/*
* psymodel.c
*
* Copyright (c) 1999-2000 Mark Taylor
* Copyright (c) 2001-2002 Naoki Shibata
* Copyright (c) 2000-2003 Takehiro Tominaga
* Copyright (c) 2000-2008 Robert Hegemann
* Copyright (c) 2000-2005 Gabriel Bouvigne
* Copyright (c) 2000-2005 Alexander Leidinger
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Library General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
* Boston, MA 02111-1307, USA.
*/
/* $Id: PsyModel.java,v 1.27 2011/05/24 20:48:06 kenchis Exp $ */
/*
PSYCHO ACOUSTICS
This routine computes the psycho acoustics, delayed by one granule.
Input: buffer of PCM data (1024 samples).
This window should be centered over the 576 sample granule window.
The routine will compute the psycho acoustics for
this granule, but return the psycho acoustics computed
for the *previous* granule. This is because the block
type of the previous granule can only be determined
after we have computed the psycho acoustics for the following
granule.
Output: maskings and energies for each scalefactor band.
block type, PE, and some correlation measures.
The PE is used by CBR modes to determine if extra bits
from the bit reservoir should be used. The correlation
measures are used to determine mid/side or regular stereo.
*/
/*
Notation:
barks: a non-linear frequency scale. Mapping from frequency to
barks is given by freq2bark()
scalefactor bands: The spectrum (frequencies) are broken into
SBMAX "scalefactor bands". Thes bands
are determined by the MPEG ISO spec. In
the noise shaping/quantization code, we allocate
bits among the partition bands to achieve the
best possible quality
partition bands: The spectrum is also broken into about
64 "partition bands". Each partition
band is about .34 barks wide. There are about 2-5
partition bands for each scalefactor band.
LAME computes all psycho acoustic information for each partition
band. Then at the end of the computations, this information
is mapped to scalefactor bands. The energy in each scalefactor
band is taken as the sum of the energy in all partition bands
which overlap the scalefactor band. The maskings can be computed
in the same way (and thus represent the average masking in that band)
or by taking the minmum value multiplied by the number of
partition bands used (which represents a minimum masking in that band).
*/
/*
The general outline is as follows:
1. compute the energy in each partition band
2. compute the tonality in each partition band
3. compute the strength of each partion band "masker"
4. compute the masking (via the spreading function applied to each masker)
5. Modifications for mid/side masking.
Each partition band is considiered a "masker". The strength
of the i'th masker in band j is given by:
s3(bark(i)-bark(j))*strength(i)
The strength of the masker is a function of the energy and tonality.
The more tonal, the less masking. LAME uses a simple linear formula
(controlled by NMT and TMN) which says the strength is given by the
energy divided by a linear function of the tonality.
*/
/*
s3() is the "spreading function". It is given by a formula
determined via listening tests.
The total masking in the j'th partition band is the sum over
all maskings i. It is thus given by the convolution of
the strength with s3(), the "spreading function."
masking(j) = sum_over_i s3(i-j)*strength(i) = s3 o strength
where "o" = convolution operator. s3 is given by a formula determined
via listening tests. It is normalized so that s3 o 1 = 1.
Note: instead of a simple convolution, LAME also has the
option of using "additive masking"
The most critical part is step 2, computing the tonality of each
partition band. LAME has two tonality estimators. The first
is based on the ISO spec, and measures how predictiable the
signal is over time. The more predictable, the more tonal.
The second measure is based on looking at the spectrum of
a single granule. The more peaky the spectrum, the more
tonal. By most indications, the latter approach is better.
Finally, in step 5, the maskings for the mid and side
channel are possibly increased. Under certain circumstances,
noise in the mid & side channels is assumed to also
be masked by strong maskers in the L or R channels.
Other data computed by the psy-model:
ms_ratio side-channel / mid-channel masking ratio (for previous granule)
ms_ratio_next side-channel / mid-channel masking ratio for this granule
percep_entropy[2] L and R values (prev granule) of PE - A measure of how
much pre-echo is in the previous granule
percep_entropy_MS[2] mid and side channel values (prev granule) of percep_entropy
energy[4] L,R,M,S energy in each channel, prev granule
blocktype_d[2] block type to use for previous granule
*/
package mp3;
import java.util.Arrays;
public class PsyModel {
private FFT fft = new FFT();
private static final float LOG10 = 2.30258509299404568402f;
private static final int rpelev = 2;
private static final int rpelev2 = 16;
private static final int rpelev_s = 2;
private static final int rpelev2_s = 16;
/* size of each partition band, in barks: */
private static final float DELBARK = .34f;
/* tuned for output level (sensitive to energy scale) */
private static final float VO_SCALE = (1.f / (14752 * 14752) / (Encoder.BLKSIZE / 2));
private static final float temporalmask_sustain_sec = 0.01f;
private static final float NS_PREECHO_ATT0 = 0.8f;
private static final float NS_PREECHO_ATT1 = 0.6f;
private static final float NS_PREECHO_ATT2 = 0.3f;
private static final float NS_MSFIX = 3.5f;
public static final float NSATTACKTHRE = 4.4f;
public static final int NSATTACKTHRE_S = 25;
private static final int NSFIRLEN = 21;
/* size of each partition band, in barks: */
private static final float LN_TO_LOG10 = 0.2302585093f;
private static final float NON_LINEAR_SCALE_ENERGY(float x) {
return x;
}
/**
*
* L3psycho_anal. Compute psycho acoustics.
*
* Data returned to the calling program must be delayed by one
* granule.
*
* This is done in two places.
* If we do not need to know the blocktype, the copying
* can be done here at the top of the program: we copy the data for
* the last granule (computed during the last call) before it is
* overwritten with the new data. It looks like this:
*
* 0. static psymodel_data
* 1. calling_program_data = psymodel_data
* 2. compute psymodel_data
*
* For data which needs to know the blocktype, the copying must be
* done at the end of this loop, and the old values must be saved:
*
* 0. static psymodel_data_old
* 1. compute psymodel_data
* 2. compute possible block type of this granule
* 3. compute final block type of previous granule based on #2.
* 4. calling_program_data = psymodel_data_old
* 5. psymodel_data_old = psymodel_data
* psycho_loudness_approx
* jd - 2001 mar 12
* in: energy - BLKSIZE/2 elements of frequency magnitudes ^ 2
* gfp - uses out_samplerate, ATHtype (also needed for ATHformula)
* returns: loudness^2 approximation, a positive value roughly tuned for a value
* of 1.0 for signals near clipping.
* notes: When calibrated, feeding this function binary white noise at sample
* values +32767 or -32768 should return values that approach 3.
* ATHformula is used to approximate an equal loudness curve.
* future: Data indicates that the shape of the equal loudness curve varies
* with intensity. This function might be improved by using an equal
* loudness curve shaped for typical playback levels (instead of the
* ATH, that is shaped for the threshold). A flexible realization might
* simply bend the existing ATH curve to achieve the desired shape.
* However, the potential gain may not be enough to justify an effort.
*
*/
private float psycho_loudness_approx(final float[] energy,
final LameInternalFlags gfc) {
float loudness_power = 0.0f;
/* apply weights to power in freq. bands */
for (int i = 0; i < Encoder.BLKSIZE / 2; ++i)
loudness_power += energy[i] * gfc.ATH.eql_w[i];
loudness_power *= VO_SCALE;
return loudness_power;
}
private void compute_ffts(final LameGlobalFlags gfp,
final float fftenergy[], final float fftenergy_s[][],
final float wsamp_l[][], final int wsamp_lPos,
final float wsamp_s[][][], final int wsamp_sPos, final int gr_out,
final int chn, final float[] buffer[], final int bufPos) {
final LameInternalFlags gfc = gfp.internal_flags;
if (chn < 2) {
fft.fft_long(gfc, wsamp_l[wsamp_lPos], chn, buffer, bufPos);
fft.fft_short(gfc, wsamp_s[wsamp_sPos], chn, buffer, bufPos);
}
/* FFT data for mid and side channel is derived from L & R */
else if (chn == 2) {
for (int j = Encoder.BLKSIZE - 1; j >= 0; --j) {
final float l = wsamp_l[wsamp_lPos + 0][j];
final float r = wsamp_l[wsamp_lPos + 1][j];
wsamp_l[wsamp_lPos + 0][j] = (l + r) * Util.SQRT2 * 0.5f;
wsamp_l[wsamp_lPos + 1][j] = (l - r) * Util.SQRT2 * 0.5f;
}
for (int b = 2; b >= 0; --b) {
for (int j = Encoder.BLKSIZE_s - 1; j >= 0; --j) {
final float l = wsamp_s[wsamp_sPos + 0][b][j];
final float r = wsamp_s[wsamp_sPos + 1][b][j];
wsamp_s[wsamp_sPos + 0][b][j] = (l + r) * Util.SQRT2 * 0.5f;
wsamp_s[wsamp_sPos + 1][b][j] = (l - r) * Util.SQRT2 * 0.5f;
}
}
}
/*********************************************************************
* compute energies
*********************************************************************/
fftenergy[0] = NON_LINEAR_SCALE_ENERGY(wsamp_l[wsamp_lPos + 0][0]);
fftenergy[0] *= fftenergy[0];
for (int j = Encoder.BLKSIZE / 2 - 1; j >= 0; --j) {
final float re = (wsamp_l[wsamp_lPos + 0])[Encoder.BLKSIZE / 2 - j];
final float im = (wsamp_l[wsamp_lPos + 0])[Encoder.BLKSIZE / 2 + j];
fftenergy[Encoder.BLKSIZE / 2 - j] = NON_LINEAR_SCALE_ENERGY((re
* re + im * im) * 0.5f);
}
for (int b = 2; b >= 0; --b) {
fftenergy_s[b][0] = (wsamp_s[wsamp_sPos + 0])[b][0];
fftenergy_s[b][0] *= fftenergy_s[b][0];
for (int j = Encoder.BLKSIZE_s / 2 - 1; j >= 0; --j) {
final float re = (wsamp_s[wsamp_sPos + 0])[b][Encoder.BLKSIZE_s
/ 2 - j];
final float im = (wsamp_s[wsamp_sPos + 0])[b][Encoder.BLKSIZE_s
/ 2 + j];
fftenergy_s[b][Encoder.BLKSIZE_s / 2 - j] = NON_LINEAR_SCALE_ENERGY((re
* re + im * im) * 0.5f);
}
}
/* total energy */
{
float totalenergy = 0.0f;
for (int j = 11; j < Encoder.HBLKSIZE; j++)
totalenergy += fftenergy[j];
gfc.tot_ener[chn] = totalenergy;
}
if (gfp.analysis) {
for (int j = 0; j < Encoder.HBLKSIZE; j++) {
gfc.pinfo.energy[gr_out][chn][j] = gfc.pinfo.energy_save[chn][j];
gfc.pinfo.energy_save[chn][j] = fftenergy[j];
}
gfc.pinfo.pe[gr_out][chn] = gfc.pe[chn];
}
/*********************************************************************
* compute loudness approximation (used for ATH auto-level adjustment)
*********************************************************************/
if (gfp.athaa_loudapprox == 2 && chn < 2) {
// no loudness for mid/side ch
gfc.loudness_sq[gr_out][chn] = gfc.loudness_sq_save[chn];
gfc.loudness_sq_save[chn] = psycho_loudness_approx(fftenergy, gfc);
}
}
/* mask_add optimization */
/* init the limit values used to avoid computing log in mask_add when it is not necessary */
/**
*
* For example, with i = 10*log10(m2/m1)/10*16 (= log10(m2/m1)*16)
*
* abs(i)>8 is equivalent (as i is an integer) to
* abs(i)>=9
* i>=9 || i<=-9
* equivalent to (as i is the biggest integer smaller than log10(m2/m1)*16
* or the smallest integer bigger than log10(m2/m1)*16 depending on the sign of log10(m2/m1)*16)
* log10(m2/m1)>=9/16 || log10(m2/m1)<=-9/16
* exp10 is strictly increasing thus this is equivalent to
* m2/m1 >= 10^(9/16) || m2/m1<=10^(-9/16) which are comparisons to constants
*
*/
/**
* as in if(i>8)
*/
private static final int I1LIMIT = 8;
/**
* as in if(i>24) . changed 23
*/
private static final int I2LIMIT = 23;
/**
* as in if(m<15)
*/
private static final int MLIMIT = 15;
private float ma_max_i1;
private float ma_max_i2;
private float ma_max_m;
/**
* This is the masking table:
* According to tonality, values are going from 0dB (TMN) to 9.3dB (NMT).
