1 /*********************************************************************** 2 Copyright (c) 2006-2011, Skype Limited. All rights reserved. 3 Redistribution and use in source and binary forms, with or without 4 modification, are permitted provided that the following conditions 5 are met: 6 - Redistributions of source code must retain the above copyright notice, 7 this list of conditions and the following disclaimer. 8 - Redistributions in binary form must reproduce the above copyright 9 notice, this list of conditions and the following disclaimer in the 10 documentation and/or other materials provided with the distribution. 11 - Neither the name of Internet Society, IETF or IETF Trust, nor the 12 names of specific contributors, may be used to endorse or promote 13 products derived from this software without specific prior written 14 permission. 15 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS AS IS 16 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 17 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 18 ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE 19 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 20 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 21 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 22 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 23 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 24 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 25 POSSIBILITY OF SUCH DAMAGE. 26 ***********************************************************************/ 27 28 #ifdef HAVE_CONFIG_H 29 #include "config.h" 30 #endif 31 32 #include "main_FIX.h" 33 #include "tuning_parameters.h" 34 35 /* Find pitch lags */ 36 void silk_find_pitch_lags_FIX( 37 silk_encoder_state_FIX *psEnc, /* I/O encoder state */ 38 silk_encoder_control_FIX *psEncCtrl, /* I/O encoder control */ 39 opus_int16 res[], /* O residual */ 40 const opus_int16 x[] /* I Speech signal */ 41 ) 42 { 43 opus_int buf_len, i, scale; 44 opus_int32 thrhld_Q15, res_nrg; 45 const opus_int16 *x_buf, *x_buf_ptr; 46 opus_int16 Wsig[ FIND_PITCH_LPC_WIN_MAX ], *Wsig_ptr; 47 opus_int32 auto_corr[ MAX_FIND_PITCH_LPC_ORDER + 1 ]; 48 opus_int16 rc_Q15[ MAX_FIND_PITCH_LPC_ORDER ]; 49 opus_int32 A_Q24[ MAX_FIND_PITCH_LPC_ORDER ]; 50 opus_int16 A_Q12[ MAX_FIND_PITCH_LPC_ORDER ]; 51 52 /******************************************/ 53 /* Set up buffer lengths etc based on Fs */ 54 /******************************************/ 55 buf_len = psEnc->sCmn.la_pitch + psEnc->sCmn.frame_length + psEnc->sCmn.ltp_mem_length; 56 57 /* Safety check */ 58 silk_assert( buf_len >= psEnc->sCmn.pitch_LPC_win_length ); 59 60 x_buf = x - psEnc->sCmn.ltp_mem_length; 61 62 /*************************************/ 63 /* Estimate LPC AR coefficients */ 64 /*************************************/ 65 66 /* Calculate windowed signal */ 67 68 /* First LA_LTP samples */ 69 x_buf_ptr = x_buf + buf_len - psEnc->sCmn.pitch_LPC_win_length; 70 Wsig_ptr = Wsig; 71 silk_apply_sine_window( Wsig_ptr, x_buf_ptr, 1, psEnc->sCmn.la_pitch ); 72 73 /* Middle un - windowed samples */ 74 Wsig_ptr += psEnc->sCmn.la_pitch; 75 x_buf_ptr += psEnc->sCmn.la_pitch; 76 