annotate core/magnatagatune/MTTMixedFeatureGenreRandom.m @ 0:e9a9cd732c1e tip

first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
rev   line source
wolffd@0 1 classdef MTTMixedFeatureGenreRandom < MTTAudioFeature & handle
wolffd@0 2 % ---
wolffd@0 3 %
wolffd@0 4 % The usual worklow for these features constist of three steps
wolffd@0 5 % 1. extract: extracts the basic single-file dependent features
wolffd@0 6 % 2. define_global_transform: calculates the global feature
wolffd@0 7 % transformation parameters
wolffd@0 8 % 3. finalise: applies the common transformations to a specific feature
wolffd@0 9 % ---
wolffd@0 10
wolffd@0 11 properties(Constant = true)
wolffd@0 12
wolffd@0 13 % svn hook
wolffd@0 14 my_revision = str2double(substr('$Rev: 455 $', 5, -1));
wolffd@0 15 end
wolffd@0 16
wolffd@0 17 properties
wolffd@0 18 % ---
wolffd@0 19 % Set default parameters
wolffd@0 20 % ---
wolffd@0 21 my_params = MTTAudioFeature.inherited_params(...
wolffd@0 22 'MTTTagFeatureGenreBasic', ...
wolffd@0 23 ... % ---
wolffd@0 24 ... % following are Random parameters
wolffd@0 25 ... % ---
wolffd@0 26 'nrandoms', 100 ...
wolffd@0 27 );
wolffd@0 28 end
wolffd@0 29
wolffd@0 30 % ---
wolffd@0 31 % member functions
wolffd@0 32 % ---
wolffd@0 33 methods
wolffd@0 34
wolffd@0 35 % ---
wolffd@0 36 % constructor: pointer to feature in database
wolffd@0 37 % ---
wolffd@0 38 function feature = MTTMixedFeatureGenreRandom(varargin)
wolffd@0 39
wolffd@0 40 feature = feature@MTTAudioFeature(varargin{:});
wolffd@0 41 end
wolffd@0 42
wolffd@0 43 % ---
wolffd@0 44 % extract feature data by combining genre adn basicsm features
wolffd@0 45 % ---
wolffd@0 46 function data = extract(feature, clip)
wolffd@0 47
wolffd@0 48 % ---
wolffd@0 49 % get MTTMixedFeatureGenreBasicSm this includes possible
wolffd@0 50 % local normalisations
wolffd@0 51 % ---
wolffd@0 52 data.tags = clip.features('MTTTagFeatureGenreBasic',feature.my_params);
wolffd@0 53
wolffd@0 54 % ---
wolffd@0 55 % get genre tag features
wolffd@0 56 % ---
wolffd@0 57
wolffd@0 58 data.random = clip.features('MTTRandomFeature',feature.my_params);
wolffd@0 59
wolffd@0 60
wolffd@0 61 % prepare field for final features
wolffd@0 62 data.final.vector = [];
wolffd@0 63 data.final.vector_info = struct();
wolffd@0 64 data.final.dim = 0;
wolffd@0 65
wolffd@0 66 % save info data
wolffd@0 67 data.info.type = class(feature);
wolffd@0 68 data.info.owner_id = clip.id;
wolffd@0 69 data.info.creatorrev = feature.my_revision;
wolffd@0 70
wolffd@0 71 % save parameters
wolffd@0 72 data.info.params = feature.my_params;
wolffd@0 73 end
wolffd@0 74
wolffd@0 75 function define_global_transform(features)
wolffd@0 76 % calculate and set normalization factors from the group of
wolffd@0 77 % input features. These features will be set for the full database
wolffd@0 78
wolffd@0 79 if numel(features) == 1
wolffd@0 80 error ('Insert feature array for this method');
wolffd@0 81 end
wolffd@0 82
wolffd@0 83 % ---
wolffd@0 84 % We collect all the relevant tag
wolffd@0 85 % features and get the transform on this basis.
wolffd@0 86 % ---
wolffd@0 87 for i = 1:numel(features)
wolffd@0 88 tags(i) = features(i).data.tags;
wolffd@0 89 end
wolffd@0 90
wolffd@0 91 % call the features own transsform function
wolffd@0 92 tags.define_global_transform();
wolffd@0 93
wolffd@0 94 % ---
wolffd@0 95 % Random features have no global transform
wolffd@0 96 % ---
wolffd@0 97
wolffd@0 98 % ---
wolffd@0 99 % set common feature values for mixed features
wolffd@0 100 % ---
wolffd@0 101 features(1).my_db.set_common([1]); %trivial common
wolffd@0 102 end
wolffd@0 103
wolffd@0 104
wolffd@0 105 function finalise(feature)
wolffd@0 106 % applies a final transformation and collects the
wolffd@0 107 % information of this feature within a single vector
wolffd@0 108 % see info for types in specific dimensions
wolffd@0 109
wolffd@0 110 for i = 1:numel(feature)
wolffd@0 111
wolffd@0 112 % check for neccesary parameters
wolffd@0 113 if isempty(feature(i).my_db.commondb)
wolffd@0 114
wolffd@0 115 error('Define the global transformation first');
wolffd@0 116 end
wolffd@0 117
wolffd@0 118 % ---
wolffd@0 119 % final data assembly
wolffd@0 120 % ---
wolffd@0 121 % concatenate vectors
wolffd@0 122 feature(i).data.final.vector = ...
wolffd@0 123 [feature(i).data.tags.vector() ;...
wolffd@0 124 feature(i).data.random.vector()];
wolffd@0 125
wolffd@0 126 % add up feature dimensions
wolffd@0 127 feature(i).data.final.dim = feature(i).data.tags.dim...
wolffd@0 128 + feature(i).data.random.dim;
wolffd@0 129
wolffd@0 130 % concatenate labels
wolffd@0 131 lbl1 = feature(i).data.tags.labels();
wolffd@0 132 lbl2 = feature(i).data.random.labels();
wolffd@0 133 feature(i).data.final.vector_info.labels = ...
wolffd@0 134 {lbl1{:}, lbl2{:}};
wolffd@0 135 end
wolffd@0 136 end
wolffd@0 137
wolffd@0 138 % ---
wolffd@0 139 % destructor: do we really want to remove this
wolffd@0 140 % from the database? No, but
wolffd@0 141 % TODO: create marker for unused objects in db, and a cleanup
wolffd@0 142 % function
wolffd@0 143 % ---
wolffd@0 144 function delete(feature)
wolffd@0 145
wolffd@0 146 end
wolffd@0 147 end
wolffd@0 148 end