* After additive masking computation, 8dB are added, so final values are
* going from 8dB to 17.3dB
*
* pow(10, -0.0..-0.6)
*/
private static final float tab[] = { 1.0f, 0.79433f, 0.63096f, 0.63096f,
0.63096f, 0.63096f, 0.63096f, 0.25119f, 0.11749f };
private void init_mask_add_max_values() {
ma_max_i1 = (float) Math.pow(10, (I1LIMIT + 1) / 16.0);
ma_max_i2 = (float) Math.pow(10, (I2LIMIT + 1) / 16.0);
ma_max_m = (float) Math.pow(10, (MLIMIT) / 10.0);
}
private static final float table1[] = { 3.3246f * 3.3246f,
3.23837f * 3.23837f, 3.15437f * 3.15437f, 3.00412f * 3.00412f,
2.86103f * 2.86103f, 2.65407f * 2.65407f, 2.46209f * 2.46209f,
2.284f * 2.284f, 2.11879f * 2.11879f, 1.96552f * 1.96552f,
1.82335f * 1.82335f, 1.69146f * 1.69146f, 1.56911f * 1.56911f,
1.46658f * 1.46658f, 1.37074f * 1.37074f, 1.31036f * 1.31036f,
1.25264f * 1.25264f, 1.20648f * 1.20648f, 1.16203f * 1.16203f,
1.12765f * 1.12765f, 1.09428f * 1.09428f, 1.0659f * 1.0659f,
1.03826f * 1.03826f, 1.01895f * 1.01895f, 1 };
private static final float table2[] = { 1.33352f * 1.33352f,
1.35879f * 1.35879f, 1.38454f * 1.38454f, 1.39497f * 1.39497f,
1.40548f * 1.40548f, 1.3537f * 1.3537f, 1.30382f * 1.30382f,
1.22321f * 1.22321f, 1.14758f * 1.14758f, 1 };
private static final float table3[] = { 2.35364f * 2.35364f,
2.29259f * 2.29259f, 2.23313f * 2.23313f, 2.12675f * 2.12675f,
2.02545f * 2.02545f, 1.87894f * 1.87894f, 1.74303f * 1.74303f,
1.61695f * 1.61695f, 1.49999f * 1.49999f, 1.39148f * 1.39148f,
1.29083f * 1.29083f, 1.19746f * 1.19746f, 1.11084f * 1.11084f,
1.03826f * 1.03826f };
/**
* addition of simultaneous masking Naoki Shibata 2000/7
*/
private float mask_add(float m1, float m2, final int kk, final int b,
final LameInternalFlags gfc, final int shortblock) {
float ratio;
if (m2 > m1) {
if (m2 < (m1 * ma_max_i2))
ratio = m2 / m1;
else
return (m1 + m2);
} else {
if (m1 >= (m2 * ma_max_i2))
return (m1 + m2);
ratio = m1 / m2;
}
/* Should always be true, just checking */
assert (m1 >= 0);
assert (m2 >= 0);
m1 += m2;
if (((long) (b + 3) & 0xffffffffL) <= 3 + 3) {
/* approximately, 1 bark = 3 partitions */
/* 65% of the cases */
/* originally 'if(i > 8)' */
if (ratio >= ma_max_i1) {
/* 43% of the total */
return m1;
}
/* 22% of the total */
int i = (int) (Util.FAST_LOG10_X(ratio, 16.0f));
return m1 * table2[i];
}
/**
*
* m<15 equ log10((m1+m2)/gfc.ATH.cb[k])<1.5
* equ (m1+m2)/gfc.ATH.cb[k]<10^1.5
* equ (m1+m2)<10^1.5 * gfc.ATH.cb[k]
*
*/
int i = (int) Util.FAST_LOG10_X(ratio, 16.0f);
if (shortblock != 0) {
m2 = gfc.ATH.cb_s[kk] * gfc.ATH.adjust;
} else {
m2 = gfc.ATH.cb_l[kk] * gfc.ATH.adjust;
}
assert (m2 >= 0);
if (m1 < ma_max_m * m2) {
/* 3% of the total */
/* Originally if (m > 0) { */
if (m1 > m2) {
float f, r;
f = 1.0f;
if (i <= 13)
f = table3[i];
r = Util.FAST_LOG10_X(m1 / m2, 10.0f / 15.0f);
return m1 * ((table1[i] - f) * r + f);
}
if (i > 13)
return m1;
return m1 * table3[i];
}
/* 10% of total */
return m1 * table1[i];
}
private static final float table2_[] = { 1.33352f * 1.33352f,
1.35879f * 1.35879f, 1.38454f * 1.38454f, 1.39497f * 1.39497f,
1.40548f * 1.40548f, 1.3537f * 1.3537f, 1.30382f * 1.30382f,
1.22321f * 1.22321f, 1.14758f * 1.14758f, 1 };
/**
* addition of simultaneous masking Naoki Shibata 2000/7
*/
private float vbrpsy_mask_add(float m1, float m2, final int b) {
float ratio;
if (m1 < 0) {
m1 = 0;
}
if (m2 < 0) {
m2 = 0;
}
if (m1 <= 0) {
return m2;
}
if (m2 <= 0) {
return m1;
}
if (m2 > m1) {
ratio = m2 / m1;
} else {
ratio = m1 / m2;
}
if (-2 <= b && b <= 2) {
/* approximately, 1 bark = 3 partitions */
/* originally 'if(i > 8)' */
if (ratio >= ma_max_i1) {
return m1 + m2;
} else {
int i = (int) (Util.FAST_LOG10_X(ratio, 16.0f));
return (m1 + m2) * table2_[i];
}
}
if (ratio < ma_max_i2) {
return m1 + m2;
}
if (m1 < m2) {
m1 = m2;
}
return m1;
}
/**
* compute interchannel masking effects
*/
private void calc_interchannel_masking(final LameGlobalFlags gfp,
final float ratio) {
final LameInternalFlags gfc = gfp.internal_flags;
if (gfc.channels_out > 1) {
for (int sb = 0; sb < Encoder.SBMAX_l; sb++) {
float l = gfc.thm[0].l[sb];
float r = gfc.thm[1].l[sb];
gfc.thm[0].l[sb] += r * ratio;
gfc.thm[1].l[sb] += l * ratio;
}
for (int sb = 0; sb < Encoder.SBMAX_s; sb++) {
for (int sblock = 0; sblock < 3; sblock++) {
float l = gfc.thm[0].s[sb][sblock];
float r = gfc.thm[1].s[sb][sblock];
gfc.thm[0].s[sb][sblock] += r * ratio;
gfc.thm[1].s[sb][sblock] += l * ratio;
}
}
}
}
/**
* compute M/S thresholds from Johnston & Ferreira 1992 ICASSP paper
*/
private void msfix1(final LameInternalFlags gfc) {
for (int sb = 0; sb < Encoder.SBMAX_l; sb++) {
/* use this fix if L & R masking differs by 2db or less */
/* if db = 10*log10(x2/x1) < 2 */
/* if (x2 < 1.58*x1) { */
if (gfc.thm[0].l[sb] > 1.58 * gfc.thm[1].l[sb]
|| gfc.thm[1].l[sb] > 1.58 * gfc.thm[0].l[sb])
continue;
float mld = gfc.mld_l[sb] * gfc.en[3].l[sb];
float rmid = Math.max(gfc.thm[2].l[sb],
Math.min(gfc.thm[3].l[sb], mld));
mld = gfc.mld_l[sb] * gfc.en[2].l[sb];
float rside = Math.max(gfc.thm[3].l[sb],
Math.min(gfc.thm[2].l[sb], mld));
gfc.thm[2].l[sb] = rmid;
gfc.thm[3].l[sb] = rside;
}
for (int sb = 0; sb < Encoder.SBMAX_s; sb++) {
for (int sblock = 0; sblock < 3; sblock++) {
if (gfc.thm[0].s[sb][sblock] > 1.58 * gfc.thm[1].s[sb][sblock]
|| gfc.thm[1].s[sb][sblock] > 1.58 * gfc.thm[0].s[sb][sblock])
continue;
float mld = gfc.mld_s[sb] * gfc.en[3].s[sb][sblock];
float rmid = Math.max(gfc.thm[2].s[sb][sblock],
Math.min(gfc.thm[3].s[sb][sblock], mld));
mld = gfc.mld_s[sb] * gfc.en[2].s[sb][sblock];
float rside = Math.max(gfc.thm[3].s[sb][sblock],
Math.min(gfc.thm[2].s[sb][sblock], mld));
gfc.thm[2].s[sb][sblock] = rmid;
gfc.thm[3].s[sb][sblock] = rside;
}
}
}
/**
* Adjust M/S maskings if user set "msfix"
*
* Naoki Shibata 2000
*/
private void ns_msfix(final LameInternalFlags gfc, float msfix,
final float athadjust) {
float msfix2 = msfix;
float athlower = (float) Math.pow(10, athadjust);
msfix *= 2.0f;
msfix2 *= 2.0f;
for (int sb = 0; sb < Encoder.SBMAX_l; sb++) {
float thmLR, thmM, thmS, ath;
ath = (gfc.ATH.cb_l[gfc.bm_l[sb]]) * athlower;
thmLR = Math.min(Math.max(gfc.thm[0].l[sb], ath),
Math.max(gfc.thm[1].l[sb], ath));
thmM = Math.max(gfc.thm[2].l[sb], ath);
thmS = Math.max(gfc.thm[3].l[sb], ath);
if (thmLR * msfix < thmM + thmS) {
float f = thmLR * msfix2 / (thmM + thmS);
thmM *= f;
thmS *= f;
assert (thmM + thmS > 0);
}
gfc.thm[2].l[sb] = Math.min(thmM, gfc.thm[2].l[sb]);
gfc.thm[3].l[sb] = Math.min(thmS, gfc.thm[3].l[sb]);
}
athlower *= ((float) Encoder.BLKSIZE_s / Encoder.BLKSIZE);
for (int sb = 0; sb < Encoder.SBMAX_s; sb++) {
for (int sblock = 0; sblock < 3; sblock++) {
float thmLR, thmM, thmS, ath;
ath = (gfc.ATH.cb_s[gfc.bm_s[sb]]) * athlower;
thmLR = Math.min(Math.max(gfc.thm[0].s[sb][sblock], ath),
Math.max(gfc.thm[1].s[sb][sblock], ath));
thmM = Math.max(gfc.thm[2].s[sb][sblock], ath);
thmS = Math.max(gfc.thm[3].s[sb][sblock], ath);
if (thmLR * msfix < thmM + thmS) {
float f = thmLR * msfix / (thmM + thmS);
thmM *= f;
thmS *= f;
assert (thmM + thmS > 0);
}
gfc.thm[2].s[sb][sblock] = Math.min(gfc.thm[2].s[sb][sblock],
thmM);
gfc.thm[3].s[sb][sblock] = Math.min(gfc.thm[3].s[sb][sblock],
thmS);
}
}
}
/**
* short block threshold calculation (part 2)
*
* partition band bo_s[sfb] is at the transition from scalefactor band sfb
* to the next one sfb+1; enn and thmm have to be split between them
*/
private void convert_partition2scalefac_s(final LameInternalFlags gfc,
final float[] eb, final float[] thr, final int chn, final int sblock) {
int sb, b;
float enn = 0.0f;
float thmm = 0.0f;
for (sb = b = 0; sb < Encoder.SBMAX_s; ++b, ++sb) {
int bo_s_sb = gfc.bo_s[sb];
int npart_s = gfc.npart_s;
int b_lim = bo_s_sb < npart_s ? bo_s_sb : npart_s;
while (b < b_lim) {
assert (eb[b] >= 0);
// iff failed, it may indicate some index error elsewhere
assert (thr[b] >= 0);
enn += eb[b];
thmm += thr[b];
b++;
}
gfc.en[chn].s[sb][sblock] = enn;
gfc.thm[chn].s[sb][sblock] = thmm;
if (b >= npart_s) {
++sb;
break;
}
assert (eb[b] >= 0);
// iff failed, it may indicate some index error elsewhere
assert (thr[b] >= 0);
{
/* at transition sfb . sfb+1 */
float w_curr = gfc.PSY.bo_s_weight[sb];
float w_next = 1.0f - w_curr;
enn = w_curr * eb[b];
thmm = w_curr * thr[b];
gfc.en[chn].s[sb][sblock] += enn;
gfc.thm[chn].s[sb][sblock] += thmm;
enn = w_next * eb[b];
thmm = w_next * thr[b];
}
}
/* zero initialize the rest */
for (; sb < Encoder.SBMAX_s; ++sb) {
gfc.en[chn].s[sb][sblock] = 0;
gfc.thm[chn].s[sb][sblock] = 0;
}
}
/**
* longblock threshold calculation (part 2)
*/
private void convert_partition2scalefac_l(final LameInternalFlags gfc,
final float[] eb, final float[] thr, final int chn) {
int sb, b;
float enn = 0.0f;
float thmm = 0.0f;
for (sb = b = 0; sb < Encoder.SBMAX_l; ++b, ++sb) {
int bo_l_sb = gfc.bo_l[sb];
int npart_l = gfc.npart_l;
int b_lim = bo_l_sb < npart_l ? bo_l_sb : npart_l;
while (b < b_lim) {
assert (eb[b] >= 0);
