hash stringlengths 32 32 | doc_id stringlengths 7 13 | section stringlengths 3 121 | content stringlengths 0 2.2M |
|---|---|---|---|
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.13.5 read_alloc_model_delta | The function read_alloc_model_delta() reads the adjustment to a reference bit allocation. The reference allocation is computed from previous and/or current stream data. The encoding of adjustments depends on the allocation model. Table 9-57: read_alloc_model_delta Syntax Reference // read adjustments to reference model... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.13.6 num_bits_step | The function num_bits_step() computes the number of bits per quantization step as a function of the band index. ETSI ETSI TS 103 491 V1.2.1 (2019-05) 171 Table 9-58: num_bits_step Syntax uint num_bits_step(uint band) { uint num_bits = 0; const uint ADJUSTMENT_STEP_SMALL_VALUE = 6; const uint ADJUSTMENT_STEP_LARGE_THRES... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.13.7 read_num_ms_mono_bands | The function read_num_ms_mono_bands() reads the number of midside-mono bands. A band is called midside-mono when only the mid or side band is encoded, and the other band is set to zero. The highest bands in a stereo stream could be encoded as midside-mono. Table 9-59: read_num_ms_mono_bands Syntax Reference // retrieve... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.13.8 read_stereo_flags | The function read_stereo_flags() reads stereo information from the bitstream, specifying the stereo coding mode for each of the coded bands in the stream. Table 9-60: read_stereo_flags Syntax Reference // read stereo flags from the bitstream void read_stereo_flags() { const uint LOW_REGION_THRESHOLD = 4; // Number of b... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.13.9 do_allocation_model | The function do_allocation_model() iteratively searches over the allocation model space until the target allocation is met. The actual model calculations are done by the sub-function do_allocation_model_step. The function returns the number of total number of allocated bits. Table 9-61: do_allocation_model Syntax Refer... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.14 do_first_lognorm_refinement | The function do_first_lognorm_refinement() shown in Table 9-72 applies a first lognorm refinement, using refinement values read from the bitstream, and controlled by _refinement_alloc. Table 9-72: do_first_lognorm_refinement Syntax Reference // apply a first lognorm refinement // // parameter[out] refinement_bits: refi... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.15 Functions Supporting do_first_lognorm_refinement | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.15.1 do_lognorm_refinement | The function do_lognorm_refinement() applies a lognorm refinement, and increases the precision of band energies. This function is called twice: the first time after completion of bit allocation computations (first refinement, clause 9.9.14), the second time after completion of band decoding (late refinement, clause 9.9... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.15.2 compute_refinement_correction | The function compute_refinement_correction() computes the refinement correction given a number of bits and a refinement level. Table 9-74: compute_refinement_correction Syntax // compute a refinement correction // // parameter[in] nbit: number of bits // parameter[in] bits: bit values // parameter[in] level: level of c... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.15.3 transfer_lognorms | The function transfer_lognorms() copies the matrix _lognorm_with_offset to the matrix _lognorm, removing the offsets introduced at encoding time. Table 9-75: transfer_lognorms Syntax // remove lognorm offsets. // transfer_lognorms() { for (uint channel=_first_effective_channel; channel <= _last_effective_channel; chann... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.16 do_late_lognorm_refinement | The function do_late_lognorm_refinement(), shown in Table 9-76 applies a late lognorm refinement, using refinement values read from the bitstream, and controlled by _late_refinement_alloc. This late refinement is applied after processing of all bands. Table 9-76: do_late_lognorm_refinement Syntax Reference // apply a l... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.17 Function Supporting do_late_lognorm_refinement | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.17.1 compute_late_lognorm_refinement_alloc | The function compute_late_lognorm_refinement_alloc(), shown in Table 9-77, computes the bit allocations for a final lognorm refinement. This function uses data from the first refinement allocations, and from bits possibly saved for late refinement during band reconstruction. Table 9-77: compute_late_lognorm_refinement_... