19.02.2013 Views

2010318163752818

2010318163752818

2010318163752818

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Short block:<br />

N_FR 128/120 MDCT block size<br />

Table 4.141 – Parameter table for the enhance and enhance_960 mode<br />

ISO/IEC 14496-3:2005(E)<br />

Parameter<br />

Common:<br />

Value Meaning<br />

AMP_MAX 8000 Maximum amplitude in the mu-law coding of the frame gain<br />

AMP_NM 1024 Normalized amplitude of flattened coefficients<br />

FW_ALF_STEP 0.5 MA prediction coefficient for the envelope coding<br />

MA_NP 1 MA prediction order (LSP coding)<br />

MU 100 Parameter mu in the mu-law coding of the frame power<br />

N_PR 20 LPC order<br />

NUM_STEP 256 Number of gain-quantization steps<br />

STEP AMP_MAX / (NUM_STEP- Step width of gain-quantization<br />

1) = 31.37<br />

SUB_AMP_MAX 6000 Maximum of subframe-gain to frame-gain ratio<br />

SUB_AMP_NM 1024 Normalized subframe amplitude of flattened coefficients<br />

SUB_GAIN_BIT 4 Number of bits for the subframe gain coding<br />

SUB_NUM_STEP 16 Number of sub-gain-quantization steps<br />

SUB_STEP SUB_AMP_MAX/ Step width of sub-gain-quantization<br />

(SUB_NUM_STEP-1)= 400.0<br />

Long block:<br />

FW_CB_LEN 7 Envelope code vector length<br />

FW_N_BIT 6 Envelope code bits<br />

FW_N_DIV 7 Number of division in the envelope coding<br />

N_CRB 42 Number of bark-scale subbands<br />

N_FR 1024/960 MDCT block size<br />

Short block:<br />

N_FR 128/120 MDCT block size<br />

4.6.11 Filterbank and block switching<br />

(Similar to ISO/IEC 13818-7)<br />

4.6.11.1 Tool description<br />

Table 4.142 – Values of isp[]<br />

split_num isp[split_num]<br />

0 0<br />

1 5<br />

2 14<br />

3 20<br />

The time/frequency representation of the signal is mapped onto the time domain by feeding it into the filterbank<br />

module. This module consists of an inverse modified discrete cosine transform (IMDCT), and a window and an<br />

overlap-add function. In order to adapt the time/frequency resolution of the filterbank to the characteristics of the<br />

input signal, a block switching tool is also adopted. N represents the window length, where N is a function of the<br />

window_sequence (see subclauses 4.5.2.3.3). For each channel, the N/2 time-frequency values Xi,k are<br />

transformed into the N time domain values xi,n via the IMDCT. After applying the window function, for each channel,<br />

the first half of the zi,n sequence is added to the second half of the previous block windowed sequence z(i-1),n to<br />

reconstruct the output samples for each channel outi,n.<br />

4.6.11.2 Definitions<br />

window_sequence 2 bit indicating which window sequence (i.e. block size) is used.<br />

window_shape 1 bit indicating which window function is selected.<br />

© ISO/IEC 2005 — All rights reserved 171

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!