// iff failed, it may indicate some index error elsewhere
assert (thr[b] >= 0);
enn += eb[b];
thmm += thr[b];
b++;
}
gfc.en[chn].l[sb] = enn;
gfc.thm[chn].l[sb] = thmm;
if (b >= npart_l) {
++sb;
break;
}
assert (eb[b] >= 0);
assert (thr[b] >= 0);
{
/* at transition sfb . sfb+1 */
float w_curr = gfc.PSY.bo_l_weight[sb];
float w_next = 1.0f - w_curr;
enn = w_curr * eb[b];
thmm = w_curr * thr[b];
gfc.en[chn].l[sb] += enn;
gfc.thm[chn].l[sb] += thmm;
enn = w_next * eb[b];
thmm = w_next * thr[b];
}
}
/* zero initialize the rest */
for (; sb < Encoder.SBMAX_l; ++sb) {
gfc.en[chn].l[sb] = 0;
gfc.thm[chn].l[sb] = 0;
}
}
private void compute_masking_s(final LameGlobalFlags gfp,
final float fftenergy_s[][], final float[] eb, final float[] thr,
final int chn, final int sblock) {
final LameInternalFlags gfc = gfp.internal_flags;
int j, b;
for (b = j = 0; b < gfc.npart_s; ++b) {
float ebb = 0, m = 0;
int n = gfc.numlines_s[b];
for (int i = 0; i < n; ++i, ++j) {
float el = fftenergy_s[sblock][j];
ebb += el;
if (m < el)
m = el;
}
eb[b] = ebb;
}
assert (b == gfc.npart_s);
assert (j == 129);
for (j = b = 0; b < gfc.npart_s; b++) {
int kk = gfc.s3ind_s[b][0];
float ecb = gfc.s3_ss[j++] * eb[kk];
++kk;
while (kk <= gfc.s3ind_s[b][1]) {
ecb += gfc.s3_ss[j] * eb[kk];
++j;
++kk;
}
{ /* limit calculated threshold by previous granule */
float x = rpelev_s * gfc.nb_s1[chn][b];
thr[b] = Math.min(ecb, x);
}
if (gfc.blocktype_old[chn & 1] == Encoder.SHORT_TYPE) {
/* limit calculated threshold by even older granule */
float x = rpelev2_s * gfc.nb_s2[chn][b];
float y = thr[b];
thr[b] = Math.min(x, y);
}
gfc.nb_s2[chn][b] = gfc.nb_s1[chn][b];
gfc.nb_s1[chn][b] = ecb;
assert (thr[b] >= 0);
}
for (; b <= Encoder.CBANDS; ++b) {
eb[b] = 0;
thr[b] = 0;
}
}
private void block_type_set(final LameGlobalFlags gfp,
final int[] uselongblock, final int[] blocktype_d,
final int[] blocktype) {
final LameInternalFlags gfc = gfp.internal_flags;
if (gfp.short_blocks == ShortBlock.short_block_coupled
/* force both channels to use the same block type */
/* this is necessary if the frame is to be encoded in ms_stereo. */
/* But even without ms_stereo, FhG does this */
&& !(uselongblock[0] != 0 && uselongblock[1] != 0))
uselongblock[0] = uselongblock[1] = 0;
/*
* update the blocktype of the previous granule, since it depends on
* what happend in this granule
*/
for (int chn = 0; chn < gfc.channels_out; chn++) {
blocktype[chn] = Encoder.NORM_TYPE;
/* disable short blocks */
if (gfp.short_blocks == ShortBlock.short_block_dispensed)
uselongblock[chn] = 1;
if (gfp.short_blocks == ShortBlock.short_block_forced)
uselongblock[chn] = 0;
if (uselongblock[chn] != 0) {
/* no attack : use long blocks */
assert (gfc.blocktype_old[chn] != Encoder.START_TYPE);
if (gfc.blocktype_old[chn] == Encoder.SHORT_TYPE)
blocktype[chn] = Encoder.STOP_TYPE;
} else {
/* attack : use short blocks */
blocktype[chn] = Encoder.SHORT_TYPE;
if (gfc.blocktype_old[chn] == Encoder.NORM_TYPE) {
gfc.blocktype_old[chn] = Encoder.START_TYPE;
}
if (gfc.blocktype_old[chn] == Encoder.STOP_TYPE)
gfc.blocktype_old[chn] = Encoder.SHORT_TYPE;
}
blocktype_d[chn] = gfc.blocktype_old[chn];
// value returned to calling program
gfc.blocktype_old[chn] = blocktype[chn];
// save for next call to l3psy_anal
}
}
private float NS_INTERP(final float x, final float y, final float r) {
/* was pow((x),(r))*pow((y),1-(r)) */
if (r >= 1.0) {
/* 99.7% of the time */
return x;
}
if (r <= 0.0)
return y;
if (y > 0.0) {
/* rest of the time */
return (float) (Math.pow(x / y, r) * y);
}
/* never happens */
return 0.0f;
}
/**
* these values are tuned only for 44.1kHz...
*/
private static final float regcoef_s[] = { 11.8f, 13.6f, 17.2f, 32f, 46.5f,
51.3f, 57.5f, 67.1f, 71.5f, 84.6f, 97.6f, 130f,
/* 255.8 */
};
private float pecalc_s(final III_psy_ratio mr,
final float masking_lower) {
float pe_s = 1236.28f / 4;
for (int sb = 0; sb < Encoder.SBMAX_s - 1; sb++) {
for (int sblock = 0; sblock < 3; sblock++) {
float thm = mr.thm.s[sb][sblock];
assert (sb < regcoef_s.length);
if (thm > 0.0) {
float x = thm * masking_lower;
float en = mr.en.s[sb][sblock];
if (en > x) {
if (en > x * 1e10) {
pe_s += regcoef_s[sb] * (10.0f * LOG10);
} else {
assert (x > 0);
pe_s += regcoef_s[sb] * Util.FAST_LOG10(en / x);
}
}
}
}
}
return pe_s;
}
/**
* these values are tuned only for 44.1kHz...
*/
private static final float regcoef_l[] = { 6.8f, 5.8f, 5.8f, 6.4f, 6.5f, 9.9f,
12.1f, 14.4f, 15f, 18.9f, 21.6f, 26.9f, 34.2f, 40.2f, 46.8f, 56.5f,
60.7f, 73.9f, 85.7f, 93.4f, 126.1f,
/* 241.3 */
};
private float pecalc_l(final III_psy_ratio mr,
final float masking_lower) {
float pe_l = 1124.23f / 4;
for (int sb = 0; sb < Encoder.SBMAX_l - 1; sb++) {
float thm = mr.thm.l[sb];
assert (sb < regcoef_l.length);
if (thm > 0.0) {
float x = thm * masking_lower;
float en = mr.en.l[sb];
if (en > x) {
if (en > x * 1e10) {
pe_l += regcoef_l[sb] * (10.0f * LOG10);
} else {
assert (x > 0);
pe_l += regcoef_l[sb] * Util.FAST_LOG10(en / x);
}
}
}
}
return pe_l;
}
private void calc_energy(final LameInternalFlags gfc,
final float[] fftenergy, final float[] eb, final float[] max,
final float[] avg) {
int b, j;
for (b = j = 0; b < gfc.npart_l; ++b) {
float ebb = 0, m = 0;
int i;
for (i = 0; i < gfc.numlines_l[b]; ++i, ++j) {
float el = fftenergy[j];
assert (el >= 0);
ebb += el;
if (m < el)
m = el;
}
eb[b] = ebb;
max[b] = m;
avg[b] = ebb * gfc.rnumlines_l[b];
assert (gfc.rnumlines_l[b] >= 0);
assert (ebb >= 0);
assert (eb[b] >= 0);
assert (max[b] >= 0);
assert (avg[b] >= 0);
}
}
private void calc_mask_index_l(final LameInternalFlags gfc,
final float[] max, final float[] avg, final int[] mask_idx) {
int last_tab_entry = tab.length - 1;
int b = 0;
float a = avg[b] + avg[b + 1];
assert (a >= 0);
if (a > 0.0) {
float m = max[b];
if (m < max[b + 1])
m = max[b + 1];
assert ((gfc.numlines_l[b] + gfc.numlines_l[b + 1] - 1) > 0);
a = 20.0f * (m * 2.0f - a)
/ (a * (gfc.numlines_l[b] + gfc.numlines_l[b + 1] - 1));
int k = (int) a;
if (k > last_tab_entry)
k = last_tab_entry;
mask_idx[b] = k;
} else {
mask_idx[b] = 0;
}
for (b = 1; b < gfc.npart_l - 1; b++) {
a = avg[b - 1] + avg[b] + avg[b + 1];
assert (a >= 0);
if (a > 0.0) {
float m = max[b - 1];
if (m < max[b])
m = max[b];
if (m < max[b + 1])
m = max[b + 1];
assert ((gfc.numlines_l[b - 1] + gfc.numlines_l[b]
+ gfc.numlines_l[b + 1] - 1) > 0);
a = 20.0f
* (m * 3.0f - a)
/ (a * (gfc.numlines_l[b - 1] + gfc.numlines_l[b]
+ gfc.numlines_l[b + 1] - 1));
int k = (int) a;
if (k > last_tab_entry)
k = last_tab_entry;
mask_idx[b] = k;
} else {
mask_idx[b] = 0;
}
}
assert (b > 0);
assert (b == gfc.npart_l - 1);
a = avg[b - 1] + avg[b];
assert (a >= 0);
if (a > 0.0) {
float m = max[b - 1];
if (m < max[b])
m = max[b];
assert ((gfc.numlines_l[b - 1] + gfc.numlines_l[b] - 1) > 0);
a = 20.0f * (m * 2.0f - a)
/ (a * (gfc.numlines_l[b - 1] + gfc.numlines_l[b] - 1));
int k = (int) a;
if (k > last_tab_entry)
k = last_tab_entry;
mask_idx[b] = k;
} else {
mask_idx[b] = 0;
}
assert (b == (gfc.npart_l - 1));
}
private static final float fircoef[] = {
-8.65163e-18f * 2, -0.00851586f * 2, -6.74764e-18f * 2, 0.0209036f * 2,
-3.36639e-17f * 2, -0.0438162f * 2, -1.54175e-17f * 2, 0.0931738f * 2,
-5.52212e-17f * 2, -0.313819f * 2
};
public final int L3psycho_anal_ns(final LameGlobalFlags gfp,
final float[] buffer[], final int bufPos, final int gr_out,
final III_psy_ratio masking_ratio[][],
final III_psy_ratio masking_MS_ratio[][],
final float percep_entropy[], final float percep_MS_entropy[],
final float energy[], final int blocktype_d[]) {
/*
* to get a good cache performance, one has to think about the sequence,
* in which the variables are used.
*/
final LameInternalFlags gfc = gfp.internal_flags;
/* fft and energy calculation */
float wsamp_L[][] = new float[2][Encoder.BLKSIZE];
float wsamp_S[][][] = new float[2][3][Encoder.BLKSIZE_s];
/* convolution */
float eb_l[] = new float[Encoder.CBANDS + 1], eb_s[] = new float[Encoder.CBANDS + 1];
float thr[] = new float[Encoder.CBANDS + 2];
/* block type */
int blocktype[] = new int[2], uselongblock[] = new int[2];
/* usual variables like loop indices, etc.. */
int numchn, chn;
int b, i, j, k;
int sb, sblock;
/* variables used for --nspsytune */
float ns_hpfsmpl[][] = new float[2][576];
float pcfact;
int mask_idx_l[] = new int[Encoder.CBANDS + 2], mask_idx_s[] = new int[Encoder.CBANDS + 2];
Arrays.fill(mask_idx_s, 0);
numchn = gfc.channels_out;
/* chn=2 and 3 = Mid and Side channels */
if (gfp.mode == MPEGMode.JOINT_STEREO)
numchn = 4;
if (gfp.VBR == VbrMode.vbr_off)
pcfact = gfc.ResvMax == 0 ? 0f : ((float) gfc.ResvSize)
/ gfc.ResvMax * 0.5f;
else if (gfp.VBR == VbrMode.vbr_rh || gfp.VBR == VbrMode.vbr_mtrh
|| gfp.VBR == VbrMode.vbr_mt) {
pcfact = 0.6f;
} else
pcfact = 1.0f;
/**********************************************************************
* Apply HPF of fs/4 to the input signal. This is used for attack
* detection / handling.