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.9.18 apply_lognorms | The function apply_lognorms() in Table 9-78 restores band spectral power values for all channels and all bands. Table 9-78: apply_lognorms Syntax // restore spectral band energies // apply_lognorms() { for ( uint channel=0; channel < _num__stream_channels; channel++ ) { for ( uint band=0; band < _num_effective_bands; b... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10 Decoding Algorithms | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.1 Band Reconstruction | The functions in this clause reconstruct the frequency domain representation of ACE streams. The top level function read_all bands() sequentially reads all the coded band for a given stream from the bitstream, adjusting the number of bits available for reading each band depending on the number of bits that have been sp... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.2 Functions to Perform Band Reconstruction | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.2.1 Band Reconstruction Constants and Definitions | Table 9-79 contains a list of constants and definitions is shared by all band reconstruction modules. Table 9-79: Band Reconstruction Constants and Definitions Syntax Reference const uint UNALLOC_UNIFORM_THRESHOLD = 4; // Threshold for fractional // forwarding const RBFP FRAC_BALANCE_FOR_NEXT_BAND = RBFP::Fraction(7,5)... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.2.2 read_all_bands | The function read_all_bands(), shown in Table 9-80, reads the frequency domain data for a mono (one channel) or stereo (two channels) stream. Ultimately this function resolves to multiple calls to read_one_band_mono(). Table 9-80: read_all_bands Syntax Reference // reconstruct band data from the bitstream. uint read_al... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.2.3 Functions Supporting read_all_bands | 9.10.2.3.1 distribute_running_balance The function distribute_running_balance() redistributes bits that have not been spent on the current band to future bands. Table 9-81: distribute_running_balance Syntax // Redistribute unspent bits to future bands. // // parameter[in] running_balance: current surplus / deficit // p... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.3 Stereo | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.3.1 Stereo Overview | The functions in this clause handle the stereo aspects of ACE stream. The function read_stereo_coding_type() establishes the stereo coding type for a given band, whereas the functions read_one_band_leftright() and read_one_band_midside() read and reconstruct a leftright stereo representation from a leftright or midside... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.3.2 Stereo Constants and Definitions | Table 9-84 lists the constants and definitions is shared by all stereo modules. Table 9-84: Stereo Constants and Definitions Syntax const uint kBandStereoCodingMono = 0; // Mono, single channel const uint kBandStereoCodingMidSideMono = 1; // Stereo, single channel encoded const uint kBandStereoCodingLeftRight = 2; // S... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.3.3 read_stereo_coding_type | read_stereo_coding_type(), shown in Table 9-85, establishes the stereo coding type for the current band. Table 9-85: read_stereo_coding_type Syntax // Establishes the stereo coding type for the current band. // // parameter[in] band: current band index. // return: stereo coding type of the current band uint read_stereo... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.3.4 read_one_band_leftright | The function read_one_band_leftright(), shown in Table 9-86, reads a frequency band of a leftright encoded stereo stream. The actual bitstream reading is handled by the function read_one_band_mono(). Table 9-86: read_one_band_leftright Syntax Reference // Read one band of a leftright encoded stereo stream. // // parame... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.3.5 Functions Called from read_one_band_left_right | 9.10.3.5.1 compute_leftright_allocations The function compute_leftright_allocations() distributes the bits available for a leftright encoded stereo stream over the left and right channel. When the left or the right channel is signaled to be silent, all bits will be allocated to the remaining channel. Table 9-87: comput... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.3.6 read_one_band_midside | The function read_one_band_midside() reads a frequency band of a midside encoded stereo stream. The actual bitstream reading is handled by the function read_one_band_mono(). Table 9-89: read_one_band_midside Syntax Reference // Read one band of a midside encoded stereo stream. // // parameter[in] band: index of frequen... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.3.7 midside_to_leftright | The function midside_to_leftright() transforms the midside representation of a stereo signal to a leftright representation of a stereo channel. Table 9-90: midside_to_leftright Syntax Reference // Transform midside representation to leftright representation. // // parameter[in] band: the index of the frequency band // ... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.9.7 normalize | (band_size,1.0,vector[1]+band_start); } 9.10.4.9.7 |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4 Vector Quantization (VQ) | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.1 Overview of VQ | The functions in this clause and sub-clauses specify how frequency domain coefficients are read from the bitstream. At the lowest level the function read_nonsplit_vector() reads a segment of data from the bitstream given an index for an unsigned pyramid code, and a sequence of sign bits. At the top-level the function r... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.2 VQ Constants and Definitions | The list of constants and definitions shown in Table 9-91 is shared by all vector quantization modules. Table 9-91: VQ Constants and Definitions Syntax const uint MIN_BIT_DIFFERENCE_FOR_REBALANCING = 3; // Threshold for rebalancing const uint LOG_MAXIMUM_QUANTIZATION_LEVEL = 8; // Log of maximum quantization // level c... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.3 read_split_vector | The function read_split_vector(), shown in Table 9-92, recursively reads a band of MDCT coefficients from the bitstream. The input parameter gain specifies the gain to be applied after reconstruction. This parameter is set to 1 at the top level. The depth parameter depth is to be set to 0 at the top level. The value of... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.4 Functions Supporting read_split_vector | 9.10.4.4.1 is_split_required The function is_split_required() indicates whether or not the segment being read needs to be read in parts. Table 9-93: is_split_required Syntax Reference // Determine whether or not a recursive split is required. // // parameter[in] allocated_bits: the bits available for reading band coeff... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.5 read_nonsplit_vector | The function read_nonsplit_vector() reads a segment from of vector of a band of MDCT coefficients from the bitstream and sets the retrieved vector a specified gain value gain. The number of time slots num_blocks (corresponding to num_band_blocks) is provided to allow appropriate reconditioning of vectors encoded with v... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.6 make_unsigned_pyramid_vector | The function make_unsigned_pyramid_vector() reconstructs an element from an unsigned pyramid code with parameters N and K given an index value index. Table 9-97: make_unsigned_pyramid_vector Syntax Reference // Reconstructs an element from an unsigned pyramid code // // parameter[in] index: the index value of the pulse... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.7 Reconditioning | 9.10.4.7.1 recondition The function recondition() reconditions a raw pulse vector to mitigate for non-sparse original vectors that are being reconstructed with only a few pulses. Reconditioning is applied separately for each of the time slots (specified by num_blocks) in the vector. The input parameter reconditioning_l... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.8 Spectral Hole Fill | 9.10.4.8.1 Overview of Spectral Hole Fill Spectral Hole Fill injects non-zero energy when no bits are available for vector quantization. The encoder indicates to the decoder using the _shf_enabled flag whether or not spectrall hole filling is appropriate. The method of Spectral Hole Filling is not normative. For the re... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.4.9 Auxiliary functions | 9.10.4.9.1 factorial Table 9-104: factorial Syntax Reference // parameter[in] n: input value // return: factorial(n) = n*(n-1)*...*1. uint factorial(uint n) { if (n == 0) { return 1; } else { return n*factorial(n-1); } } 9.10.4.9.2 choose Table 9-105: choose Syntax Reference // parameter[in] n: input value // parameter... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.5 Frequency Transforms | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.5.1 Overview of Frequency Transforms | The spectral coefficients recovered from the bitstream are converted to time-domain samples using an overlapped inverse MDCT transform, implemented as a block DCT4 transform followed by phase-shifted window transform. The DCT4 transform operates directly on the band data as a single block transform, applied to the whol... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.5.2 do_windowed_imdct | The windowed IMDCT function operates on a previous and current vector. After completion, the previous vector is ready for further processing, whereas the current vector will only be ready when the next frame has become available. This delay results in a one frame delay for the ACE codec output. The function do_windowed... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.5.3.1 make_window | (N_curr,K,curr_window); 9.10.5.3.1 do_multiple_idct4(N_cur, length, curr_vector); 9.10.5.3.2 do_multipe_window(N_curr,length,prev_window,curr_window,prev_vector,curr_vector); 9.10.5.3.4 } |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.5.3 Functions Supporting do_windowed_imdct | 9.10.5.3.1 make_window The function make_window() populates a window array window according to parameters N and K, where the former is the length of the half-window window and the latter is the size of the transition region from 0 to 1. Table 9-112: make_window Syntax // parameter[in] N: size of half-window // paramete... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.5.4 Band Reshaping | The function reshape_band() reshapes the time-frequency representation of a vector vector from an original number of time slots old_num_blocks to a new number of time slots new_num_blocks. The in/out vector holds the values for a given frequency band, with start bin band_start (included) and end bin band_end (excluded)... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.5.5 Haar Transform | 9.10.5.5.1 Overview of Haar Transform The Haar transform used is the sequentially ordered Haar transform used for reconstruction from time-frequency reshaping in the ACE encoder. 9.10.5.5.2 do_multiple_haar The do_multiple_haar() function applies the Haar transform to contiguous blocks within a vector vector. ETSI ETSI... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.6 Long Term Synthesis (LTS) | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.6.1 Overview of LTS | The long term synthesis module restores 'predictable' signal components that have been removed as a preprocessing step at encoding time. The missing signal components are restored using 1 024 samples previously decoded. A current buffer of samples is restored by time-varying IIR filtering. |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.6.2 LTS Constants and Definitions | The list of constants and definitions shared by all LTS modules is shown in Table 9-121. Table 9-121: LTS constants Syntax // Pi const float PI_2 = acos(0); // Length of history buffer const uint LTS_PREDICTION_RANGE = 1024; // Length of transition window const uint WINDOW_TRANSITION = 102; // Thresholds for window dec... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.6.3 do_longterm_synthesis | The function do_longterm_synthesis(), shown in Table 9-122, is the top level function for long term synthesis. ETSI ETSI TS 103 491 V1.2.1 (2019-05) 212 Table 9-122: do_longterm_synthesis Syntax Reference // Apply Long Term Synthesis to buffer of time samples in current frame. // // parameter[in] pcm1: the correspondin... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.6.4 Functions Supporting do_longterm_synthesis | 9.10.6.4.1 lts_make_buffer The function lts_make_buffer() concatenates the current buffer of samples with the corresponding buffer of one or two previous frames. Table 9-123: lts_make_buffer Syntax // Concatenate three input buffers in a single buffer. // // parameter[in] pcm2: sample array corresponding to twice previ... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.7 Temporal Hole Fill (informative) | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.7.1 Overview of Temporal Hole Filling | Temporal Hole Filling (THF) is an optional part of ACE Stream decoding and applies only in case of a short transform (i.e. 8 MDCT blocks within a frame). Temporal Hole Filling is any process that injects energy in all-zero bands of an MDCT block, in combination with rescaling the modified band to the original norm. The... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.7.2 do_temporal_hole_fill (informative) | The function do_temporal_hole_fill(), in Table 9-126, runs over all effective channels and all effective bands and applies temporal hole filling. This function only has effect when temporal hole filling is enabled and only applies when _is_short is true. Table 9-126: do_temporal_hole_fill Syntax Reference // Temporal H... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.7.3 do_band_temporal_hole_fill (informative) | The function do_band_temporal_hole_fill() in Table 9-127 injects energy into an all-zero band of a short MDCT block. Table 9-127: do_band_temporal_hole_fill Syntax Reference // Temporal hole filling (THF) // Signal energy is inserted in all-zero blocks. // This processing only applies to short frames. // // parameter[i... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.7.4 random_sign | The function random_sign(), shown in Table 9-128, draws with equal probability between the values -1 and +1. Table 9-128: random_sign Syntax Reference // random equal probability drawing of -1 and +1 int random_sign() { bool result = -1; float value = Rand(0.0,2.0); if (value >= 1.0) { result = +1; } return result; } 8... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.8 Deemphasis | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.8.1 Overview of Deemphasis | A final step in the decoding process boosts lower frequencies and attenuates higher frequencies, compensating for the inverse pre-processing step in the encoding process. The deemphasis filter is implemented as a one-tap IIR filter. The filter state is maintained in the _deemphasis_prev variable. |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.8.2 do_deemphasis | The function de_demphasis() shown in Table 9-129, applies a demphasis filter to all output streams of an ACE stream decoder. Table 9-129: do_deemphasis Syntax Reference // Deemphasis filtering over all channels // updating state for the next frame. void do_demphasis() { for (int channel = 0 ; channel < _num_stream_chan... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.10.8.3 Function Called by do_deemphasis | 9.10.8.3.1 do_channel_deemphasis The function do_channel_deemphasis(float prev, float buffer[]) operates on a single channel, boosting lower frequencies and attenuating higher frequencies. Table 9-130: do_channel_deemphasis Syntax // Deemphasis filtering for one channel // // Low frequency boost / high frequency attenu... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11 LFE Decoding | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11.1 Overview of LFE Decoding | An LFE channel is reconstructed in two steps: 1) reconstruction of a decimated representation; and 2) upsampling to full sampling frequency. Reconstruction optionally involves insertion of synthetic noise. An informative example for random noise generation is included for completeness. All functions in this clause have... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11.2 LFE Constants and Definition | The constants defined in Table 9-131 are used throughout the remainder of the LFE pseudo-code examples. Table 9-131: LFE Constants LFE Constants and DefinitionSyntax const uint kLFEAbsoluteMode = 0; // LFE samples at full resolution const uint kLFEPredictMode = 1; // LFE samples as residuals const uint kLFEReducedMode ... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11.3 decode_lfe_channel | decode_lfe_channel(), shown in Table 9-132 reconstructs an LFE channel in two steps. In the first step a subsampled version of the LFE channel is reconstructed from the data read from the bitstream, either predictively (kLFEPredictMode) or absolutely (kLFEAbsoluteMode, kLFEReducedMode). If the lfe_mode is given as kLFE... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11.4 Functions Supporting decode_lfe_channel | |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11.4.1 lfe_predict | The function lfe_predict() reconstructs the LFE decimated channel from the residual sample values and prediction coefficients that have been read from the bitstream. Table 9-133: lfe_predict Syntax // Reconstruct the LFE decimated channel in predictive mode lfe_predict() { _lfe_decimated_channel[-1] = previous->_lfe_de... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11.4.2 lfe_reconstruct | The function lfe_reconstruct() reconstructs the LFE decimated channel from the quantized sample values that have been read from the bitstream. Table 9-134: lfe_reconstruct Syntax // Reconstruct LFE decimated channel from quantized sample values lfe_reconstruct() { for (int i=0;i<_decimated_frame_duration;i++) { _lfe_de... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11.4.3 lfe_synthesize (informative) | The function lfe_synthesize() inserts synthetic noise of the appropriate power as determined by the value of db_norm. Various methods of pseudo-random noise generation may be used here. Table 9-135: lfe_synthesize Syntax Reference // Inject synthetic noise lfe_synthesize() { const float LFE_NORM_FACTOR = 1.73; for (int... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11.4.4 lfe_upsample | The function lfe_upsample() interpolates the LFE decimated channel to full frame duration. This function is called from the next (future) frame after extending the LFE decimated channel with current decimated sample values. Table 9-136: lfe_upsample Syntax Reference // Interpolation of LFE decimated channel to full fra... |
e16d334f11519b9671e26952047d356b | 103 491 | 9.11.4.5 db_to_linear | The function db_to_linear(int n) converts a integer dB value to a linear (floating point) scalar. Table 9-137: db_to_linear Syntax // Converting integer dB to linear // // parameter[in] n: input value // return 10^(val/10) float db_to_linear(int n) { float crt2[] = {1.0, 1.25992107, 1.58740103}; int prem = (n % 3 + 3) ... |
e16d334f11519b9671e26952047d356b | 103 491 | 10 Tables and Constants | |
e16d334f11519b9671e26952047d356b | 103 491 | 10.1 Overview | The clauses throughout clause 10 contain constants and look-up tables used by the ACE bitstream extraction and decoding algorithms in clause 9. ETSI ETSI TS 103 491 V1.2.1 (2019-05) 220 |
e16d334f11519b9671e26952047d356b | 103 491 | 10.2 System Constants and Tables | |
e16d334f11519b9671e26952047d356b | 103 491 | 10.2.1 Basic Constants | Table 10-1: System Constants Syntax const uint NUM_ACE_BANDS = 22; // Number of spectral bands |
e16d334f11519b9671e26952047d356b | 103 491 | 10.2.2 Coding Modes | Table 10-2: Coding Modes Syntax const uint kCodingFull = 0; const uint kCodingOff = 1 const uint kCodingLeft = 2; const uint kCodingRight = 3; |
e16d334f11519b9671e26952047d356b | 103 491 | 10.2.3 BandsConfig | |
e16d334f11519b9671e26952047d356b | 103 491 | 10.2.3.1 Structure BandsConfig | The BandsConfig datatype defines the basic aspects of spectral representation. Table 10-3: BandsConfig Syntax struct BandsConfig { const uint num_bands; // number of bands const uint index_largest_band; // index of the largest band const uint num_narrow_bands; // number of minimal size bands const band_boundaries[NUM_A... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.2.3.2 Bands Configuration Tables | 10.2.3.2.1 bands_config_48khz Table 10-4 lists the configuration tables for 48 KHz. Table 10-4: bands_config_48khz Syntax const BandsConfig BANDS_CONFIG_48KHZ = { // num_bands 22, // index_largest_band 21, // num_narrow_bands 8, // band_boundaries { 0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 144, 176, 208, 240, 288... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.3 Long Term Synthesis Constants and Tables | |
e16d334f11519b9671e26952047d356b | 103 491 | 10.3.1 LTS_FILTER_CODEBOOK | Table 10-6: LTS Filter Codebook Syntax // LTS_FILTER_CODEBOOK float LTS_FILTER_CODEBOOK[29[3] = { {0, 0, 0}, {0.064995, 0.040695, 0.020557}, {0.079613, 0.047869, 0.006069}, {0.115744, 0.029569, 0.004913}, {0.15375, 0.01125, 0.005625}, {0.097493, 0.061043, 0.030836}, {0.119419, 0.071803, 0.009104}, {0.173616, 0.044353, ... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.4 Lognorm Tables and Constants | |
e16d334f11519b9671e26952047d356b | 103 491 | 10.4.1 Lognorm Basic Constants | Table 10-7: Lognorm Constants Syntax const LNFP LOGNORM_MINUS_INFINITY = -32; |
e16d334f11519b9671e26952047d356b | 103 491 | 10.4.2 LOGNORM_GOLOMB_PARAMS | Table 10-8: Lognorm Golomb Parameters Syntax static const GolombCodeParams golomb_params_64_0_15 = { 64, 0, 15}; static const GolombCodeParams golomb_params_64_1_31 = { 64, 1, 31}; static const GolombCodeParams golomb_params_64_2_31 = { 64, 2, 31}; static const GolombCodeParams golomb_params_64_3_31 = { 64, 3, 31}; uin... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.4.3 F_PREDICTOR_TABLE | Table 10-9: F Predictor Table Syntax uint const NUM_PRED_SETS = 2; const LNFP F_PREDICTOR_TABLE[NUM_PRED_SETS][NUM_ACE_BANDS] = { // predictive frame { 0.5996093750, 0.5800781250, 0.5654296875, 0.5498046875, 0.5371093750, 0.5253906250, 0.5146484375, 0.5048828125, 0.5000000000, 0.5000000000, 0.5000000000, 0.5000000000, ... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.4.4 B_PREDICTOR_TABLE | Table 10-10: B Predictor Table Syntax uint const NUM_PRED_SETS = 2; const LNFP B_PREDICTOR_TABLE[NUM_PRED_SETS][NUM_ACE_BANDS] = { // predictive frame { 0.0000000000, 1.0000000000, 0.7500000000, 0.5000000000, 0.5000000000, 0.4501953125, 0.4003906250, 0.4003906250, 0.3496093750, 0.5000000000, 0.5000000000, 0.5000000000,... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.4.5 LOGNORM_REFINEMENT_BOOST_TABLE | Table 10-11: Lognorm Refinement Boost Table Syntax LNFP LOGNORM_REFINEMENT_BOOST_TABLE[NUM_ACE_BANDS] = { 0.375000, 0.375000, 0.375000, 0.375000, 0.375000, 0.375000, 0.375000, 0.375000, 0.500000, 0.500000, 0.500000, 0.625000, 0.625000, 0.625000, 0.625000, 0.679688, 0.734375, 0.789063, 0.843750, 0.898438, 0.937500, 0.93... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.5 Stereo Constants and Tables | |