silk_memcpy( Wsig_ptr, x_buf_ptr, ( psEnc->sCmn.pitch_LPC_win_length - silk_LSHIFT( psEnc->sCmn.la_pitch, 1 ) ) * sizeof( opus_int16 ) ); 77 78 /* Last LA_LTP samples */ 79 Wsig_ptr += psEnc->sCmn.pitch_LPC_win_length - silk_LSHIFT( psEnc->sCmn.la_pitch, 1 ); 80 x_buf_ptr += psEnc->sCmn.pitch_LPC_win_length - silk_LSHIFT( psEnc->sCmn.la_pitch, 1 ); 81 silk_apply_sine_window( Wsig_ptr, x_buf_ptr, 2, psEnc->sCmn.la_pitch ); 82 83 /* Calculate autocorrelation sequence */ 84 silk_autocorr( auto_corr, &scale, Wsig, psEnc->sCmn.pitch_LPC_win_length, psEnc->sCmn.pitchEstimationLPCOrder + 1 ); 85 86 /* Add white noise, as fraction of energy */ 87 auto_corr[ 0 ] = silk_SMLAWB( auto_corr[ 0 ], auto_corr[ 0 ], SILK_FIX_CONST( FIND_PITCH_WHITE_NOISE_FRACTION, 16 ) ) + 1; 88 89 /* Calculate the reflection coefficients using schur */ 90 res_nrg = silk_schur( rc_Q15, auto_corr, psEnc->sCmn.pitchEstimationLPCOrder ); 91 92 /* Prediction gain */ 93 psEncCtrl->predGain_Q16 = silk_DIV32_varQ( auto_corr[ 0 ], silk_max_int( res_nrg, 1 ), 16 ); 94 95 /* Convert reflection coefficients to prediction coefficients */ 96 silk_k2a( A_Q24, rc_Q15, psEnc->sCmn.pitchEstimationLPCOrder ); 97 98 /* Convert From 32 bit Q24 to 16 bit Q12 coefs */ 99 for( i = 0; i < psEnc->sCmn.pitchEstimationLPCOrder; i++ ) { 100 A_Q12[ i ] = (opus_int16)silk_SAT16( silk_RSHIFT( A_Q24[ i ], 12 ) ); 101 } 102 103 /* Do BWE */ 104 silk_bwexpander( A_Q12, psEnc->sCmn.pitchEstimationLPCOrder, SILK_FIX_CONST( FIND_PITCH_BANDWIDTH_EXPANSION, 16 ) ); 105 106 /*****************************************/ 107 /* LPC analysis filtering */ 108 /*****************************************/ 109 silk_LPC_analysis_filter( res, x_buf, A_Q12, buf_len, psEnc->sCmn.pitchEstimationLPCOrder ); 110 111 if( psEnc->sCmn.indices.signalType != TYPE_NO_VOICE_ACTIVITY && psEnc->sCmn.first_frame_after_reset == 0 ) { 112 /* Threshold for pitch estimator */ 113 thrhld_Q15 = SILK_FIX_CONST( 0.6, 15 ); 114 thrhld_Q15 = silk_SMLABB( thrhld_Q15, SILK_FIX_CONST( -0.004, 15 ), psEnc->sCmn.pitchEstimationLPCOrder ); 115 thrhld_Q15 = silk_SMLABB( thrhld_Q15, SILK_FIX_CONST( -0.1, 7 ), psEnc->sCmn.speech_activity_Q8 ); 116 thrhld_Q15 = silk_SMLABB( thrhld_Q15, SILK_FIX_CONST( -0.15, 15 ), silk_RSHIFT( psEnc->sCmn.prevSignalType, 1 ) ); 117 thrhld_Q15 = silk_SMLAWB( thrhld_Q15, SILK_FIX_CONST( -0.1, 16 ), psEnc->sCmn.input_tilt_Q15 ); 118 thrhld_Q15 = silk_SAT16( thrhld_Q15 ); 119 120 /*****************************************/ 121 /* Call pitch estimator */ 122 /*****************************************/ 123 if( silk_pitch_analysis_core( res, psEncCtrl->pitchL, &psEnc->sCmn.indices.lagIndex, &psEnc->sCmn.indices.contourIndex, 124 &psEnc->LTPCorr_Q15, psEnc->sCmn.prevLag, psEnc->sCmn.pitchEstimationThreshold_Q16, 125 (opus_int16)thrhld_Q15, psEnc->sCmn.fs_kHz, psEnc->sCmn.pitchEstimationComplexity, psEnc->sCmn.nb_subfr ) == 0 ) 126 { 127 psEnc->sCmn.indices.signalType = TYPE_VOICED; 128 } else { 129 psEnc->sCmn.indices.signalType = TYPE_UNVOICED; 130 } 131 } else { 132 silk_memset( psEncCtrl->pitchL, 0, sizeof( psEncCtrl->pitchL ) ); 133 psEnc->sCmn.indices.lagIndex = 0; 134 psEnc->sCmn.indices.contourIndex = 0; 135 psEnc->LTPCorr_Q15 = 0; 136 } 137 } 138