**********************************************************************/
/* Don't copy the input buffer into a temporary buffer */
/* unroll the loop 2 times */
for (chn = 0; chn < gfc.channels_out; chn++) {
/* apply high pass filter of fs/4 */
final float[] firbuf = buffer[chn];
int firbufPos = bufPos + 576 - 350 - NSFIRLEN + 192;
assert (fircoef.length == ((NSFIRLEN - 1) / 2));
for (i = 0; i < 576; i++) {
float sum1, sum2;
sum1 = firbuf[firbufPos + i + 10];
sum2 = 0.0f;
for (j = 0; j < ((NSFIRLEN - 1) / 2) - 1; j += 2) {
sum1 += fircoef[j]
* (firbuf[firbufPos + i + j] + firbuf[firbufPos + i
+ NSFIRLEN - j]);
sum2 += fircoef[j + 1]
* (firbuf[firbufPos + i + j + 1] + firbuf[firbufPos
+ i + NSFIRLEN - j - 1]);
}
ns_hpfsmpl[chn][i] = sum1 + sum2;
}
masking_ratio[gr_out][chn].en.assign(gfc.en[chn]);
masking_ratio[gr_out][chn].thm.assign(gfc.thm[chn]);
if (numchn > 2) {
/* MS maskings */
/* percep_MS_entropy [chn-2] = gfc . pe [chn]; */
masking_MS_ratio[gr_out][chn].en.assign(gfc.en[chn + 2]);
masking_MS_ratio[gr_out][chn].thm.assign(gfc.thm[chn + 2]);
}
}
for (chn = 0; chn < numchn; chn++) {
float wsamp_l[][];
float wsamp_s[][][];
float en_subshort[] = new float[12];
float en_short[] = { 0, 0, 0, 0 };
float attack_intensity[] = new float[12];
int ns_uselongblock = 1;
float attackThreshold;
float max[] = new float[Encoder.CBANDS], avg[] = new float[Encoder.CBANDS];
int ns_attacks[] = { 0, 0, 0, 0 };
float fftenergy[] = new float[Encoder.HBLKSIZE];
float fftenergy_s[][] = new float[3][Encoder.HBLKSIZE_s];
/*
* rh 20040301: the following loops do access one off the limits so
* I increase the array dimensions by one and initialize the
* accessed values to zero
*/
assert (gfc.npart_s <= Encoder.CBANDS);
assert (gfc.npart_l <= Encoder.CBANDS);
/***************************************************************
* determine the block type (window type)
***************************************************************/
/* calculate energies of each sub-shortblocks */
for (i = 0; i < 3; i++) {
en_subshort[i] = gfc.nsPsy.last_en_subshort[chn][i + 6];
assert (gfc.nsPsy.last_en_subshort[chn][i + 4] > 0);
attack_intensity[i] = en_subshort[i]
/ gfc.nsPsy.last_en_subshort[chn][i + 4];
en_short[0] += en_subshort[i];
}
if (chn == 2) {
for (i = 0; i < 576; i++) {
float l, r;
l = ns_hpfsmpl[0][i];
r = ns_hpfsmpl[1][i];
ns_hpfsmpl[0][i] = l + r;
ns_hpfsmpl[1][i] = l - r;
}
}
{
float[] pf = ns_hpfsmpl[chn & 1];
int pfPos = 0;
for (i = 0; i < 9; i++) {
int pfe = pfPos + 576 / 9;
float p = 1.f;
for (; pfPos < pfe; pfPos++)
if (p < Math.abs(pf[pfPos]))
p = Math.abs(pf[pfPos]);
gfc.nsPsy.last_en_subshort[chn][i] = en_subshort[i + 3] = p;
en_short[1 + i / 3] += p;
if (p > en_subshort[i + 3 - 2]) {
assert (en_subshort[i + 3 - 2] > 0);
p = p / en_subshort[i + 3 - 2];
} else if (en_subshort[i + 3 - 2] > p * 10.0f) {
assert (p > 0);
p = en_subshort[i + 3 - 2] / (p * 10.0f);
} else
p = 0.0f;
attack_intensity[i + 3] = p;
}
}
if (gfp.analysis) {
float x = attack_intensity[0];
for (i = 1; i < 12; i++)
if (x < attack_intensity[i])
x = attack_intensity[i];
gfc.pinfo.ers[gr_out][chn] = gfc.pinfo.ers_save[chn];
gfc.pinfo.ers_save[chn] = x;
}
/* compare energies between sub-shortblocks */
attackThreshold = (chn == 3) ? gfc.nsPsy.attackthre_s
: gfc.nsPsy.attackthre;
for (i = 0; i < 12; i++)
if (0 == ns_attacks[i / 3]
&& attack_intensity[i] > attackThreshold)
ns_attacks[i / 3] = (i % 3) + 1;
/*
* should have energy change between short blocks, in order to avoid
* periodic signals
*/
for (i = 1; i < 4; i++) {
float ratio;
if (en_short[i - 1] > en_short[i]) {
assert (en_short[i] > 0);
ratio = en_short[i - 1] / en_short[i];
} else {
assert (en_short[i - 1] > 0);
ratio = en_short[i] / en_short[i - 1];
}
if (ratio < 1.7) {
ns_attacks[i] = 0;
if (i == 1)
ns_attacks[0] = 0;
}
}
if (ns_attacks[0] != 0 && gfc.nsPsy.lastAttacks[chn] != 0)
ns_attacks[0] = 0;
if (gfc.nsPsy.lastAttacks[chn] == 3
|| (ns_attacks[0] + ns_attacks[1] + ns_attacks[2] + ns_attacks[3]) != 0) {
ns_uselongblock = 0;
if (ns_attacks[1] != 0 && ns_attacks[0] != 0)
ns_attacks[1] = 0;
if (ns_attacks[2] != 0 && ns_attacks[1] != 0)
ns_attacks[2] = 0;
if (ns_attacks[3] != 0 && ns_attacks[2] != 0)
ns_attacks[3] = 0;
}
if (chn < 2) {
uselongblock[chn] = ns_uselongblock;
} else {
if (ns_uselongblock == 0) {
uselongblock[0] = uselongblock[1] = 0;
}
}
/*
* there is a one granule delay. Copy maskings computed last call
* into masking_ratio to return to calling program.
*/
energy[chn] = gfc.tot_ener[chn];
/*********************************************************************
* compute FFTs
*********************************************************************/
wsamp_s = wsamp_S;
wsamp_l = wsamp_L;
compute_ffts(gfp, fftenergy, fftenergy_s, wsamp_l, (chn & 1),
wsamp_s, (chn & 1), gr_out, chn, buffer, bufPos);
/*********************************************************************
* Calculate the energy and the tonality of each partition.
*********************************************************************/
calc_energy(gfc, fftenergy, eb_l, max, avg);
calc_mask_index_l(gfc, max, avg, mask_idx_l);
/* compute masking thresholds for short blocks */
for (sblock = 0; sblock < 3; sblock++) {
float enn, thmm;
compute_masking_s(gfp, fftenergy_s, eb_s, thr, chn, sblock);
convert_partition2scalefac_s(gfc, eb_s, thr, chn, sblock);
/**** short block pre-echo control ****/
for (sb = 0; sb < Encoder.SBMAX_s; sb++) {
thmm = gfc.thm[chn].s[sb][sblock];
thmm *= NS_PREECHO_ATT0;
if (ns_attacks[sblock] >= 2 || ns_attacks[sblock + 1] == 1) {
int idx = (sblock != 0) ? sblock - 1 : 2;
double p = NS_INTERP(gfc.thm[chn].s[sb][idx], thmm,
NS_PREECHO_ATT1 * pcfact);
thmm = (float) Math.min(thmm, p);
}
if (ns_attacks[sblock] == 1) {
int idx = (sblock != 0) ? sblock - 1 : 2;
double p = NS_INTERP(gfc.thm[chn].s[sb][idx], thmm,
NS_PREECHO_ATT2 * pcfact);
thmm = (float) Math.min(thmm, p);
} else if ((sblock != 0 && ns_attacks[sblock - 1] == 3)
|| (sblock == 0 && gfc.nsPsy.lastAttacks[chn] == 3)) {
int idx = (sblock != 2) ? sblock + 1 : 0;
double p = NS_INTERP(gfc.thm[chn].s[sb][idx], thmm,
NS_PREECHO_ATT2 * pcfact);
thmm = (float) Math.min(thmm, p);
}
/* pulse like signal detection for fatboy.wav and so on */
enn = en_subshort[sblock * 3 + 3]
+ en_subshort[sblock * 3 + 4]
+ en_subshort[sblock * 3 + 5];
if (en_subshort[sblock * 3 + 5] * 6 < enn) {
thmm *= 0.5;
if (en_subshort[sblock * 3 + 4] * 6 < enn)
thmm *= 0.5;
}
gfc.thm[chn].s[sb][sblock] = thmm;
}
}
gfc.nsPsy.lastAttacks[chn] = ns_attacks[2];
/*********************************************************************
* convolve the partitioned energy and unpredictability with the
* spreading function, s3_l[b][k]
********************************************************************/
k = 0;
{
for (b = 0; b < gfc.npart_l; b++) {
/*
* convolve the partitioned energy with the spreading
* function
*/
int kk = gfc.s3ind[b][0];
float eb2 = eb_l[kk] * tab[mask_idx_l[kk]];
float ecb = gfc.s3_ll[k++] * eb2;
while (++kk <= gfc.s3ind[b][1]) {
eb2 = eb_l[kk] * tab[mask_idx_l[kk]];
ecb = mask_add(ecb, gfc.s3_ll[k++] * eb2, kk, kk - b,
gfc, 0);
}
ecb *= 0.158489319246111; /* pow(10,-0.8) */
/**** long block pre-echo control ****/
/**
*
* dont use long block pre-echo control if previous granule was
* a short block. This is to avoid the situation:
* frame0: quiet (very low masking)
* frame1: surge (triggers short blocks)
* frame2: regular frame. looks like pre-echo when compared to
* frame0, but all pre-echo was in frame1.
*
*/
/*
* chn=0,1 L and R channels
*
* chn=2,3 S and M channels.
*/
if (gfc.blocktype_old[chn & 1] == Encoder.SHORT_TYPE)
thr[b] = ecb;
else
thr[b] = NS_INTERP(
Math.min(ecb, Math.min(rpelev
* gfc.nb_1[chn][b], rpelev2
* gfc.nb_2[chn][b])), ecb, pcfact);
gfc.nb_2[chn][b] = gfc.nb_1[chn][b];
gfc.nb_1[chn][b] = ecb;
}
}
for (; b <= Encoder.CBANDS; ++b) {
eb_l[b] = 0;
thr[b] = 0;
}
/* compute masking thresholds for long blocks */
convert_partition2scalefac_l(gfc, eb_l, thr, chn);
} /* end loop over chn */
if (gfp.mode == MPEGMode.STEREO || gfp.mode == MPEGMode.JOINT_STEREO) {
if (gfp.interChRatio > 0.0) {
calc_interchannel_masking(gfp, gfp.interChRatio);
}
}
if (gfp.mode == MPEGMode.JOINT_STEREO) {
float msfix;
msfix1(gfc);
msfix = gfp.msfix;
if (Math.abs(msfix) > 0.0)
ns_msfix(gfc, msfix, gfp.ATHlower * gfc.ATH.adjust);
}
/***************************************************************
* determine final block type
***************************************************************/
block_type_set(gfp, uselongblock, blocktype_d, blocktype);
/*********************************************************************
* compute the value of PE to return ... no delay and advance
*********************************************************************/
for (chn = 0; chn < numchn; chn++) {
float[] ppe;
int ppePos = 0;
int type;
III_psy_ratio mr;
if (chn > 1) {
ppe = percep_MS_entropy;
ppePos = -2;
type = Encoder.NORM_TYPE;
if (blocktype_d[0] == Encoder.SHORT_TYPE
|| blocktype_d[1] == Encoder.SHORT_TYPE)
type = Encoder.SHORT_TYPE;
mr = masking_MS_ratio[gr_out][chn - 2];
} else {
ppe = percep_entropy;
ppePos = 0;
type = blocktype_d[chn];
mr = masking_ratio[gr_out][chn];
}
if (type == Encoder.SHORT_TYPE)
ppe[ppePos + chn] = pecalc_s(mr, gfc.masking_lower);
else
ppe[ppePos + chn] = pecalc_l(mr, gfc.masking_lower);
if (gfp.analysis)
gfc.pinfo.pe[gr_out][chn] = ppe[ppePos + chn];
}
return 0;
}
private void vbrpsy_compute_fft_l(final LameGlobalFlags gfp,
final float[] buffer[], final int bufPos, final int chn,
final int gr_out, final float fftenergy[], final float wsamp_l[][],
final int wsamp_lPos) {
final LameInternalFlags gfc = gfp.internal_flags;
if (chn < 2) {
fft.fft_long(gfc, wsamp_l[wsamp_lPos], chn, buffer, bufPos);
} else if (chn == 2) {
/* FFT data for mid and side channel is derived from L & R */
for (int j = Encoder.BLKSIZE - 1; j >= 0; --j) {
float l = wsamp_l[wsamp_lPos + 0][j];
float r = wsamp_l[wsamp_lPos + 1][j];
wsamp_l[wsamp_lPos + 0][j] = (l + r) * Util.SQRT2 * 0.5f;
wsamp_l[wsamp_lPos + 1][j] = (l - r) * Util.SQRT2 * 0.5f;
}
}
/*********************************************************************
* compute energies
*********************************************************************/
fftenergy[0] = NON_LINEAR_SCALE_ENERGY(wsamp_l[wsamp_lPos + 0][0]);
fftenergy[0] *= fftenergy[0];
for (int j = Encoder.BLKSIZE / 2 - 1; j >= 0; --j) {
float re = wsamp_l[wsamp_lPos + 0][Encoder.BLKSIZE / 2 - j];
float im = wsamp_l[wsamp_lPos + 0][Encoder.BLKSIZE / 2 + j];
fftenergy[Encoder.BLKSIZE / 2 - j] = NON_LINEAR_SCALE_ENERGY((re
* re + im * im) * 0.5f);
}
/* total energy */
{
float totalenergy = 0.0f;
for (int j = 11; j < Encoder.HBLKSIZE; j++)
totalenergy += fftenergy[j];
gfc.tot_ener[chn] = totalenergy;
}
if (gfp.analysis) {
for (int j = 0; j < Encoder.HBLKSIZE; j++) {
gfc.pinfo.energy[gr_out][chn][j] = gfc.pinfo.energy_save[chn][j];
gfc.pinfo.energy_save[chn][j] = fftenergy[j];
}
gfc.pinfo.pe[gr_out][chn] = gfc.pe[chn];
}
}
private void vbrpsy_compute_fft_s(final LameGlobalFlags gfp,
final float[] buffer[], final int bufPos, final int chn,
final int sblock, final float fftenergy_s[][],
final float wsamp_s[][][], int wsamp_sPos) {
final LameInternalFlags gfc = gfp.internal_flags;
if (sblock == 0 && chn < 2) {
fft.fft_short(gfc, wsamp_s[wsamp_sPos], chn, buffer, bufPos);
}
if (chn == 2) {
/* FFT data for mid and side channel is derived from L & R */
for (int j = Encoder.BLKSIZE_s - 1; j >= 0; --j) {
float l = wsamp_s[wsamp_sPos+0][sblock][j];
float r = wsamp_s[wsamp_sPos+1][sblock][j];
wsamp_s[wsamp_sPos+0][sblock][j] = (l + r) * Util.SQRT2 * 0.5f;
wsamp_s[wsamp_sPos+1][sblock][j] = (l - r) * Util.SQRT2 * 0.5f;
}
}
/*********************************************************************
* compute energies
*********************************************************************/
fftenergy_s[sblock][0] = wsamp_s[wsamp_sPos+0][sblock][0];
fftenergy_s[sblock][0] *= fftenergy_s[sblock][0];
for (int j = Encoder.BLKSIZE_s / 2 - 1; j >= 0; --j) {
float re = wsamp_s[wsamp_sPos+0][sblock][Encoder.BLKSIZE_s / 2 - j];
float im = wsamp_s[wsamp_sPos+0][sblock][Encoder.BLKSIZE_s / 2 + j];
fftenergy_s[sblock][Encoder.BLKSIZE_s / 2 - j] = NON_LINEAR_SCALE_ENERGY((re
* re + im * im) * 0.5f);
}
}
/**
* compute loudness approximation (used for ATH auto-level adjustment)
*/
private void vbrpsy_compute_loudness_approximation_l(LameGlobalFlags gfp,
int gr_out, int chn, float fftenergy[]) {
final LameInternalFlags gfc = gfp.internal_flags;
if (gfp.athaa_loudapprox == 2 && chn < 2) {
// no loudness for mid/side ch
gfc.loudness_sq[gr_out][chn] = gfc.loudness_sq_save[chn];
gfc.loudness_sq_save[chn] = psycho_loudness_approx(fftenergy, gfc);
}
}
private static final float fircoef_[] = { -8.65163e-18f * 2,
-0.00851586f * 2, -6.74764e-18f * 2, 0.0209036f * 2,
-3.36639e-17f * 2, -0.0438162f * 2, -1.54175e-17f * 2,
0.0931738f * 2, -5.52212e-17f * 2, -0.313819f * 2 };
/**
* Apply HPF of fs/4 to the input signal. This is used for attack detection
* / handling.
