Continuous - can take any value on a scale.

Discrete - can only take particular values.

1. Analogue signals can take any values on a scale.
2. Analogue signals can change as frequently as required.

1. Digital signals can only take one of a specified range of values.
2. Digital signals can only change at specified intervals.

1. The device measures the amplitude of the analogue wave at regular time intervals to produce samples.
2. Each sample is stored in binary and a sequence a samples represents sound.

1. The device reads the sequence of binary samples and outputs an alternating, analogue, electrical current.
2. A speaker dome will move in/ out depending on the amplitude of this signal.

The number of samples measured per second in Hz.

The number of bits used to record each sample.

1. Stereo * length in seconds * sample rate in Hz * sample resolution in bits = file size in bits.

2 channels ∗ 60 seconds ∗ 44,000 Hz ∗ 24 bits = 126,720,000bits

2. Convert to MB

bits → bytes → kb → mb

1. 126,720,000 bits / 8 = 15,840,000 bytes
2. 15,840,000 / 1,000,000 = 15.84 MB

File size will increase, but the quality of audio will be higher, that is to say more accurate.

Sampling rate must be at least twice the highest frequency of the sound, in order for an accurate binary representation.

MIDI stores sound as a series of event messagesEach event message holds data as a series of instructions to reproduce that sound.

1. Duration of a note.
2. How loud a note is.
3. The pitch of a note.
4. The octave within which the note is to be played.

1. Easy manipulation: event messages can be played by many instruments.
1. event messages can be transposed to different instruments.
2. The duration and timing of notes can be altered.
2. MIDI files are often smaller than sampled audio files.
3. MIDI files are a lossless format, so no information is lost.

1. MIDI music can sound less realistic and "faked" which listeners might not appreciate.
2. MIDI music cannot accurately represent human speech.

Images are series of ordered picture elements, known as pixels.

picture element, the smallest identifiable area of an raster image, containing a binary value representing a single colour.

The size of the image in pixels or The Pixels Per Inch (PPI).

Image resolution = height in pixels * width in pixels.

1. The number of bits per pixel used to store colour information.
2. A measure of the range of colours which are present in the image.

Bit depth = log(256) / log(2) = 8 bits.

1. Use formula: log(colourRange) / log(2) = colour or bit depth.

log(8) / log(2) = 3 bits

2. Use formula: height * width * colour depth + metadata = size in bits.

100 *250 * 3 + 32 = 75,032 bits

3. Convert to KB
1. 75,032 bits / 8 = 9379 bytes
2. 9379 bytes / 1000 = 9.379 KB

1. Resolution
2. Colour depth
3. Date created
4. Geolocation co-ordinates
5. Author

A list of geometric objects and shapes and their corresponding properties.

1. Stroke width
2. Stroke colour
3. Fill colour
4. Start position, end position

1. Can be scaled up or down (larger, or smaller) without becoming blurry or pixelated, because objects are simply redrawn.
2. Often uses less storage space than bitmapped graphics.

1. Fine details such as strands of hair are more accurately represented with pixels than geometric shapes.
2. Modification of raster images is more precise than geometric shapes, so unsuitable for applications such as photography.

1. Reduces file size.

Meaning more data can be stored per unit of storage space.

2. More economical

Therefore...

1. Speeds of transmission across a network are faster.
2. More data can be transferred per unit of time.

Some information is lost: such as reducing sample resolution or the size of a image in order to decrease file size.

No information is lost, patterns in data are identified and made concise.

1. Lossy compression is faster to execute, so only lossy is suitable for network transmissions which require a small latency (such as phone call).
2. Can in some cases result in smaller file size than lossless.

1. As some information is lost, the accuracy of the digital representation is reduced.
2. This means ugly artifacts in images and a perceptible loss of quality.

No information is lost, no reduction in quality of image or sound.

1. Lossless compression algorithms are often slower.
2. Lossless compression algorithms may not in all cases produce a great reduction in file size.
3. There is a limit to how much a file can be compressed.

Run Length Encoding (RLE), Dictionary

1. A series of contiguous repeated pixels are replaced with one pixel value and a the length of repetitions.
2. Data is stored as a series of these elementary units.

1. A dictionary is used to store large patterns of repeating data.
2. The file references the dictionary so such patterns are only stored once.
3. The included dictionary may increase file size.

1. Best: Many sets of repeating sequential data.
2. Worst: No repeating sets of sequential data at all.

The transformation of data into another form to prevent unauthorised parties from understanding the message.

The algorithm used to encrypt data/ transform data into another form.

The original data or message to be sent and understood by authorised users.

The encrypted data to be transmitted.

This is a substitution cipher, where each character of the alphabet is shifted to the left or right by a set value.

1. Align the characters of the plaintext and the key.
2. Convert each character to binary.
3. Apply a logical xor operation on each of the two bit-patterns.
4. Convert the result back into a (ASCII or UTF8) characters

One-time pad.

1. OTP must be at least as long as the plaintext message.
2. OTP must only be used once.
3. OTP must be chosen uniformly at random.
4. OTP must be transmitted using other communication channels.

1. Given enough time all cipher algorithms can be broken by cryptanalytic attack.
2. Cryptanalytic attack relies on the frequency analysis.
3. Since computers cannot produce random numbers any key will eventually be found.
4. A cipher is more secure the longer it takes for cryptanalysis.
5. Some ciphers may take longer than age of the universe but can theoretically be cracked.

1. Since the key is chosen uniformly at random, the ciphertext data is also uniformly randomly distributed.
2. This means frequency analysis and cryptanalytic attacks will always fail mathematically.