User contributions for Superdump
Jump to navigation
Jump to search
16 September 2007
- 11:0911:09, 16 September 2007 diff hist +8 AMR-NB-WIP →The LSP vector is converted to LP filter coefficients (c.f. §5.2.4)
15 September 2007
- 07:4807:48, 15 September 2007 diff hist +8 AMR-NB-WIP →Calculation of the quantified fixed gain
- 07:4807:48, 15 September 2007 diff hist +39 N File:Amrnb fixedgain.gif AMR-NB: calculate quantified fixed gain current
- 05:5405:54, 15 September 2007 diff hist −9 AMR-NB-WIP →Pitch gain
- 05:5205:52, 15 September 2007 diff hist −10 AMR-NB-WIP →Fixed gain prediction
- 05:5105:51, 15 September 2007 diff hist +3 AMR-NB-WIP →Pitch sharpening
- 05:5005:50, 15 September 2007 diff hist +37 N File:Amrnb fixvecpitshp.gif AMR-NB: fixed vector pitch sharpening current
- 05:0005:00, 15 September 2007 diff hist +3 AMR-NB-WIP →The available LSP vector(s) are used to linearly interpolate vectors for the other subframes (c.f. §5.2.6)
- 05:0005:00, 15 September 2007 diff hist +6 AMR-NB-WIP →The available LSP vector(s) are used to linearly interpolate vectors for the other subframes (c.f. §5.2.6)
- 05:0005:00, 15 September 2007 diff hist +59 N File:Amrnb lspinterpq3-2.gif AMR-NB: linear interpolation of LSP vector ^q at subframe 3 current
- 04:5904:59, 15 September 2007 diff hist +59 N File:Amrnb lspinterpq2.gif AMR-NB: linear interpolation of LSP vector ^q at subframe 2 current
- 04:5804:58, 15 September 2007 diff hist +59 N File:Amrnb lspinterpq1-2.gif AMR-NB: linear interpolation of LSP vector ^q at subframe 1 current
- 04:5604:56, 15 September 2007 diff hist +59 N File:Amrnb lspinterpq3-1.gif AMR-NB: linear interpolation of LSP vector ^q at subframe 3 current
- 04:5404:54, 15 September 2007 diff hist +59 N File:Amrnb lspinterpq1-1.gif AMR-NB: linear interpolation of LSP vector ^q at subframe 1 current
- 04:5004:50, 15 September 2007 diff hist +137 AMR-NB-WIP →Prediction from the previous frame is added to obtain the mean-removed LSF vector
- 04:5004:50, 15 September 2007 diff hist +394 N File:Amrnb lsfmeanrem2.gif AMR-NB: mean-removed LSF vector ; z_j(n)[i] : the ith coefficient of the mean-removed LSF vector at the jth subframe ; r_j(n)[i] : the ith coefficient of the prediction residual vector of frame n at the jth subframe ; pred_fac[i] : the ith coefficient of current
- 04:4004:40, 15 September 2007 diff hist +24 AMR-NB-WIP →Prediction from the previous frame is added to obtain the mean-removed LSF vectors
- 04:3804:38, 15 September 2007 diff hist −1 AMR-NB-WIP →Prediction from the previous frame is added to obtain the mean-removed LSF vectors
14 September 2007
- 18:3818:38, 14 September 2007 diff hist +26 AMR-NB-WIP →The LSF vectors are converted to cosine domain LSP vectors
- 18:3818:38, 14 September 2007 diff hist +302 N File:Amrnb lspvec.gif AMR-NB: LSP vectors ; q_k[i] : the ith coefficient of the kth line spectral pair (LSP) in the cosine domain ; k : the two lsf vectors give the LSP vectors q2, q4 at the 2nd and 4th subframes; k = 2*j ; f_j[i] : ith coefficient of the jth LSF vector; [0,4 current
- 18:3218:32, 14 September 2007 diff hist +7 AMR-NB-WIP →The mean is added
- 18:3218:32, 14 September 2007 diff hist +173 N File:Amrnb lsfvec.gif AMR-NB: add mean to lsf vectors ; lsf_mean_m : a table of the means of the LSF coefficients ; m : the number of indices parsed according to the mode ; f_j : the LSF vectors current
- 15:3615:36, 14 September 2007 diff hist +45 AMR-NB-WIP →Prediction from the previous frame is added to obtain the mean-removed LSF vectors
- 15:3415:34, 14 September 2007 diff hist +229 N File:Amrnb lsfmeanrem.gif AMR-NB: find the mean-removed LSF vector z_j(n) : a mean-removed LSF vector from the current frame r_j(n) : prediction residual vector of frame n at subframe j ^r_2(n-1) : the quantified 2nd residual vector in the previous frame current
- 15:0915:09, 14 September 2007 diff hist +1 AMR-NB-WIP →Anti-sparseness processing
- 15:0715:07, 14 September 2007 diff hist +1 AMR-NB-WIP →Smoothing of the fixed codebook gain
