DIGITAL AUDIO |
|
Frequency
A means of describing in which intervals (how often) a particular tone or sound induces periodic cycles.
We can hear these cycles as sound waves, as they exist in the Natural world, and if they are within our frequency range of hearing.
Frequency is measured in Hertz (Hz), the common term for cycles per second. |
|
The frequency of human hearing is generally considered 20 Hz to 20,000Hz (20Khz) |
|
A Digital audio system consists of numbers that represents a waveform. |
|
|
|
Waveform: A visual representation of a sound, representing the changing amplitude of the region over time.
Notice how the sound evolves over time in a cyclic way. |
|
|
|
|
|
Amplitude: A term used by audio engineers to describe relative levels.
(levels meaning volume)
Sampling rate: Defines an audio file’s upper frequency limit.
|
How many intervals per time unit.
The CD format uses 44,100 samples per second. |
THE DIGITAL AUDIO PROCESS. Transforming an Analog signal to Digital, through Discrete Time sampling.
A continuous analog audio signal (represented by a waveform here) is SAMPLED at regular intervals, and its AMPLITUDE is determined at each point. Based on the amplitude each sample is given a numerical value, which constitutes the DIGITAL AUDIO FILE. |
|
|
|
|
Dynamic Range: The variation between a signal’s quiet and loud passages, measured in Decibels (dB).
MORE
A symphony Orchestra is capable of great dynamic variation.
At its loudest may be able to produce 110 to 115 dB.
At its quietest moments, the most pianissimo flute passage, the sound might measure only 50dB.
Given this span, we could say the orchestra has a dynamic range of roughly 60 to 65dB.
Compare the dynamic range of an orchestra to the dynamic range of a cellphone, that has a very limited amount of values for representing the whole sound waveform. |
|
|
|
|
|
|
|
Bit Rate: Samples are coded using BINARY BITS. The more bits in each digital “word” the greater the total number of words (discrete values) available to code each sample, and so the greater of the precision of that word. |
|
|
|
2 bit |
4 values or possible levels |
|
3 bit |
8 values or possible levels |
4 bit |
16 values or possible levels |
8 bit |
256 values or possible levels |
16 bit |
65,536 values or possible levels |
|
|
Values in a 2 bit word: |
11 |
10 |
00 |
01 |
|
|
|
|
|
|
Bandwidth: The range between the lower and upper frequencies, also the maximum amount of digital data capable of transmission or storage. |
|
|
|
To normalize : Process used to boost the sound signal.
The file is scanned for the loudest peak.
That peak’s amplitude is brought to a 100% full scale.
The rest of the volume levels are raised accordingly. |
In Red the normalized
waveform.
In Blue the original Amplitude. |
|
|
|
|
TO EQUALIZE: The Process of cutting or boosting one or more areas of a particular sound’s frequency range. EQ.
You select a certain section of the sound (high frequencies, mid values, low range) and you boost their amplitude, so you can hear them better.
You can also use this technique to isolate certain undesirable frequencies (cars passing by, an alarm or telephone ringing, the washing machine running in the back room...) and cut them of or lower their levels so they become less noticeable.
|
|
|
|
Remember that our hearing process is a highly sophisticated system that happens for most of us in a totally unconscious level.
We think very little about sound from a rational point of view, but we can extract very sophisticated information about our surrounding Space through the different qualities of the perceived sound.
when you modify the different frequencies of a sound you need to pay close attention to the Spatial result of your experiment: How is that sound now being perceived? Am I creating an unnatural situation, and if I am, Is this "creative license" producing an aesthetic experience or is it being perceived as an error?
This is a very delicate balance that you need to keep in place, and it defines the difference between an art-piece and a failed experiment.
|
|
|
|
|
|
|
|
|