*/
private void vbrpsy_attack_detection(final LameGlobalFlags gfp,
final float[] buffer[], final int bufPos, final int gr_out,
final III_psy_ratio masking_ratio[][],
final III_psy_ratio masking_MS_ratio[][],
final float energy[], float sub_short_factor[][],
final int ns_attacks[][], final int uselongblock[]) {
float ns_hpfsmpl[][] = new float[2][576];
final LameInternalFlags gfc = gfp.internal_flags;
int n_chn_out = gfc.channels_out;
/* chn=2 and 3 = Mid and Side channels */
int n_chn_psy = (gfp.mode == MPEGMode.JOINT_STEREO) ? 4 : n_chn_out;
/* Don't copy the input buffer into a temporary buffer */
/* unroll the loop 2 times */
for (int chn = 0; chn < n_chn_out; chn++) {
/* apply high pass filter of fs/4 */
final float[] firbuf = buffer[chn];
int firbufPos = bufPos + 576 - 350 - NSFIRLEN + 192;
assert (fircoef_.length == ((NSFIRLEN - 1) / 2));
for (int i = 0; i < 576; i++) {
float sum1, sum2;
sum1 = firbuf[firbufPos + i + 10];
sum2 = 0.0f;
for (int j = 0; j < ((NSFIRLEN - 1) / 2) - 1; j += 2) {
sum1 += fircoef_[j]
* (firbuf[firbufPos + i + j] + firbuf[firbufPos + i
+ NSFIRLEN - j]);
sum2 += fircoef_[j + 1]
* (firbuf[firbufPos + i + j + 1] + firbuf[firbufPos
+ i + NSFIRLEN - j - 1]);
}
ns_hpfsmpl[chn][i] = sum1 + sum2;
}
masking_ratio[gr_out][chn].en.assign(gfc.en[chn]);
masking_ratio[gr_out][chn].thm.assign(gfc.thm[chn]);
if (n_chn_psy > 2) {
/* MS maskings */
/* percep_MS_entropy [chn-2] = gfc . pe [chn]; */
masking_MS_ratio[gr_out][chn].en.assign(gfc.en[chn + 2]);
masking_MS_ratio[gr_out][chn].thm.assign(gfc.thm[chn + 2]);
}
}
for (int chn = 0; chn < n_chn_psy; chn++) {
float attack_intensity[] = new float[12];
float en_subshort[] = new float[12];
float en_short[] = { 0, 0, 0, 0 };
float[] pf = ns_hpfsmpl[chn & 1];
int pfPos = 0;
final float attackThreshold = (chn == 3) ? gfc.nsPsy.attackthre_s
: gfc.nsPsy.attackthre;
int ns_uselongblock = 1;
if (chn == 2) {
for (int i = 0, j = 576; j > 0; ++i, --j) {
final float l = ns_hpfsmpl[0][i];
final float r = ns_hpfsmpl[1][i];
ns_hpfsmpl[0][i] = l + r;
ns_hpfsmpl[1][i] = l - r;
}
}
/***************************************************************
* determine the block type (window type)
***************************************************************/
/* calculate energies of each sub-shortblocks */
for (int i = 0; i < 3; i++) {
en_subshort[i] = gfc.nsPsy.last_en_subshort[chn][i + 6];
assert (gfc.nsPsy.last_en_subshort[chn][i + 4] > 0);
attack_intensity[i] = en_subshort[i]
/ gfc.nsPsy.last_en_subshort[chn][i + 4];
en_short[0] += en_subshort[i];
}
for (int i = 0; i < 9; i++) {
final int pfe = pfPos + 576 / 9;
float p = 1.f;
for (; pfPos < pfe; pfPos++)
if (p < Math.abs(pf[pfPos]))
p = Math.abs(pf[pfPos]);
gfc.nsPsy.last_en_subshort[chn][i] = en_subshort[i + 3] = p;
en_short[1 + i / 3] += p;
if (p > en_subshort[i + 3 - 2]) {
assert (en_subshort[i + 3 - 2] > 0);
p = p / en_subshort[i + 3 - 2];
} else if (en_subshort[i + 3 - 2] > p * 10.0) {
assert (p > 0);
p = en_subshort[i + 3 - 2] / (p * 10.0f);
} else {
p = 0.0f;
}
attack_intensity[i + 3] = p;
}
/* pulse like signal detection for fatboy.wav and so on */
for (int i = 0; i < 3; ++i) {
final float enn = en_subshort[i * 3 + 3]
+ en_subshort[i * 3 + 4] + en_subshort[i * 3 + 5];
float factor = 1.f;
if (en_subshort[i * 3 + 5] * 6 < enn) {
factor *= 0.5f;
if (en_subshort[i * 3 + 4] * 6 < enn) {
factor *= 0.5f;
}
}
sub_short_factor[chn][i] = factor;
}
if (gfp.analysis) {
float x = attack_intensity[0];
for (int i = 1; i < 12; i++) {
if (x < attack_intensity[i]) {
x = attack_intensity[i];
}
}
gfc.pinfo.ers[gr_out][chn] = gfc.pinfo.ers_save[chn];
gfc.pinfo.ers_save[chn] = x;
}
/* compare energies between sub-shortblocks */
for (int i = 0; i < 12; i++) {
if (0 == ns_attacks[chn][i / 3]
&& attack_intensity[i] > attackThreshold) {
ns_attacks[chn][i / 3] = (i % 3) + 1;
}
}
/*
* should have energy change between short blocks, in order to avoid
* periodic signals
*/
/* Good samples to show the effect are Trumpet test songs */
/*
* GB: tuned (1) to avoid too many short blocks for test sample
* TRUMPET
*/
/*
* RH: tuned (2) to let enough short blocks through for test sample
* FSOL and SNAPS
*/
for (int i = 1; i < 4; i++) {
final float u = en_short[i - 1];
final float v = en_short[i];
final float m = Math.max(u, v);
if (m < 40000) { /* (2) */
if (u < 1.7 * v && v < 1.7 * u) { /* (1) */
if (i == 1 && ns_attacks[chn][0] <= ns_attacks[chn][i]) {
ns_attacks[chn][0] = 0;
}
ns_attacks[chn][i] = 0;
}
}
}
if (ns_attacks[chn][0] <= gfc.nsPsy.lastAttacks[chn]) {
ns_attacks[chn][0] = 0;
}
if (gfc.nsPsy.lastAttacks[chn] == 3
|| (ns_attacks[chn][0] + ns_attacks[chn][1]
+ ns_attacks[chn][2] + ns_attacks[chn][3]) != 0) {
ns_uselongblock = 0;
if (ns_attacks[chn][1] != 0 && ns_attacks[chn][0] != 0) {
ns_attacks[chn][1] = 0;
}
if (ns_attacks[chn][2] != 0 && ns_attacks[chn][1] != 0) {
ns_attacks[chn][2] = 0;
}
if (ns_attacks[chn][3] != 0 && ns_attacks[chn][2] != 0) {
ns_attacks[chn][3] = 0;
}
}
if (chn < 2) {
uselongblock[chn] = ns_uselongblock;
} else {
if (ns_uselongblock == 0) {
uselongblock[0] = uselongblock[1] = 0;
}
}
/*
* there is a one granule delay. Copy maskings computed last call
* into masking_ratio to return to calling program.
*/
energy[chn] = gfc.tot_ener[chn];
}
}
private void vbrpsy_skip_masking_s(final LameInternalFlags gfc,
final int chn, final int sblock) {
if (sblock == 0) {
for (int b = 0; b < gfc.npart_s; b++) {
gfc.nb_s2[chn][b] = gfc.nb_s1[chn][b];
gfc.nb_s1[chn][b] = 0;
}
}
}
private void vbrpsy_skip_masking_l(final LameInternalFlags gfc,
final int chn) {
for (int b = 0; b < gfc.npart_l; b++) {
gfc.nb_2[chn][b] = gfc.nb_1[chn][b];
gfc.nb_1[chn][b] = 0;
}
}
private void psyvbr_calc_mask_index_s(final LameInternalFlags gfc,
final float[] max, final float[] avg, final int[] mask_idx) {
int last_tab_entry = tab.length - 1;
int b = 0;
float a = avg[b] + avg[b + 1];
assert (a >= 0);
if (a > 0.0) {
float m = max[b];
if (m < max[b + 1])
m = max[b + 1];
assert ((gfc.numlines_s[b] + gfc.numlines_s[b + 1] - 1) > 0);
a = 20.0f * (m * 2.0f - a)
/ (a * (gfc.numlines_s[b] + gfc.numlines_s[b + 1] - 1));
int k = (int) a;
if (k > last_tab_entry)
k = last_tab_entry;
mask_idx[b] = k;
} else {
mask_idx[b] = 0;
}
for (b = 1; b < gfc.npart_s - 1; b++) {
a = avg[b - 1] + avg[b] + avg[b + 1];
assert (b + 1 < gfc.npart_s);
assert (a >= 0);
if (a > 0.0) {
float m = max[b - 1];
if (m < max[b])
m = max[b];
if (m < max[b + 1])
m = max[b + 1];
assert ((gfc.numlines_s[b - 1] + gfc.numlines_s[b]
+ gfc.numlines_s[b + 1] - 1) > 0);
a = 20.0f
* (m * 3.0f - a)
/ (a * (gfc.numlines_s[b - 1] + gfc.numlines_s[b]
+ gfc.numlines_s[b + 1] - 1));
int k = (int) a;
if (k > last_tab_entry)
k = last_tab_entry;
mask_idx[b] = k;
} else {
mask_idx[b] = 0;
}
}
assert (b > 0);
assert (b == gfc.npart_s - 1);
a = avg[b - 1] + avg[b];
assert (a >= 0);
if (a > 0.0) {
float m = max[b - 1];
if (m < max[b])
m = max[b];
assert ((gfc.numlines_s[b - 1] + gfc.numlines_s[b] - 1) > 0);
a = 20.0f * (m * 2.0f - a)
/ (a * (gfc.numlines_s[b - 1] + gfc.numlines_s[b] - 1));
int k = (int) a;
if (k > last_tab_entry)
k = last_tab_entry;
mask_idx[b] = k;
} else {
mask_idx[b] = 0;
}
assert (b == (gfc.npart_s - 1));
}
private void vbrpsy_compute_masking_s(final LameGlobalFlags gfp,
final float[] fftenergy_s[], final float[] eb, final float[] thr,
final int chn, final int sblock) {
final LameInternalFlags gfc = gfp.internal_flags;
float max[] = new float[Encoder.CBANDS], avg[] = new float[Encoder.CBANDS];
int i, j, b;
int mask_idx_s[] = new int[Encoder.CBANDS];
for (b = j = 0; b < gfc.npart_s; ++b) {
float ebb = 0, m = 0;
int n = gfc.numlines_s[b];
for (i = 0; i < n; ++i, ++j) {
float el = fftenergy_s[sblock][j];
ebb += el;
if (m < el)
m = el;
}
eb[b] = ebb;
assert (ebb >= 0);
max[b] = m;
assert (n > 0);
avg[b] = ebb / n;
assert (avg[b] >= 0);
}
assert (b == gfc.npart_s);
assert (j == 129);
for (; b < Encoder.CBANDS; ++b) {
max[b] = 0;
avg[b] = 0;
}
psyvbr_calc_mask_index_s(gfc, max, avg, mask_idx_s);
for (j = b = 0; b < gfc.npart_s; b++) {
int kk = gfc.s3ind_s[b][0];
int last = gfc.s3ind_s[b][1];
int dd, dd_n;
float x, ecb, avg_mask;
dd = mask_idx_s[kk];
dd_n = 1;
ecb = gfc.s3_ss[j] * eb[kk] * tab[mask_idx_s[kk]];
++j;
++kk;
while (kk <= last) {
dd += mask_idx_s[kk];
dd_n += 1;
x = gfc.s3_ss[j] * eb[kk] * tab[mask_idx_s[kk]];
ecb = vbrpsy_mask_add(ecb, x, kk - b);
++j;
++kk;
}
dd = (1 + 2 * dd) / (2 * dd_n);
avg_mask = tab[dd] * 0.5f;
ecb *= avg_mask;
thr[b] = ecb;
gfc.nb_s2[chn][b] = gfc.nb_s1[chn][b];
gfc.nb_s1[chn][b] = ecb;
{
/*
* if THR exceeds EB, the quantization routines will take the
* difference from other bands. in case of strong tonal samples
* (tonaltest.wav) this leads to heavy distortions. that's why
* we limit THR here.
