrease when Compression is applied to an audio signal?

Answer 6

The Crest factor decreases because the Peaks in the signal are lowered.

Question 7

What is the (generally accepted) smallest volume change that the average person can perceive? (in dB)

Answer 7

1 dB, although this can actually vary from 0.1-5dB, depending on the listening material and the person.

Question 8

A 15dB boost in volume is perceived as twice as loud, true or false?

Answer 8

False, anywhere between 6-10dB is perceived as twice as loud.

Question 9

Briefly describe the ‘loudness wars’

Answer 9

The increase in use of compression, based on the belief that ‘loud is better’.

Question 10

Who created the K-System and to what hard/software does it apply?

Answer 10

Bob Katz, It applies to meters.

Question 11

What makes it unique as a meter?

Answer 11

It includes 0dB as a monitoring reference.

Question 12

What are the 3 different K-System scales? and what do the numbers represent?

Answer 12

K-20, K-14, K-12. The number represents how far below 0dBfs the 0dB reading is placed.

Question 13

What does MFIT stand for?

Answer 13

Mastering-For Itunes

Question 14

What is ‘normalising’?

Answer 14

Boosting the volume of a track to bring it’s highest peak to a level near 0 dBfs

Question 15

What could be an advantage of normalising all songs?

Answer 15

The use of Compression may be reduced and the ‘loudness wars’ may reach an end.

Question 16

What is a quantisation error?

Answer 16

When the signal amplitude being sampled falls between two ‘bits’, the amplitude cannot be re-constructed and is truncated.

Question 17

What process could overcome quantisation errors?

Answer 17

Dithering/Dither

Question 18

Describe dithering.

Answer 18

The addition of random broadband noise to a waveform to break up the ‘statistical determinability’ of the truncated wave. Although low level noise is added, this is often more desirable than the harmonic distortion of truncation.

Question 19

What is POW-r (#1, #2 and #3)?

Answer 19

A Dither plug-in and it’s 3 settings.

Question 20

In summary, What four topics can affect audio quality?

Answer 20

• Sample Rate/Bit-Rate and Dither
• Level and Dynamic Range
• File type
• MP3 Codec