dsp.VectorQuantizerEncoder

Description

The VectorQuantizerEncoder object performs vector quantization encoding. The object finds the nearest codeword by computing a distortion based on Euclidean or weighted Euclidean distance.

To perform vector quantization encoding:

Define and set up your vector quantizer encoder. See Construction.
Call step to perform the quantization encoding according to the properties of dsp.VectorQuantizerEncoder. The behavior of step is specific to each object in the toolbox.

Note: Starting in R2016b, instead of using the step method to perform the operation defined by the System object™, you can call the object with arguments, as if it were a function. For example, y = step(obj,x) and y = obj(x) perform equivalent operations.

Construction

H = dsp.VectorQuantizerEncoder returns a vector quantizer encoder System object, H. This object finds a zero-based index of the nearest codeword for each given input column vector.

H = dsp.VectorQuantizerEncoder('PropertyName',PropertyValue,...) returns a vector quantizer encoder System object, H, with each specified property set to the specified value.

Properties

`CodebookSource`	Source of codebook values Specify how to determine the codebook values as `Property` or `Input port`. The default is `Property`.
`Codebook`	Codebook Specify the codebook to which the input column vector or matrix is compared, as a k-by-N matrix. Each column of the codebook matrix is a codeword, and each codeword corresponds to an index value. The codeword vectors must have the same number of rows as the input. The first codeword vector corresponds to an index value of `0`, the second codeword vector corresponds to an index value of `1`, and so on. This property applies when you set the `CodebookSource` property to `Property`. The default is: $[\begin{array}{l} 1 .5 13 .3 136 .4 6 .8 \\ 2 .5 14 .3 137 .4 7 .8 \\ 3 .5 15 .3 138 .4 8 .8 \end{array}]$ This property is tunable.
`DistortionMeasure`	Distortion calculation method Specify how to calculate the distortion as `Squared error` or `Weighted squared error`. If you set this property to `Squared error`, the object calculates the distortion by evaluating the Euclidean distance between the input column vector and each codeword in the codebook. If you set this property to `Weighted squared error`, the object calculates the distortion by evaluating a weighted Euclidean distance using a weighting factor to emphasize or deemphasize certain input values. The default is `Squared error`.
`WeightsSource`	Source of weighting factor Specify how to determine weighting factor as `Property` or `Input port`. This property applies when you set the `DistortionMeasure` property to `Weighted squared error`. The default is `Property`.
`Weights`	Weighting factor Specify the weighting factor as a vector of length equal to the number of rows of the input. This property applies when you set the `DistortionMeasure` property to `Weighted squared error` and `WeightsSource` property is `Property`. The default is [1 1 1]. This property is tunable.
`TiebreakerRule`	Behavior when input column vector is equidistant from two codewords. Specify whether to represent the input column vector by the lower index valued codeword or higher indexed valued codeword when an input column vector is equidistant from two codewords. You can set this property to `Choose the lower index` or `Choose the higher index`. The default is `Choose the lower index`.
`CodewordOutputPort`	Enable output of codeword value Set this property to `true` to output the codeword vectors nearest to the input column vectors. The default is `false`.
`QuantizationErrorOutputPort`	Enable output of quantization error Set this property to `true` to output the quantization error value that results when the object represents the input column vector by the nearest codeword. The default is `false`.
`OutputIndexDataType`	Data type of index output Specify the data type of the index output as: `int8`, `uint8`, `int16`, `uint16`, `int32`, `uint32`. The default is `int32`.

Fixed-Point Properties

Methods

clone	Create vector quantizer encoder object with same property values
getNumInputs	Number of expected inputs to step method
getNumOutputs	Number of outputs of step method
isLocked	Locked status for input attributes and nontunable properties
release	Allow property value and input characteristics changes
step	Perform vector quantization encoding

Examples

expand all

Find Indices of Nearest Codewords

Open Script

Note: This example runs only in R2016b or later. If you are using an earlier release, replace each call to the function with the equivalent step syntax. For example, myObject(x) becomes step(myObject,x).

Find the indices of nearest codewords based on Euclidean distances.

hvqenc = dsp.VectorQuantizerEncoder(...
    'Codebook', [-1 -1 1 1;1 -1 -1 1], ...
    'CodewordOutputPort', true, ...
    'QuantizationErrorOutputPort', true, ...
    'OutputIndexDataType', 'uint8');

Generate an input signal with some additive noise

x = sign(rand(2,40)-0.5) + 0.1*randn(2,40);
[ind, cw, err] = hvqenc(x);
plot(cw(1,:), cw(2,:), 'rO', x(1,:), x(2,:), 'g.');
legend('Quantized', 'Inputs', 'location', 'best');

Algorithms

This object implements the algorithm, inputs, and outputs described on the Vector Quantizer Encoder block reference page. The object properties correspond to the block parameters.

`RoundingMethod`	Rounding method for fixed-point operations Specify the rounding method as `Ceiling`, `Convergent`, `Floor`, `Nearest`, `Round`, `Simplest` or `Zero`. The default is `Floor`.
`OverflowAction`	Overflow action for fixed-point operations Specify the overflow action as `Wrap` or `Saturate`. The default is `Wrap`.
`ProductDataType`	Product word and fraction lengths Specify the product fixed-point data type as `Same as input` or `Custom`. The default is `Same as input`.
`CustomProductDataType`	Product word and fraction lengths Specify the product fixed-point type as a scaled `numerictype` object with a `Signedness` of `Auto`. This property applies when you set the `ProductDataType` property to `Custom`.
`AccumulatorDataType`	Accumulator word and fraction lengths Specify the accumulator fixed-point data type as `Same as input`, or `Custom`. The default is `Same as product`.
`CustomAccumulatorDataType`	Accumulator word and fraction lengths Specify the accumulator fixed-point type as a scaled `numerictype` object with a `Signedness` of `Auto`. This property applies when you set the `AccumulatorDataType` property to`Custom`.

Documentation

dsp.VectorQuantizerEncoder System object

Description

Construction

Properties

Methods

Examples

Find Indices of Nearest Codewords

Algorithms

See Also

Introduced in R2012a

DSP System Toolbox Documentation

Other Documentation

Support