*/
x = max[b];
x *= gfc.minval_s[b];
x *= avg_mask;
if (thr[b] > x) {
thr[b] = x;
}
}
if (gfc.masking_lower > 1) {
thr[b] *= gfc.masking_lower;
}
if (thr[b] > eb[b]) {
thr[b] = eb[b];
}
if (gfc.masking_lower < 1) {
thr[b] *= gfc.masking_lower;
}
assert (thr[b] >= 0);
}
for (; b < Encoder.CBANDS; ++b) {
eb[b] = 0;
thr[b] = 0;
}
}
private void vbrpsy_compute_masking_l(final LameInternalFlags gfc,
final float fftenergy[], final float eb_l[], final float thr[],
final int chn) {
float max[] = new float[Encoder.CBANDS], avg[] = new float[Encoder.CBANDS];
int mask_idx_l[] = new int[Encoder.CBANDS + 2];
int b;
/*********************************************************************
* Calculate the energy and the tonality of each partition.
*********************************************************************/
calc_energy(gfc, fftenergy, eb_l, max, avg);
calc_mask_index_l(gfc, max, avg, mask_idx_l);
/*********************************************************************
* convolve the partitioned energy and unpredictability with the
* spreading function, s3_l[b][k]
********************************************************************/
int k = 0;
for (b = 0; b < gfc.npart_l; b++) {
float x, ecb, avg_mask, t;
/* convolve the partitioned energy with the spreading function */
int kk = gfc.s3ind[b][0];
int last = gfc.s3ind[b][1];
int dd = 0, dd_n = 0;
dd = mask_idx_l[kk];
dd_n += 1;
ecb = gfc.s3_ll[k] * eb_l[kk] * tab[mask_idx_l[kk]];
++k;
++kk;
while (kk <= last) {
dd += mask_idx_l[kk];
dd_n += 1;
x = gfc.s3_ll[k] * eb_l[kk] * tab[mask_idx_l[kk]];
t = vbrpsy_mask_add(ecb, x, kk - b);
ecb = t;
++k;
++kk;
}
dd = (1 + 2 * dd) / (2 * dd_n);
avg_mask = tab[dd] * 0.5f;
ecb *= avg_mask;
/**** long block pre-echo control ****/
/**
*
* dont use long block pre-echo control if previous granule was
* a short block. This is to avoid the situation:
* frame0: quiet (very low masking)
* frame1: surge (triggers short blocks)
* frame2: regular frame. looks like pre-echo when compared to
* frame0, but all pre-echo was in frame1.
*
*/
/*
* chn=0,1 L and R channels chn=2,3 S and M channels.
*/
if (gfc.blocktype_old[chn & 0x01] == Encoder.SHORT_TYPE) {
float ecb_limit = rpelev * gfc.nb_1[chn][b];
if (ecb_limit > 0) {
thr[b] = Math.min(ecb, ecb_limit);
} else {
/**
*
* Robert 071209:
* Because we don't calculate long block psy when we know a granule
* should be of short blocks, we don't have any clue how the granule
* before would have looked like as a long block. So we have to guess
* a little bit for this END_TYPE block.
* Most of the time we get away with this sloppyness. (fingers crossed :)
* The speed increase is worth it.
*
*/
thr[b] = Math.min(ecb, eb_l[b] * NS_PREECHO_ATT2);
}
} else {
float ecb_limit_2 = rpelev2 * gfc.nb_2[chn][b];
float ecb_limit_1 = rpelev * gfc.nb_1[chn][b];
float ecb_limit;
if (ecb_limit_2 <= 0) {
ecb_limit_2 = ecb;
}
if (ecb_limit_1 <= 0) {
ecb_limit_1 = ecb;
}
if (gfc.blocktype_old[chn & 0x01] == Encoder.NORM_TYPE) {
ecb_limit = Math.min(ecb_limit_1, ecb_limit_2);
} else {
ecb_limit = ecb_limit_1;
}
thr[b] = Math.min(ecb, ecb_limit);
}
gfc.nb_2[chn][b] = gfc.nb_1[chn][b];
gfc.nb_1[chn][b] = ecb;
{
/*
* if THR exceeds EB, the quantization routines will take the
* difference from other bands. in case of strong tonal samples
* (tonaltest.wav) this leads to heavy distortions. that's why
* we limit THR here.
*/
x = max[b];
x *= gfc.minval_l[b];
x *= avg_mask;
if (thr[b] > x) {
thr[b] = x;
}
}
if (gfc.masking_lower > 1) {
thr[b] *= gfc.masking_lower;
}
if (thr[b] > eb_l[b]) {
thr[b] = eb_l[b];
}
if (gfc.masking_lower < 1) {
thr[b] *= gfc.masking_lower;
}
assert (thr[b] >= 0);
}
for (; b < Encoder.CBANDS; ++b) {
eb_l[b] = 0;
thr[b] = 0;
}
}
private void vbrpsy_compute_block_type(final LameGlobalFlags gfp,
final int[] uselongblock) {
final LameInternalFlags gfc = gfp.internal_flags;
if (gfp.short_blocks == ShortBlock.short_block_coupled
/* force both channels to use the same block type */
/* this is necessary if the frame is to be encoded in ms_stereo. */
/* But even without ms_stereo, FhG does this */
&& !(uselongblock[0] != 0 && uselongblock[1] != 0))
uselongblock[0] = uselongblock[1] = 0;
for (int chn = 0; chn < gfc.channels_out; chn++) {
/* disable short blocks */
if (gfp.short_blocks == ShortBlock.short_block_dispensed) {
uselongblock[chn] = 1;
}
if (gfp.short_blocks == ShortBlock.short_block_forced) {
uselongblock[chn] = 0;
}
}
}
private void vbrpsy_apply_block_type(final LameGlobalFlags gfp,
final int[] uselongblock, final int[] blocktype_d) {
final LameInternalFlags gfc = gfp.internal_flags;
/*
* update the blocktype of the previous granule, since it depends on
* what happend in this granule
*/
for (int chn = 0; chn < gfc.channels_out; chn++) {
int blocktype = Encoder.NORM_TYPE;
/* disable short blocks */
if (uselongblock[chn] != 0) {
/* no attack : use long blocks */
assert (gfc.blocktype_old[chn] != Encoder.START_TYPE);
if (gfc.blocktype_old[chn] == Encoder.SHORT_TYPE)
blocktype = Encoder.STOP_TYPE;
} else {
/* attack : use short blocks */
blocktype = Encoder.SHORT_TYPE;
if (gfc.blocktype_old[chn] == Encoder.NORM_TYPE) {
gfc.blocktype_old[chn] = Encoder.START_TYPE;
}
if (gfc.blocktype_old[chn] == Encoder.STOP_TYPE)
gfc.blocktype_old[chn] = Encoder.SHORT_TYPE;
}
blocktype_d[chn] = gfc.blocktype_old[chn];
// value returned to calling program
gfc.blocktype_old[chn] = blocktype;
// save for next call to l3psy_anal
}
}
/**
* compute M/S thresholds from Johnston & Ferreira 1992 ICASSP paper
*/
private void vbrpsy_compute_MS_thresholds(final float eb[][],
final float thr[][], final float cb_mld[], final float ath_cb[],
final float athadjust, final float msfix, final int n) {
float msfix2 = msfix * 2;
float athlower = msfix > 0 ? (float) Math.pow(10, athadjust) : 1f;
float rside, rmid;
for (int b = 0; b < n; ++b) {
float ebM = eb[2][b];
float ebS = eb[3][b];
float thmL = thr[0][b];
float thmR = thr[1][b];
float thmM = thr[2][b];
float thmS = thr[3][b];
/* use this fix if L & R masking differs by 2db or less */
if (thmL <= 1.58 * thmR && thmR <= 1.58 * thmL) {
float mld_m = cb_mld[b] * ebS;
float mld_s = cb_mld[b] * ebM;
rmid = Math.max(thmM, Math.min(thmS, mld_m));
rside = Math.max(thmS, Math.min(thmM, mld_s));
} else {
rmid = thmM;
rside = thmS;
}
if (msfix > 0) {
/***************************************************************/
/* Adjust M/S maskings if user set "msfix" */
/***************************************************************/
/* Naoki Shibata 2000 */
float thmLR, thmMS;
float ath = ath_cb[b] * athlower;
thmLR = Math.min(Math.max(thmL, ath), Math.max(thmR, ath));
thmM = Math.max(rmid, ath);
thmS = Math.max(rside, ath);
thmMS = thmM + thmS;
if (thmMS > 0 && (thmLR * msfix2) < thmMS) {
float f = thmLR * msfix2 / thmMS;
thmM *= f;
thmS *= f;
assert (thmMS > 0);
}
rmid = Math.min(thmM, rmid);
rside = Math.min(thmS, rside);
}
if (rmid > ebM) {
rmid = ebM;
}
if (rside > ebS) {
rside = ebS;
}
thr[2][b] = rmid;
thr[3][b] = rside;
}
}
public final int L3psycho_anal_vbr(final LameGlobalFlags gfp,
final float[] buffer[], final int bufPos, final int gr_out,
final III_psy_ratio masking_ratio[][],
final III_psy_ratio masking_MS_ratio[][],
final float percep_entropy[], final float percep_MS_entropy[],
final float energy[], final int blocktype_d[]) {
final LameInternalFlags gfc = gfp.internal_flags;
/* fft and energy calculation */
float wsamp_l[][];
float wsamp_s[][][];
float fftenergy[] = new float[Encoder.HBLKSIZE];
float fftenergy_s[][] = new float[3][Encoder.HBLKSIZE_s];
float wsamp_L[][] = new float[2][Encoder.BLKSIZE];
float wsamp_S[][][] = new float[2][3][Encoder.BLKSIZE_s];
float eb[][] = new float[4][Encoder.CBANDS], thr[][] = new float[4][Encoder.CBANDS];
float sub_short_factor[][] = new float[4][3];
float pcfact = 0.6f;
/* block type */
int ns_attacks[][] = { { 0, 0, 0, 0 }, { 0, 0, 0, 0 }, { 0, 0, 0, 0 },
{ 0, 0, 0, 0 } };
int uselongblock[] = new int[2];
/* usual variables like loop indices, etc.. */
/* chn=2 and 3 = Mid and Side channels */
int n_chn_psy = (gfp.mode == MPEGMode.JOINT_STEREO) ? 4
: gfc.channels_out;
vbrpsy_attack_detection(gfp, buffer, bufPos, gr_out, masking_ratio,
masking_MS_ratio, energy, sub_short_factor, ns_attacks,
uselongblock);
vbrpsy_compute_block_type(gfp, uselongblock);
/* LONG BLOCK CASE */
{
for (int chn = 0; chn < n_chn_psy; chn++) {
int ch01 = chn & 0x01;
wsamp_l = wsamp_L;
vbrpsy_compute_fft_l(gfp, buffer, bufPos, chn, gr_out,
fftenergy, wsamp_l, ch01);
vbrpsy_compute_loudness_approximation_l(gfp, gr_out, chn,
fftenergy);
if (uselongblock[ch01] != 0) {
