LSFQuant_G729

Performs quantization of LSF coefficients.

Syntax

IppStatus ippsLSFQuant_G729_16s (const Ipp16s* pLSF, Ipp16s* pQuantLSFTable, Ipp16s* pQuantLSF, Ipp16s* quantIndex);

IppStatus ippsLSFQuant_G729B_16s (const Ipp16s* pLSF, Ipp16s* pQuantLSFTable, Ipp16s* pQuantLSF, Ipp16s* quantIndex);

IppStatus ippsLSFQuant_G729B_32f (const Ipp32f* pLSF, Ipp32f* pQuantLSFTable, Ipp32f* pQuantLSF, int* quantIndex);

Parameters

pLSF

Pointer to the LSF vector [10], in Q13 in the range [0,π] for 16s data.

pQuantLSFTable

Pointer to a matrix of 4 rows and 10 columns, containing previously quantized LSFs.

pQuantLSF

Pointer to the quantized LSF vector, in Q13 in the range [0,π ] for 16s data.

quantIndex

Pointer to the output combined codebook indices L0, L1, L2, and L3.

Description

These functions are declared in the ippsc.h file. The function ippsLSFQuant_G729_16s quantizes the difference between the input LSF coefficients and previously quantized LSF coefficients, using the switched Moving Average (MA) predictor. Quantization is performed using a two-stage Vector Quantizer (VQ): 10-dimensional VQ using codebook L1 (128 entries), and 10-bit split VQ (5 dimensions each) on two codebooks, L2 and L3 (32 entries each).

The function ippsLSFQuant_G729B_16s quantizes the LSF coefficients using a two-stage split VQ, (in 5 and 4 bits). The second order MA predictor used in the SID-LPC quantization procedure is calculated as a linear interpolation of the first and second MA predictors, with weight values 0.6 and 0.4, respectively. The first stage of quantization is the same as that in G.729, but only part of the quantization table (32 entries) is used. The second stage is different. No splitting is done and also only portion of the second table (16 entries) is used.

Return Values

ippStsNoErr

Indicates no error.

ippStsNullPtrErr

Indicates an error when one of the specified pointers is NULL.

ippStsLSFLow

Indicates a warning that the corresponding filter has low stability.

ippStsLSFHigh

Indicates a warning that the corresponding filter has high stability.

ippStsLSFLowAndHigh

Indicates a warning that the corresponding filter has both low and high stability.

Submit feedback on this help topic

Copyright © 2000 - 2011, Intel Corporation. All rights reserved.