e16d334f11519b9671e26952047d356b | 103 491 | 10.5.1 Stereo Coding Modes | Table 10-12: Stereo Coding Modes Syntax // Stereo coding modes const uint kStereoMono = 0; const uint kStereoMidSideMono = 1; const uint kStereoLeftRight = 2; const uint kStereoMidSide = 3; ETSI ETSI TS 103 491 V1.2.1 (2019-05) 226 |
e16d334f11519b9671e26952047d356b | 103 491 | 10.5.2 Stereo Allocation Modes | Table 10-13: Stereo Allocation Modes Syntax // Stereo allocation modes const uint kLrStereoAllocation5050 = 0; // 50/50 bit distribution const uint kLrStereoAllocation6040 = 1; // 60/40 bit distribution const uint kLrStereoAllocation4060 = 2; // 40/60 bit distribution const uint kLrStereoAllocation7525 = 3; // 75/25 bi... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.6 Bit Allocation Constants and Tables | |
e16d334f11519b9671e26952047d356b | 103 491 | 10.6.1 Bit Allocation Models | Table 10-14: Bit Allocation Models Syntax // bit allocations are incrementally adjusted from values derived // from the primary lognorms, with a standard interpretation // of adjustment values. const uint kUnsignedDeltaFromPrimaryLSQ = 0; // bit allocations are derived as (non-negative) // modifications to a bit alloca... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.7 Vector Quantizer Constants and Tables | |
e16d334f11519b9671e26952047d356b | 103 491 | 10.7.1 Vector Quantizer Basic Constants | Table 10-15: Stereo Allocation Modes Syntax // VQ constants cont uint MAX_BAND_SIZE = 208; ETSI ETSI TS 103 491 V1.2.1 (2019-05) 227 |
e16d334f11519b9671e26952047d356b | 103 491 | 10.7.2 MAX_BITS_FOR_N | Table 10-16: Max Bits for N Syntax // entries for N=0 and N=1 are omitted; first entry is for N=2. const uint MAX_BITS_FOR_N[MAX_BAND_SIZE - 1] = { 11, 19, 26, 33, 38, 39, 40, 40, 41, 41, 41, 41, 42, 42, 42, 42, 41, 41, 42, 41, 41, 41, 41, 41, 40, 41, 41, 40, 41, 41, 41, 39, 39, 40, 40, 41, 39, 39, 39, 39, 40, 40, 40, ... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.7.3 MAX_K_FOR_N | Table 10-17: Max K for N Syntax // Maximum value of K such that pyramid(N,K) fits in 32 bit. const uint MAX_K_FOR_N[MAX_BAND_SIZE] = { 256, 256, 256, 256, 256, 217, 117, 76, 55, 43, 36, 31, 27, 24, 22, 20, 19, 17, 16, 16, 15, 14, 14, 13, 13, 12, 12, 12, 11, 11, 11, 11, 10, 10, 10, 10, 10, 9, 9, 9, 9, 9, 9, 9, 8, 8, 8, ... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.7.4 K_ESTIMATION_COEFFICIENTS | Table 10-18: K Estimation Coefficients Syntax // K-estimation coefficients. const uint K_ESTIMATION_COEFFICIENTS[] = { {411, 128}, {733, 255}, {671, 367}, {726, 471}, {597, 571}, {243, 656}, {118, 751}, {520, 887}, {679, 1007}, {831, 1121}, {1044, 1251}, {1166, 1370}, {1329, 1493}, {1458, 1623}, {1637, 1762}, {1761, 18... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.7.5 BETA_DEQUANT_TABLE | Table 10-19: Beta Dequantization Table Syntax const uint MAX_QUANT_LEVEL = 256; int BETA_DEQUANT_TABLE[MAX_QUANT_LEVEL/2][MAX_QUANT_LEVEL+1] = // The entry corresponding to a given pair (quantization_level, index) is found // in BETA_DEQUANT_TABLE[quantization_level/2-1][index] if // index <= quantization_level/2; othe... |
e16d334f11519b9671e26952047d356b | 103 491 | 10.8 LFE Constants and Tables | |
e16d334f11519b9671e26952047d356b | 103 491 | 10.8.1 INTERPOLATION_FILTER_TABLE | Table 10-20: Interpolation_filter_table Syntax // INTERPOLATION_FILTER_TABLE const float INTERPOLATION_FILTER_TABLE[512] = { -7.6688190671960729e-06, -7.2122179843474751e-06, 9.1138652731199750e-07, 9.1714805409316167e-07, 9.3845793823653087e-07, 9.7469236870201287e-07, 1.0253352563411658e-06, 1.0899559063179447e-06, 1... |
3a3e023f0bbc7bf133c249f8b740dda6 | 103 464 | 1 Scope | The present document augments clause 6 of ETSI TS 102 796 [1], which states that broadcast-related applications are signalled as part of the broadcast. It defines a method for discovery of HbbTV® applications in settings where AIT signalling via the broadcast network is not available to the terminal. In this situation,... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.