vbrpsy_compute_masking_l(gfc, fftenergy, eb[chn], thr[chn],
chn);
} else {
vbrpsy_skip_masking_l(gfc, chn);
}
}
if ((uselongblock[0] + uselongblock[1]) == 2) {
/* M/S channel */
if (gfp.mode == MPEGMode.JOINT_STEREO) {
vbrpsy_compute_MS_thresholds(eb, thr, gfc.mld_cb_l,
gfc.ATH.cb_l, gfp.ATHlower * gfc.ATH.adjust,
gfp.msfix, gfc.npart_l);
}
}
/* TODO: apply adaptive ATH masking here ?? */
for (int chn = 0; chn < n_chn_psy; chn++) {
int ch01 = chn & 0x01;
if (uselongblock[ch01] != 0) {
convert_partition2scalefac_l(gfc, eb[chn], thr[chn], chn);
}
}
}
/* SHORT BLOCKS CASE */
{
for (int sblock = 0; sblock < 3; sblock++) {
for (int chn = 0; chn < n_chn_psy; ++chn) {
int ch01 = chn & 0x01;
if (uselongblock[ch01] != 0) {
vbrpsy_skip_masking_s(gfc, chn, sblock);
} else {
/* compute masking thresholds for short blocks */
wsamp_s = wsamp_S;
vbrpsy_compute_fft_s(gfp, buffer, bufPos, chn, sblock,
fftenergy_s, wsamp_s, ch01);
vbrpsy_compute_masking_s(gfp, fftenergy_s, eb[chn],
thr[chn], chn, sblock);
}
}
if ((uselongblock[0] + uselongblock[1]) == 0) {
/* M/S channel */
if (gfp.mode == MPEGMode.JOINT_STEREO) {
vbrpsy_compute_MS_thresholds(eb, thr, gfc.mld_cb_s,
gfc.ATH.cb_s, gfp.ATHlower * gfc.ATH.adjust,
gfp.msfix, gfc.npart_s);
}
/* L/R channel */
}
/* TODO: apply adaptive ATH masking here ?? */
for (int chn = 0; chn < n_chn_psy; ++chn) {
int ch01 = chn & 0x01;
if (0 == uselongblock[ch01]) {
convert_partition2scalefac_s(gfc, eb[chn], thr[chn],
chn, sblock);
}
}
}
/**** short block pre-echo control ****/
for (int chn = 0; chn < n_chn_psy; chn++) {
int ch01 = chn & 0x01;
if (uselongblock[ch01] != 0) {
continue;
}
for (int sb = 0; sb < Encoder.SBMAX_s; sb++) {
float new_thmm[] = new float[3];
for (int sblock = 0; sblock < 3; sblock++) {
float thmm = gfc.thm[chn].s[sb][sblock];
thmm *= NS_PREECHO_ATT0;
if (ns_attacks[chn][sblock] >= 2
|| ns_attacks[chn][sblock + 1] == 1) {
int idx = (sblock != 0) ? sblock - 1 : 2;
double p = NS_INTERP(gfc.thm[chn].s[sb][idx], thmm,
NS_PREECHO_ATT1 * pcfact);
thmm = (float) Math.min(thmm, p);
} else if (ns_attacks[chn][sblock] == 1) {
int idx = (sblock != 0) ? sblock - 1 : 2;
double p = NS_INTERP(gfc.thm[chn].s[sb][idx], thmm,
NS_PREECHO_ATT2 * pcfact);
thmm = (float) Math.min(thmm, p);
} else if ((sblock != 0 && ns_attacks[chn][sblock - 1] == 3)
|| (sblock == 0 && gfc.nsPsy.lastAttacks[chn] == 3)) {
int idx = (sblock != 2) ? sblock + 1 : 0;
double p = NS_INTERP(gfc.thm[chn].s[sb][idx], thmm,
NS_PREECHO_ATT2 * pcfact);
thmm = (float) Math.min(thmm, p);
}
/* pulse like signal detection for fatboy.wav and so on */
thmm *= sub_short_factor[chn][sblock];
new_thmm[sblock] = thmm;
}
for (int sblock = 0; sblock < 3; sblock++) {
gfc.thm[chn].s[sb][sblock] = new_thmm[sblock];
}
}
}
}
for (int chn = 0; chn < n_chn_psy; chn++) {
gfc.nsPsy.lastAttacks[chn] = ns_attacks[chn][2];
}
/***************************************************************
* determine final block type
***************************************************************/
vbrpsy_apply_block_type(gfp, uselongblock, blocktype_d);
/*********************************************************************
* compute the value of PE to return ... no delay and advance
*********************************************************************/
for (int chn = 0; chn < n_chn_psy; chn++) {
float[] ppe;
int ppePos;
int type;
III_psy_ratio mr;
if (chn > 1) {
ppe = percep_MS_entropy;
ppePos = -2;
type = Encoder.NORM_TYPE;
if (blocktype_d[0] == Encoder.SHORT_TYPE
|| blocktype_d[1] == Encoder.SHORT_TYPE)
type = Encoder.SHORT_TYPE;
mr = masking_MS_ratio[gr_out][chn - 2];
} else {
ppe = percep_entropy;
ppePos = 0;
type = blocktype_d[chn];
mr = masking_ratio[gr_out][chn];
}
if (type == Encoder.SHORT_TYPE) {
ppe[ppePos + chn] = pecalc_s(mr, gfc.masking_lower);
} else {
ppe[ppePos + chn] = pecalc_l(mr, gfc.masking_lower);
}
if (gfp.analysis) {
gfc.pinfo.pe[gr_out][chn] = ppe[ppePos + chn];
}
}
return 0;
}
private float s3_func_x(final float bark, final float hf_slope) {
float tempx = bark, tempy;
if (tempx >= 0) {
tempy = -tempx * 27;
} else {
tempy = tempx * hf_slope;
}
if (tempy <= -72.0) {
return 0;
}
return (float) Math.exp(tempy * LN_TO_LOG10);
}
private float norm_s3_func_x(final float hf_slope) {
double lim_a = 0, lim_b = 0;
{
double x = 0, l, h;
for (x = 0; s3_func_x((float) x, hf_slope) > 1e-20; x -= 1)
;
l = x;
h = 0;
while (Math.abs(h - l) > 1e-12) {
x = (h + l) / 2;
if (s3_func_x((float) x, hf_slope) > 0) {
h = x;
} else {
l = x;
}
}
lim_a = l;
}
{
double x = 0, l, h;
for (x = 0; s3_func_x((float) x, hf_slope) > 1e-20; x += 1)
;
l = 0;
h = x;
while (Math.abs(h - l) > 1e-12) {
x = (h + l) / 2;
if (s3_func_x((float) x, hf_slope) > 0) {
l = x;
} else {
h = x;
}
}
lim_b = h;
}
{
double sum = 0;
final int m = 1000;
int i;
for (i = 0; i <= m; ++i) {
double x = lim_a + i * (lim_b - lim_a) / m;
double y = s3_func_x((float) x, hf_slope);
sum += y;
}
{
double norm = (m + 1) / (sum * (lim_b - lim_a));
/* printf( "norm = %lf\n",norm); */
return (float) norm;
}
}
}
/**
* The spreading function. Values returned in units of energy
*/
private float s3_func(final float bark) {
float tempx, x, tempy, temp;
tempx = bark;
if (tempx >= 0)
tempx *= 3;
else
tempx *= 1.5;
if (tempx >= 0.5 && tempx <= 2.5) {
temp = tempx - 0.5f;
x = 8.0f * (temp * temp - 2.0f * temp);
} else
x = 0.0f;
tempx += 0.474;
tempy = 15.811389f + 7.5f * tempx - 17.5f
* (float) Math.sqrt(1.0 + tempx * tempx);
if (tempy <= -60.0)
return 0.0f;
tempx = (float) Math.exp((x + tempy) * LN_TO_LOG10);
/**
*
* Normalization. The spreading function should be normalized so that:
* +inf
* /
* | s3 [ bark ] d(bark) = 1
* /
* -inf
*
*/
tempx /= .6609193;
return tempx;
}
/**
* see for example "Zwicker: Psychoakustik, 1982; ISBN 3-540-11401-7
*/
private float freq2bark(float freq) {
/* input: freq in hz output: barks */
if (freq < 0)
freq = 0;
freq = freq * 0.001f;
return 13.0f * (float) Math.atan(.76 * freq) + 3.5f
* (float) Math.atan(freq * freq / (7.5 * 7.5));
}
private int init_numline(final int[] numlines, final int[] bo,
final int[] bm, final float[] bval, final float[] bval_width,
final float[] mld, final float[] bo_w, float sfreq,
final int blksize, final int[] scalepos, final float deltafreq,
final int sbmax) {
float b_frq[] = new float[Encoder.CBANDS + 1];
float sample_freq_frac = sfreq / (sbmax > 15 ? 2 * 576 : 2 * 192);
int partition[] = new int[Encoder.HBLKSIZE];
int i;
sfreq /= blksize;
int j = 0;
int ni = 0;
/* compute numlines, the number of spectral lines in each partition band */
/* each partition band should be about DELBARK wide. */
for (i = 0; i < Encoder.CBANDS; i++) {
float bark1;
int j2;
bark1 = freq2bark(sfreq * j);
b_frq[i] = sfreq * j;
for (j2 = j; freq2bark(sfreq * j2) - bark1 < DELBARK
&& j2 <= blksize / 2; j2++)
;
numlines[i] = j2 - j;
ni = i + 1;
while (j < j2) {
assert (j < Encoder.HBLKSIZE);
partition[j++] = i;
}
if (j > blksize / 2) {
j = blksize / 2;
++i;
break;
}
}
assert (i < Encoder.CBANDS);
b_frq[i] = sfreq * j;
for (int sfb = 0; sfb < sbmax; sfb++) {
int i1, i2, start, end;
float arg;
start = scalepos[sfb];
end = scalepos[sfb + 1];
i1 = (int) Math.floor(.5 + deltafreq * (start - .5));
if (i1 < 0)
i1 = 0;
i2 = (int) Math.floor(.5 + deltafreq * (end - .5));
if (i2 > blksize / 2)
i2 = blksize / 2;
bm[sfb] = (partition[i1] + partition[i2]) / 2;
bo[sfb] = partition[i2];
float f_tmp = sample_freq_frac * end;
/*
* calculate how much of this band belongs to current scalefactor
* band
*/
bo_w[sfb] = (f_tmp - b_frq[bo[sfb]])
/ (b_frq[bo[sfb] + 1] - b_frq[bo[sfb]]);
if (bo_w[sfb] < 0) {
bo_w[sfb] = 0;
} else {
if (bo_w[sfb] > 1) {
bo_w[sfb] = 1;
}
}
/* setup stereo demasking thresholds */
/* formula reverse enginerred from plot in paper */
arg = freq2bark(sfreq * scalepos[sfb] * deltafreq);
arg = ((float) Math.min(arg, 15.5) / 15.5f);
mld[sfb] = (float) Math.pow(10.0,
1.25 * (1 - Math.cos(Math.PI * arg)) - 2.5);
}
/* compute bark values of each critical band */
j = 0;
for (int k = 0; k < ni; k++) {
final int w = numlines[k];
float bark1, bark2;
bark1 = freq2bark(sfreq * (j));
bark2 = freq2bark(sfreq * (j + w - 1));
bval[k] = .5f * (bark1 + bark2);
bark1 = freq2bark(sfreq * (j - .5f));
bark2 = freq2bark(sfreq * (j + w - .5f));
bval_width[k] = bark2 - bark1;
j += w;
}
return ni;
}
private float[] init_s3_values(final int s3ind[][], final int npart,
final float[] bval, final float[] bval_width, final float[] norm,
final boolean use_old_s3) {
float s3[][] = new float[Encoder.CBANDS][Encoder.CBANDS];
/*
* The s3 array is not linear in the bark scale.
*
* bval[x] should be used to get the bark value.
