quantize

Functions

int main(int argc, char *argv[])

quantize [ option ] [ infile ]

• -v double

  • absolute maximum value \((0 < V)\)

• -b int

  • number of bits \((1 \le B)\)

• -t int

  • quantization type

    • 0 mid-rise

    • 1 mid-tread

• -o int

  • output format

    • 0 non-negative integer

    • 1 integer

• infile str

  • double-type data sequence

• stdout

  • int-type quantized data sequence

echo -2 -1 0 1 2 | x2x +ad | quantize -b 2 -v 2 -t 0 | x2x +ia
# 0, 1, 2, 3, 3
echo -2 -1 0 1 2 | x2x +ad | quantize -b 2 -v 2 -t 1 | x2x +ia
# 0, 0, 1, 2, 2

Parameters:

• argc – [in] Number of arguments.

• argv – [in] Argument vector.

Returns:

0 on success, 1 on failure.

See also

dequantize

class UniformQuantization

Perform uniform quantization.

The input is a value and the output is the corresponding index. There are two types of quantization, i.e., mid-rise and mid-tread.

Public Types

enum QuantizationType

Quantization type.

Values:

enumerator kMidRise

enumerator kMidTread

enumerator kNumQuantizationTypes

Public Functions

UniformQuantization(double absolute_maximum_value, int num_bit, QuantizationType quantization_type)

Parameters:

• absolute_maximum_value – [in] Absolute maximum value.

• num_bit – [in] Number of quantization bits.

• quantization_type – [in] Quantization type.

inline double GetAbsoluteMaximumValue() const

Returns:

Absolute maximum value.

inline int GetNumBit() const

Returns:

Number of quantization bits.

inline QuantizationType GetQuantizationType() const

Returns:

Quantization type.

inline bool IsValid() const

Returns:

True if this object is valid.

inline int GetQuantizationLevels() const

Returns:

Quantization levels.

bool Run(double input, int *output) const

Parameters:

• input – [in] Input value.

• output – [out] Output index.

Returns:

True on success, false on failure.