- 15:0615:06, 14 September 2007 diff hist −118 AMR-NB-WIP No edit summary
- 14:5714:57, 14 September 2007 diff hist +218 AMR-NB-WIP →Other modes - vector quantised gains
- 11:3311:33, 14 September 2007 diff hist +172 AMR-NB-WIP →Smoothing of the fixed codebook gain
- 11:2911:29, 14 September 2007 diff hist +6 AMR-NB-WIP →Smoothing of the fixed codebook gain
- 11:2911:29, 14 September 2007 diff hist +549 AMR-NB-WIP →Smoothing of the fixed codebook gain
- 11:2811:28, 14 September 2007 diff hist +133 N File:Amrnb smoothfac.gif AMR-NB: fixed gain smoothing factor k_m - fixed gain smoothing factor K_1 - 0.4 K_2 - 0.25 diff_m - difference measure at subframe m current
- 11:2511:25, 14 September 2007 diff hist +147 N File:Amrnb meangc.gif AMR-NB: averaged fixed gain used for fixed gain smoothing \bar{g}(m) - mean fixed gain at subframe m ^g_c(k) - quantified fixed gain at subframe k current
- 11:2211:22, 14 September 2007 diff hist +171 N File:Amrnb lspsmooth.gif AMR-NB: averaged LSP vector used for fixed gain smoothing \bar{q}(n) - averaged LSP vector at frame n \hat{q}_4(n) - quantified LSP vector for the 4th subframe at frame n current
- 11:1811:18, 14 September 2007 diff hist +184 N File:Amrnb lspdiff.gif AMR-NB: difference measure for adaptive fixed gain smoothing j - loops over LSPs m - loops over subframes \bar{q} - averaged LSP vector \hat{q}_m - quantified LSP vector at subframe m current
- 11:0911:09, 14 September 2007 diff hist +113 N File:Amrnb gcsmooth.gif AMR-NB: fixed gain smoothing ^g_c - quantified fixed gain k_m - smoothing factor \bar{g_c} - averaged fixed gain current
12 September 2007
- 20:3020:30, 12 September 2007 diff hist +1,066 AMR-NB-WIP →Other modes - vector quantised gains
- 17:2717:27, 12 September 2007 diff hist +19 AMR-NB-WIP →Fixed gain prediction
- 12:1812:18, 12 September 2007 diff hist +460 AMR-NB-WIP →Dequantisation of the gains
- 08:5808:58, 12 September 2007 diff hist +114 AMR-NB-WIP →Decoding of the pitch and fixed codebook gains
- 07:1307:13, 12 September 2007 diff hist −679 AMR-NB-WIP →Fixed gain prediction
- 07:1207:12, 12 September 2007 diff hist +1,536 AMR-NB-WIP →12.2kbps and 7.95kbps - scalar quantised gains
- 07:1007:10, 12 September 2007 diff hist +148 N File:Amrnb meaniner.gif AMR-NB: calculated mean innovation (fixed) energy E_I - calculated mean innovation (fixed) energy N - subframe size 40 c(n) - fixed codebook vector current
- 06:0706:07, 12 September 2007 diff hist +169 N File:Amrnb predener.gif AMR-NB: predicted energy \tilde{E} - predicted energy b = [0.68, 0.58, 0.34, 0.19] - 4-tap MA prediction coefficients ^R(k) - quantified prediction errors at subframe k current
- 05:5105:51, 12 September 2007 diff hist +228 N File:Amrnb fixedgainpred.gif AMR-NB: fixed gain prediction g_c' - predicted fixed gain \tilde{E} - predicted energy \bar{E} - desired mean innovation (fixed) energy (pre-specified, differs with mode) E_I - actual mean innovation (fixed) energy (calculated) current
11 September 2007
- 17:0817:08, 11 September 2007 diff hist +204 AMR-NB-WIP →Decoding of the adaptive and fixed codebook gains
- 15:3215:32, 11 September 2007 diff hist +6 AMR-NB-WIP →Adaptive codebook vector is found by interpolating the past excitation at the pitch lag using an FIR filter
- 15:3015:30, 11 September 2007 diff hist +226 N File:Amrnb pitchvector.gif Finding the adaptive code book vector v(n) by interpolating the past excitation, u(n), at the pitch lag, k, using the b60 FIR filter based on a Hamming windowed sinc function truncated at ±59 and zero-padded at ±60. (AMR-NB) current
- 15:2315:23, 11 September 2007 diff hist +3 AMR-NB-WIP →Computing the reconstructed speech
- 15:1815:18, 11 September 2007 diff hist +210 N File:Amrnb excitation.gif AMR-NB how to construct the excitation vector u(n) - excitation vector ^g_p - quantified pitch gain ^g_c - quantified fixed code book gain v(n) - adaptive (pitch) code book vector c(n) - fixed code book vector current