*/
int j;
int numberOfNoneZero = 0;
/**
*
* s[i][j], the value of the spreading function,
* centered at band j (masker), for band i (maskee)
*
* i.e.: sum over j to spread into signal barkval=i
* NOTE: i and j are used opposite as in the ISO docs
*
*/
if (use_old_s3) {
for (int i = 0; i < npart; i++) {
for (j = 0; j < npart; j++) {
float v = s3_func(bval[i] - bval[j]) * bval_width[j];
s3[i][j] = v * norm[i];
}
}
} else {
for (j = 0; j < npart; j++) {
float hf_slope = 15 + Math.min(21 / bval[j], 12);
float s3_x_norm = norm_s3_func_x(hf_slope);
for (int i = 0; i < npart; i++) {
float v = s3_x_norm
* s3_func_x(bval[i] - bval[j], hf_slope)
* bval_width[j];
s3[i][j] = v * norm[i];
}
}
}
for (int i = 0; i < npart; i++) {
for (j = 0; j < npart; j++) {
if (s3[i][j] > 0.0f)
break;
}
s3ind[i][0] = j;
for (j = npart - 1; j > 0; j--) {
if (s3[i][j] > 0.0f)
break;
}
s3ind[i][1] = j;
numberOfNoneZero += (s3ind[i][1] - s3ind[i][0] + 1);
}
float[] p = new float[numberOfNoneZero];
int k = 0;
for (int i = 0; i < npart; i++)
for (j = s3ind[i][0]; j <= s3ind[i][1]; j++)
p[k++] = s3[i][j];
return p;
}
private float stereo_demask(double f) {
/* setup stereo demasking thresholds */
/* formula reverse enginerred from plot in paper */
double arg = freq2bark((float) f);
arg = (Math.min(arg, 15.5) / 15.5);
return (float) Math.pow(10.0,
1.25 * (1 - Math.cos(Math.PI * arg)) - 2.5);
}
/**
* NOTE: the bitrate reduction from the inter-channel masking effect is low
* compared to the chance of getting annyoing artefacts. L3psycho_anal_vbr
* does not use this feature. (Robert 071216)
*/
public final int psymodel_init(final LameGlobalFlags gfp) {
final LameInternalFlags gfc = gfp.internal_flags;
int i;
boolean useOldS3 = true;
float bvl_a = 13, bvl_b = 24;
float snr_l_a = 0, snr_l_b = 0;
float snr_s_a = -8.25f, snr_s_b = -4.5f;
float bval[] = new float[Encoder.CBANDS];
float bval_width[] = new float[Encoder.CBANDS];
float norm[] = new float[Encoder.CBANDS];
final float sfreq = gfp.out_samplerate;
switch (gfp.experimentalZ) {
default:
case 0:
useOldS3 = true;
break;
case 1:
useOldS3 = (gfp.VBR == VbrMode.vbr_mtrh || gfp.VBR == VbrMode.vbr_mt) ? false
: true;
break;
case 2:
useOldS3 = false;
break;
case 3:
bvl_a = 8;
snr_l_a = -1.75f;
snr_l_b = -0.0125f;
snr_s_a = -8.25f;
snr_s_b = -2.25f;
break;
}
gfc.ms_ener_ratio_old = .25f;
gfc.blocktype_old[0] = gfc.blocktype_old[1] = Encoder.NORM_TYPE;
// the vbr header is long blocks
for (i = 0; i < 4; ++i) {
for (int j = 0; j < Encoder.CBANDS; ++j) {
gfc.nb_1[i][j] = 1e20f;
gfc.nb_2[i][j] = 1e20f;
gfc.nb_s1[i][j] = gfc.nb_s2[i][j] = 1.0f;
}
for (int sb = 0; sb < Encoder.SBMAX_l; sb++) {
gfc.en[i].l[sb] = 1e20f;
gfc.thm[i].l[sb] = 1e20f;
}
for (int j = 0; j < 3; ++j) {
for (int sb = 0; sb < Encoder.SBMAX_s; sb++) {
gfc.en[i].s[sb][j] = 1e20f;
gfc.thm[i].s[sb][j] = 1e20f;
}
gfc.nsPsy.lastAttacks[i] = 0;
}
for (int j = 0; j < 9; j++)
gfc.nsPsy.last_en_subshort[i][j] = 10.f;
}
/* init. for loudness approx. -jd 2001 mar 27 */
gfc.loudness_sq_save[0] = gfc.loudness_sq_save[1] = 0.0f;
/*************************************************************************
* now compute the psychoacoustic model specific constants
************************************************************************/
/* compute numlines, bo, bm, bval, bval_width, mld */
gfc.npart_l = init_numline(gfc.numlines_l, gfc.bo_l, gfc.bm_l, bval,
bval_width, gfc.mld_l, gfc.PSY.bo_l_weight, sfreq,
Encoder.BLKSIZE, gfc.scalefac_band.l, Encoder.BLKSIZE
/ (2.0f * 576f), Encoder.SBMAX_l);
assert (gfc.npart_l < Encoder.CBANDS);
/* compute the spreading function */
for (i = 0; i < gfc.npart_l; i++) {
double snr = snr_l_a;
if (bval[i] >= bvl_a) {
snr = snr_l_b * (bval[i] - bvl_a) / (bvl_b - bvl_a) + snr_l_a
* (bvl_b - bval[i]) / (bvl_b - bvl_a);
}
norm[i] = (float) Math.pow(10.0, snr / 10.0);
if (gfc.numlines_l[i] > 0) {
gfc.rnumlines_l[i] = 1.0f / gfc.numlines_l[i];
} else {
gfc.rnumlines_l[i] = 0;
}
}
gfc.s3_ll = init_s3_values(gfc.s3ind, gfc.npart_l, bval, bval_width,
norm, useOldS3);
/* compute long block specific values, ATH and MINVAL */
int j = 0;
for (i = 0; i < gfc.npart_l; i++) {
double x;
/* ATH */
x = Float.MAX_VALUE;
for (int k = 0; k < gfc.numlines_l[i]; k++, j++) {
final float freq = sfreq * j / (1000.0f * Encoder.BLKSIZE);
float level;
/*
* ATH below 100 Hz constant, not further climbing
*/
level = ATHformula(freq * 1000, gfp) - 20;
// scale to FFT units; returned value is in dB
level = (float) Math.pow(10., 0.1 * level);
// convert from dB . energy
level *= gfc.numlines_l[i];
if (x > level)
x = level;
}
gfc.ATH.cb_l[i] = (float) x;
/*
* MINVAL. For low freq, the strength of the masking is limited by
* minval this is an ISO MPEG1 thing, dont know if it is really
* needed
*/
/*
* FIXME: it does work to reduce low-freq problems in S53-Wind-Sax
* and lead-voice samples, but introduces some 3 kbps bit bloat too.
* TODO: Further refinement of the shape of this hack.
*/
x = -20 + bval[i] * 20 / 10;
if (x > 6) {
x = 100;
}
if (x < -15) {
x = -15;
}
x -= 8.;
gfc.minval_l[i] = (float) (Math.pow(10.0, x / 10.) * gfc.numlines_l[i]);
}
/************************************************************************
* do the same things for short blocks
************************************************************************/
gfc.npart_s = init_numline(gfc.numlines_s, gfc.bo_s, gfc.bm_s, bval,
bval_width, gfc.mld_s, gfc.PSY.bo_s_weight, sfreq,
Encoder.BLKSIZE_s, gfc.scalefac_band.s, Encoder.BLKSIZE_s
/ (2.0f * 192), Encoder.SBMAX_s);
assert (gfc.npart_s < Encoder.CBANDS);
/* SNR formula. short block is normalized by SNR. is it still right ? */
j = 0;
for (i = 0; i < gfc.npart_s; i++) {
double x;
double snr = snr_s_a;
if (bval[i] >= bvl_a) {
snr = snr_s_b * (bval[i] - bvl_a) / (bvl_b - bvl_a) + snr_s_a
* (bvl_b - bval[i]) / (bvl_b - bvl_a);
}
norm[i] = (float) Math.pow(10.0, snr / 10.0);
/* ATH */
x = Float.MAX_VALUE;
for (int k = 0; k < gfc.numlines_s[i]; k++, j++) {
final float freq = sfreq * j / (1000.0f * Encoder.BLKSIZE_s);
float level;
/* freq = Min(.1,freq); *//*
* ATH below 100 Hz constant, not
* further climbing
*/
level = ATHformula(freq * 1000, gfp) - 20;
// scale to FFT units; returned value is in dB
level = (float) Math.pow(10., 0.1 * level);
// convert from dB . energy
level *= gfc.numlines_s[i];
if (x > level)
x = level;
}
gfc.ATH.cb_s[i] = (float) x;
/*
* MINVAL. For low freq, the strength of the masking is limited by
* minval this is an ISO MPEG1 thing, dont know if it is really
* needed
*/
x = (-7.0 + bval[i] * 7.0 / 12.0);
if (bval[i] > 12) {
x *= 1 + Math.log(1 + x) * 3.1;
}
if (bval[i] < 12) {
x *= 1 + Math.log(1 - x) * 2.3;
}
if (x < -15) {
x = -15;
}
x -= 8;
gfc.minval_s[i] = (float) Math.pow(10.0, x / 10)
* gfc.numlines_s[i];
}
gfc.s3_ss = init_s3_values(gfc.s3ind_s, gfc.npart_s, bval, bval_width,
norm, useOldS3);
init_mask_add_max_values();
fft.init_fft(gfc);
/* setup temporal masking */
gfc.decay = (float) Math.exp(-1.0 * LOG10
/ (temporalmask_sustain_sec * sfreq / 192.0));
{
float msfix;
msfix = NS_MSFIX;
if ((gfp.exp_nspsytune & 2) != 0)
msfix = 1.0f;
if (Math.abs(gfp.msfix) > 0.0)
msfix = gfp.msfix;
gfp.msfix = msfix;
/*
* spread only from npart_l bands. Normally, we use the spreading
* function to convolve from npart_l down to npart_l bands
*/
for (int b = 0; b < gfc.npart_l; b++)
if (gfc.s3ind[b][1] > gfc.npart_l - 1)
gfc.s3ind[b][1] = gfc.npart_l - 1;
}
/*
* prepare for ATH auto adjustment: we want to decrease the ATH by 12 dB
* per second
*/
float frame_duration = (576.f * gfc.mode_gr / sfreq);
gfc.ATH.decay = (float) Math.pow(10., -12. / 10. * frame_duration);
gfc.ATH.adjust = 0.01f; /* minimum, for leading low loudness */
gfc.ATH.adjustLimit = 1.0f; /* on lead, allow adjust up to maximum */
assert (gfc.bo_l[Encoder.SBMAX_l - 1] <= gfc.npart_l);
assert (gfc.bo_s[Encoder.SBMAX_s - 1] <= gfc.npart_s);
if (gfp.ATHtype != -1) {
/* compute equal loudness weights (eql_w) */
float freq;
final float freq_inc = (float) gfp.out_samplerate
/ (float) (Encoder.BLKSIZE);
float eql_balance = 0.0f;
freq = 0.0f;
for (i = 0; i < Encoder.BLKSIZE / 2; ++i) {
/* convert ATH dB to relative power (not dB) */
/* to determine eql_w */
freq += freq_inc;
gfc.ATH.eql_w[i] = 1.f / (float) Math.pow(10,
ATHformula(freq, gfp) / 10);
eql_balance += gfc.ATH.eql_w[i];
}
eql_balance = 1.0f / eql_balance;
for (i = Encoder.BLKSIZE / 2; --i >= 0;) { /* scale weights */
gfc.ATH.eql_w[i] *= eql_balance;
}
}
{
for (int b = j = 0; b < gfc.npart_s; ++b) {
for (i = 0; i < gfc.numlines_s[b]; ++i) {
++j;
}
}
assert (j == 129);
for (int b = j = 0; b < gfc.npart_l; ++b) {
for (i = 0; i < gfc.numlines_l[b]; ++i) {
++j;
}
}
assert (j == 513);
}
j = 0;
for (i = 0; i < gfc.npart_l; i++) {
final float freq = sfreq * (j + gfc.numlines_l[i] / 2)
/ (1.0f * Encoder.BLKSIZE);
gfc.mld_cb_l[i] = stereo_demask(freq);
j += gfc.numlines_l[i];
}
for (; i < Encoder.CBANDS; ++i) {
gfc.mld_cb_l[i] = 1;
}
j = 0;
for (i = 0; i < gfc.npart_s; i++) {
final float freq = sfreq * (j + gfc.numlines_s[i] / 2)
/ (1.0f * Encoder.BLKSIZE_s);
gfc.mld_cb_s[i] = stereo_demask(freq);
j += gfc.numlines_s[i];
}
for (; i < Encoder.CBANDS; ++i) {
gfc.mld_cb_s[i] = 1;
}
return 0;
}
/**
* Those ATH formulas are returning their minimum value for input = -1
*/
private float ATHformula_GB(float f, final float value) {
/**
*
* from Painter & Spanias
* modified by Gabriel Bouvigne to better fit the reality
* ath = 3.640 * pow(f,-0.8)
* - 6.800 * exp(-0.6*pow(f-3.4,2.0))
* + 6.000 * exp(-0.15*pow(f-8.7,2.0))
* + 0.6* 0.001 * pow(f,4.0);
*
*
* In the past LAME was using the Painter &Spanias formula.
* But we had some recurrent problems with HF content.
* We measured real ATH values, and found the older formula
* to be inaccurate in the higher part. So we made this new
* formula and this solved most of HF problematic test cases.
* The tradeoff is that in VBR mode it increases a lot the
* bitrate.
*
*/
/*
* This curve can be adjusted according to the VBR scale: it adjusts
* from something close to Painter & Spanias on V9 up to Bouvigne's
* formula for V0. This way the VBR bitrate is more balanced according
* to the -V value.
*/
// the following Hack allows to ask for the lowest value
if (f < -.3)
f = 3410;
// convert to khz
f /= 1000;
f = (float) Math.max(0.1, f);
float ath = 3.640f * (float) Math.pow(f, -0.8) - 6.800f
* (float) Math.exp(-0.6 * Math.pow(f - 3.4, 2.0)) + 6.000f
* (float) Math.exp(-0.15 * Math.pow(f - 8.7, 2.0))
+ (0.6f + 0.04f * value) * 0.001f * (float) Math.pow(f, 4.0);
return ath;
}
public final float ATHformula(final float f, final LameGlobalFlags gfp) {
float ath;
switch (gfp.ATHtype) {
case 0:
ath = ATHformula_GB(f, 9);
break;
case 1:
// over sensitive, should probably be removed
ath = ATHformula_GB(f, -1);
break;
case 2:
ath = ATHformula_GB(f, 0);
break;
case 3:
// modification of GB formula by Roel
ath = ATHformula_GB(f, 1) + 6;
break;
case 4:
ath = ATHformula_GB(f, gfp.ATHcurve);
break;
default:
ath = ATHformula_GB(f, 0);
break;
}
return ath;
}
}