search for: lpc_to_lsp

Saju Palayur <saju_pa at hotmail.com> writes: > Hi > > > > We have ported the Narrow band 8 Khz Speex encoder and decoder on to TI DSP 5505 > > > > To compress 20ms of 16 bit audio sampled at 8Khz it takes 4.2ms. > > Is there any optimizations I can do to make it run faster. The first thing to do would be to profile the code (I assume you're using

...5% vq_nbest_sign 11.12% filter_mem16 4.14% inner_prod 4.07% iir_mem16 2.75% qmf_synth 2.32% lsp_to_lpc 2.32% open_loop_nbest_pitch 1.41% compute_impulse_response 1.37% qmf_decomp 1.28% lpc_to_lsp 1.26% fir_mem16 1.16% speex_bits_pack 1.07% speex_bits_unpack_unsigned 0.86% compute_rms16 0.79% 4. I'm using the echo-canceller + preprocessor, I'd really like to improve performance here: - I would like to use ADI's FFT, but it...

Hi When we run the profiler these 3 functions take more than 8500 cycles. split_cb_search_shape_sign_N1() 22782 pitch_search_3tap() 19208 lpc_to_lsp() 16007 So can I configure these in a different way so it takes less cycles. I do not need very high sound quality Thanks for your time Warm Regards Saju > From: yates at ieee.org > To: saju_pa at hotmail.com > CC: speex-dev at xiph.org > Subject: Re: [Speex-dev] Speex on TI DSP...

Saju Palayur <saju_pa at hotmail.com> writes: > Hi > > > > When we run the profiler these 3 functions take more than 8500 cycles. > > > > split_cb_search_shape_sign_N1() 22782 > pitch_search_3tap() 19208 > lpc_to_lsp() 16007 > > So can I configure these in a different way so it takes less cycles. I am not familiar with the code, so I can't say. I was trying to teach you how to optimize, not give you an optimization. --Randy > I do not need very high sound quality > > Thanks for your ti...

...al:Excl. compute_weighted_codebook 200 4511420 4511420 iir_mem2 599 3338308 3338308 filter_mem2 799 2323655 2323655 compute_impulse_response 200 1800518 1800518 pitch_gain_search_3tap 199 4726604 1744952 open_loop_nbest_pitch 199 4204121 1641016 vq_nbest 800 1626252 1626252 lpc_to_lsp 50 1612650 1558133 nb_encode 50 27412845 1179551 fir_mem2 50 1097300 1097300 inner_prod 27469 1072299 1072299 split_cb_search_shape_sign_N1200 7310588 1007711 normalize16 597 303378 303378, A lower cost option would be to use a floating point 6713. I thought that a...

Hello! I'v downloaded speex lib 1.1.11.1. I am trying to port speex lib to Blackfin processor. I am using VisualDSP++ 4.0. If I am compiling source codes with using floating point everything ok. When I am compiling with FIXED_POINT defined everything's ok and code works about two times faster. But when I am defining BFIN_ASM I am getting several compiling errors in Blackfin assembler

...R lpc_window -R shift_filt -R vbr_hb_thresh -R vbr_nb_thresh -R vbr_uhb_thresh -T bw_lpc -T compute_impulse_response -T compute_rms -T compute_rms16 -T filter_mem16 T _fini -T fir_mem16 -T forced_pitch_quant -T forced_pitch_unquant -T highpass -T iir_mem16 T _init -T inner_prod -T interp_pitch -T lpc_to_lsp -T lsp_enforce_margin -T lsp_interpolate -T lsp_quant_high -T lsp_quant_lbr -T lsp_quant_nb -T lsp_to_lpc -T lsp_unquant_high -T lsp_unquant_lbr -T lsp_unquant_nb -T multicomb -T nb_decode -T nb_decoder_ctl -T nb_decoder_destroy -T nb_decoder_init -T nb_encode -T nb_encoder_ctl -T nb_encoder_destro...

...R lpc_window -R shift_filt -R vbr_hb_thresh -R vbr_nb_thresh -R vbr_uhb_thresh -T bw_lpc -T compute_impulse_response -T compute_rms -T compute_rms16 -T filter_mem16 T _fini -T fir_mem16 -T forced_pitch_quant -T forced_pitch_unquant -T highpass -T iir_mem16 T _init -T inner_prod -T interp_pitch -T lpc_to_lsp -T lsp_enforce_margin -T lsp_interpolate -T lsp_quant_high -T lsp_quant_lbr -T lsp_quant_nb -T lsp_to_lpc -T lsp_unquant_high -T lsp_unquant_lbr -T lsp_unquant_nb -T multicomb -T nb_decode -T nb_decoder_ctl -T nb_decoder_destroy -T nb_decoder_init -T nb_encode -T nb_encoder_ctl -T nb_encoder_destro...

...for (i=0;i<st->lpcSize+1;i++) st->autocorr[i] = MULT16_16_Q14(st->autocorr[i],st->lagWindow[i]); /* Levinson-Durbin */ _spx_lpc(st->lpc+1, st->autocorr, st->lpcSize); st->lpc[0]=(spx_coef_t)LPC_SCALING; /* LPC to LSPs (x-domain) transform */ roots=lpc_to_lsp (st->lpc, st->lpcSize, st->lsp, 15, LSP_DELTA1, stack); /* Check if we found all the roots */ if (roots!=st->lpcSize) { /* Search again if we can afford it */ if (st->complexity>1) roots = lpc_to_lsp (st->lpc, st->lpcSize, st->lsp, 11, LSP_DE...