shithub: opus

Info  •  Files  •  Log  •  Branches

Scaling back the pitch filter when most of the energy is above 3.2 kHz

That corresponds to the fundamental for the shortest pitch period allowed

--- a/celt/celt.h

+++ b/celt/celt.h

@@ -63,6 +63,7 @@

    float music_prob_max;

    int   bandwidth;

    float activity_probability;

+   float max_pitch_ratio;

    /* Store as Q6 char to save space. */

    unsigned char leak_boost[LEAK_BANDS];

 } AnalysisInfo;

--- a/celt/celt_encoder.c

+++ b/celt/celt_encoder.c

@@ -1109,7 +1109,7 @@

 static int run_prefilter(CELTEncoder *st, celt_sig *in, celt_sig *prefilter_mem, int CC, int N,

-      int prefilter_tapset, int *pitch, opus_val16 *gain, int *qgain, int enabled, int nbAvailableBytes)

+      int prefilter_tapset, int *pitch, opus_val16 *gain, int *qgain, int enabled, int nbAvailableBytes, AnalysisInfo *analysis)

 {

    int c;

    VARDECL(celt_sig, _pre);

@@ -1165,7 +1165,10 @@

       gain1 = 0;

       pitch_index = COMBFILTER_MINPERIOD;

    }

-

+#ifndef DISABLE_FLOAT_API

+   if (analysis->valid)

+      gain1 *= analysis->max_pitch_ratio;

+#endif

    /* Gain threshold for enabling the prefilter/postfilter */

    pf_threshold = QCONST16(.2f,15);

@@ -1603,7 +1606,7 @@

             && st->complexity >= 5;

       prefilter_tapset = st->tapset_decision;

-      pf_on = run_prefilter(st, in, prefilter_mem, CC, N, prefilter_tapset, &pitch_index, &gain1, &qg, enabled, nbAvailableBytes);

+      pf_on = run_prefilter(st, in, prefilter_mem, CC, N, prefilter_tapset, &pitch_index, &gain1, &qg, enabled, nbAvailableBytes, &st->analysis);

       if ((gain1 > QCONST16(.4f,15) || st->prefilter_gain > QCONST16(.4f,15)) && (!st->analysis.valid || st->analysis.tonality > .3)

             && (pitch_index > 1.26*st->prefilter_period || pitch_index < .79*st->prefilter_period))

          pitch_change = 1;

--- a/src/analysis.c

+++ b/src/analysis.c

@@ -446,6 +446,8 @@

     float leakage_from[NB_TBANDS+1];

     float leakage_to[NB_TBANDS+1];

     float layer_out[MAX_NEURONS];

+    float below_max_pitch;

+    float above_max_pitch;

     SAVE_STACK;

     alpha = 1.f/IMIN(10, 1+tonal->count);

@@ -722,6 +724,8 @@

     maxE = 0;

     noise_floor = 5.7e-4f/(1<<(IMAX(0,lsb_depth-8)));

     noise_floor *= noise_floor;

+    below_max_pitch=0;

+    above_max_pitch=0;

     for (b=0;b<NB_TBANDS;b++)

     {

        float E=0;

@@ -738,6 +742,12 @@

        }

        E = SCALE_ENER(E);

        maxE = MAX32(maxE, E);

+       if (band_start < 64)

+       {

+          below_max_pitch += E;

+       } else {

+          above_max_pitch += E;

+       }

        tonal->meanE[b] = MAX32((1-alphaE2)*tonal->meanE[b], E);

        Em = MAX32(E, tonal->meanE[b]);

        /* Consider the band "active" only if all these conditions are met:

@@ -767,6 +777,7 @@

        /* silk_resampler_down2_hp() shifted right by an extra 8 bits. */

        E *= 256.f*(1.f/Q15ONE)*(1.f/Q15ONE);

 #endif

+       above_max_pitch += E;

        tonal->meanE[b] = MAX32((1-alphaE2)*tonal->meanE[b], E);

        Em = MAX32(E, tonal->meanE[b]);

        if (Em > 3*noise_ratio*noise_floor*160 || E > noise_ratio*noise_floor*160)

@@ -774,6 +785,10 @@

        /* Check if the band is masked (see below). */

        is_masked[b] = E < (tonal->prev_bandwidth == 20  ? .01f : .05f)*bandwidth_mask;

     }

+    if (above_max_pitch > below_max_pitch)

+       info->max_pitch_ratio = below_max_pitch/above_max_pitch;

+    else

+       info->max_pitch_ratio = 1;

     /* In some cases, resampling aliasing can create a small amount of energy in the first band

        being cut. So if the last band is masked, we don't include it.  */

     if (bandwidth == 20 && is_masked[NB_TBANDS])