1. Explain what is lossless source coding?
- A data compression technique, which reverts an exact copy of original file.
- Lossless Source Coding is used for compressing text files in modems.
- Lossless Source Coding is a building block for designing lossy compressors.
- Lossy compression is implemented for images, sound and video files for effective data compression.
- Many compression techniques have a lossless mode.
- The lossless source coding involves a sequence of fixed length symbols.
- Each of these symbols is easily manipulated independently.
2. Can you explain information theory plays an important role in field of compression
- Information Theory is about quantification of information.
- It is used in compressing data.
- Entropy is a key measure of information.
- It is expressed in terms of average number of bits that are required to store a message.
- Entropy is used to quantify the uncertainty which is a process in predicting the random variable values.
- Lossless data compression, Lossy data compression and channel coding are the fundamental topics of information theory.
3. Do you know instantaneous variable length codes?
- A code that maps source symbols into a set of variable number of bits.
- A VL code compresses the sources and decompresses with zero error.
- By implementing a right coding strategy, an identically distributed source might be compressed almost close to its entropy.
- This process is in contrast to fixed length coding methods.
- Examples of variable-length codes are Huffman coding, LempelZiv code.
4. Can you explain what is file compression and why is it necessary to compress files?
- File compression is a process to reduce the disk space to store that file.
- File compression enables data to be transferred quickly.
- Disk space needed on internet servers is reduced. This allows the servers to store more files / information with less disk space.
- File compression reduces the amount of time on internet to upload or download a file.
- Compression hides data so that not all computers can read the information stored.
- File compression is a mandatory preference for some of the internet servers to transfer files.
5. What is unique decipherability?
- Data symbols are encoded with coding schemes for fixed length codes.
- Every coding scheme has unique code.
- This unique encoded character ensures unambiguous.
- The encoded strings have fixed length.
- The fixed length codes are always uniquely decipherable.
6. Do you know non-binary Hoffman Codes?
- The non-binary Hoffman code elements are derived from an alphabet 'm' is > 2 letters.
- All the symbols ‘m' which occur least frequently will be having the same length.
- The lowest probability of the symbols ‘m' will differ only in the last position.
- The letters that combine have code words of the same length.
- The symbols that have lowest probability will have code words with long length.
7. Tell me what are the parameters that are used in silence compression?
- Silence compression is used in compressing sound files.
- It is equivalent to run length coding on normal data files.
- The parameters are:
1. A threshold value. It is a parameter that specifies, below which the compression can be considered as silence.
2. A silence code followed by a single byte. It indicates the numbers of consecutive silence codes are present.
3. To specify the start of a run of silence, which is a threshold.
8. Do you know what is rate distortion theory?
- Distortion theory is about trade-offs between the rate and distortion.
- It is applied for compression schemes.
- An average number of bits are utilized to represent each sample value.
- If the rate of bits is decreased it is known as increase in distortion.
- If the rate of bits to represent each value is increased it is known decrease in distortion.
- An entropy encoding algorithm.
- It uses variable length code table for encoding source symbol.
10. What is Shannon Fano Coding?
- It is used to construct a prefix code that is based on a set of symbols.
- It suboptimal. The lowest expected code word length will not be achieved.
11. What is Arithmetic Coding?
- A variable-length entropy encoding form.
- It is used for implementing loss less data compression.
- Fewer bits are occupied when frequently used characters are represented. More bits are occupied when not-so-frequently used characters are represented.
- Sub Band Coding(SBC) is a transform coding.
- A transform code can break a signal that may result in many different ‘frequency bands'.
- Every transform code is encoded independently.
- Most of the time, it is used for compressing audio and video signals.
13. What is Wavelet Based Compression?
- It is a process where a well defined temporal support for ‘wiggles' about X-axis.
- The inner-product of the input signal is multiplied with a set of ortho-normal basis functions.
- Then the coefficients are computed by this inner-product.
14. What is Vector Quantization?
- Vector Quantization has quantizes as inputs and outputs.
- The vector quantization results in a distortion rate.
- This rate is lower than the scalar quantization.
15. Explain the taxonomy of compression techniques?
- They are classified based on the requirements of reconstruction and compression of data.
- They are Lossy Compression and Lossless Compression.
16. What do you know about companded quantization?
- Companded Quantization maps the input through compressor function.
- This function expands the probability to the high level regions.
- These regions are close to the origin.
- These regions are compresses the corresponding lower probability regions that are away from the origin.
- The output of Companded Quantization is resulted by using uniform quantizer.
- An expander function is used to transform the quantized value.
17. Tell me about optimum prefix codes?
- Prefix coding is known as optimum coding.
- More frequently occurred symbols have shorter code words.
- Less frequently occurred symbols have longer code words.
- The less occurred frequently symbols will have equal length.
- Optimum prefix codes enhance the efficiency of data compression.
18. Explain forward adaptive quantization?
- The source output is divided into various blocks of data.
- Every block is analyzed prior to quantization.
- As per the block analyses, the quantizer parameters are set.
- These settings are transmitted later to the receiver.
- The transmitted settings are served as side information at the receiver end.
19. Explain composite source model?
- It is not simple to use a single model to describe the source in many applications.
- In these scenarios, a composite source model is used.
- Composite Source Model uses only one source.
- Only single source is activated at a given point of time.
- A prefix code is a code which does not require code word as a prefix to another code word.
- Huffman code is an example for Prefix Code.
21. Explain characteristics of a code?
- A code should be decodable.
- The code words are shorter than the letters which occur less frequently, has code word letters that occur more frequently.
22. Tell me two types of quantization errors?
1. Granular error
2. Slope over load error.
23. Explain two types of adaptive quantization?
1. Forward Adaptive Quantization
2. Backward Adaptive Quantization.
24. Explain offset in LZ77 approach?
- The sequence encoding in the look ahead buffer is encoded in this technique.
- The encoding id done by moving the encoder to a search pointer.
- The search pointer is through until a match to the first symbol is encountered.
- This symbol is available in the look ahead buffer.
- The actual distance between the pointer and the look ahead buffer is known as offset.
25. What is Progressive Transmission?
- A low resolution of an image is sent first.
- IT needs only few bits for the purpose of encoding.
- The image is then updated to the required fidelity.
- This is done by transmitting more information.
