File: MATCHAligner.cpp

package info (click to toggle)
sonic-visualiser 5.2.1-2
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid
  • size: 24,744 kB
  • sloc: cpp: 158,888; ansic: 11,920; sh: 1,785; makefile: 517; xml: 64; perl: 31
file content (428 lines) | stat: -rw-r--r-- 14,523 bytes parent folder | download | duplicates (2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
/* -*- c-basic-offset: 4 indent-tabs-mode: nil -*-  vi:set ts=8 sts=4 sw=4: */

/*
    Sonic Visualiser
    An audio file viewer and annotation editor.
    Centre for Digital Music, Queen Mary, University of London.
    
    This program is free software; you can redistribute it and/or
    modify it under the terms of the GNU General Public License as
    published by the Free Software Foundation; either version 2 of the
    License, or (at your option) any later version.  See the file
    COPYING included with this distribution for more information.
*/

#include "MATCHAligner.h"

#include "data/model/SparseTimeValueModel.h"
#include "data/model/RangeSummarisableTimeValueModel.h"
#include "data/model/AlignmentModel.h"
#include "data/model/AggregateWaveModel.h"

#include "framework/Document.h"

#include "transform/TransformFactory.h"
#include "transform/ModelTransformerFactory.h"
#include "transform/FeatureExtractionModelTransformer.h"

#include <QSettings>

namespace sv {

MATCHAligner::MATCHAligner(Document *doc,
                           ModelId reference,
                           ModelId toAlign,
                           bool subsequence,
                           bool withTuningDifference) :
    m_document(doc),
    m_reference(reference),
    m_toAlign(toAlign),
    m_subsequence(subsequence),
    m_withTuningDifference(withTuningDifference),
    m_tuningFrequency(440.f),
    m_incomplete(true)
{
}

MATCHAligner::~MATCHAligner()
{
    if (m_incomplete) {
        auto other =
            ModelById::getAs<RangeSummarisableTimeValueModel>(m_toAlign);
        if (other) {
            other->setAlignment({});
        }
    }

    ModelById::release(m_tuningDiffOutputModel);
    ModelById::release(m_pathOutputModel);
}

QString
MATCHAligner::getAlignmentTransformName(bool subsequence)
{
    QSettings settings;
    settings.beginGroup("Alignment");
    TransformId id;
    if (subsequence) {
        id = settings.value
            ("transform-id-subsequence",
             "vamp:match-vamp-plugin:match-subsequence:path").toString();
    } else {
        id = settings.value
            ("transform-id",
             "vamp:match-vamp-plugin:match:path").toString();
    }
    settings.endGroup();
    return id;
}

QString
MATCHAligner::getTuningDifferenceTransformName()
{
    QSettings settings;
    settings.beginGroup("Alignment");
    TransformId id = settings.value
        ("tuning-difference-transform-id",
         "vamp:tuning-difference:tuning-difference:tuningfreq")
        .toString();
    settings.endGroup();
    return id;
}

bool
MATCHAligner::isAvailable(bool subsequence,
                          bool withTuningDifference)
{
    TransformFactory *factory = TransformFactory::getInstance();
    TransformId id = getAlignmentTransformName(false);

    TransformId subId, tdId;

    if (subsequence) {
        subId = getAlignmentTransformName(true);
    }
    if (withTuningDifference) {
        tdId = getTuningDifferenceTransformName();
    }

    bool haveId = factory->haveTransform(id);
    bool haveSubId = (subId == "" || factory->haveTransform(subId));
    bool haveTdId = (tdId == "" || factory->haveTransform(tdId));
    bool available = haveId && haveSubId && haveTdId;

    SVDEBUG << "MATCHAligner::isAvailable: "
            << "transform id = " << id << " (have: " << haveId
            << "), subId = " << subId << " (have: " << haveSubId
            << "), tdId = " << tdId << " (have: " << haveTdId << ")"
            << ": available = " << available << endl;

    return available;
}

void
MATCHAligner::begin()
{
    auto reference =
        ModelById::getAs<RangeSummarisableTimeValueModel>(m_reference);
    auto other =
        ModelById::getAs<RangeSummarisableTimeValueModel>(m_toAlign);

    if (!reference || !other) return;

    // This involves creating a number of new models:
    //
    // 1. an AggregateWaveModel to provide the mixdowns of the main
    // model and the new model in its two channels, as input to the
    // MATCH plugin. We just call this one aggregateModel
    //
    // 2a. a SparseTimeValueModel which will be automatically created
    // by FeatureExtractionModelTransformer when running the
    // TuningDifference plugin to receive the relative tuning of the
    // second model (if pitch-aware alignment is enabled in the
    // preferences). This is m_tuningDiffOutputModel.
    //
    // 2b. a SparseTimeValueModel which will be automatically created
    // by FeatureExtractionPluginTransformer when running the MATCH
    // plugin to perform alignment (so containing the alignment path).
    // This is m_pathOutputModel.
    //
    // 3. an AlignmentModel, which stores the path and carries out
    // alignment lookups on it. This one is m_alignmentModel.
    //
    // Models 1 and 3 are registered with the document, which will
    // eventually release them. We don't release them here except in
    // the case where an activity fails before the point where we
    // would otherwise have registered them with the document.
    //
    // Models 2a (m_tuningDiffOutputModel) and 2b (m_pathOutputModel)
    // are not registered with the document, because they are not
    // intended to persist. These have to be released by us when
    // finished with, but their lifespans do not extend beyond the end
    // of the alignment procedure, so this should be ok.

    AggregateWaveModel::ChannelSpecList components;
    components.push_back
        (AggregateWaveModel::ModelChannelSpec(m_reference, -1));

    components.push_back
        (AggregateWaveModel::ModelChannelSpec(m_toAlign, -1));

    auto aggregateModel = std::make_shared<AggregateWaveModel>(components);
    m_aggregateModel = ModelById::add(aggregateModel);
    m_document->addNonDerivedModel(m_aggregateModel);

    auto alignmentModel = std::make_shared<AlignmentModel>
        (m_reference, m_toAlign, ModelId());
    m_alignmentModel = ModelById::add(alignmentModel);

    TransformId tdId;
    if (m_withTuningDifference) {
        tdId = getTuningDifferenceTransformName();
    }

    if (tdId == "") {
        
        if (beginAlignmentPhase()) {
            other->setAlignment(m_alignmentModel);
            m_document->addNonDerivedModel(m_alignmentModel);
        } else {
            QString error = alignmentModel->getError();
            ModelById::release(alignmentModel);
            emit failed(m_toAlign, error);
            return;
        }

    } else {

        // Have a tuning-difference transform id, so run it
        // asynchronously first
        
        TransformFactory *tf = TransformFactory::getInstance();

        Transform transform = tf->getDefaultTransformFor
            (tdId, aggregateModel->getSampleRate());

        transform.setParameter("maxduration", 60);
        transform.setParameter("maxrange", 6);
        transform.setParameter("finetuning", false);
    
        SVDEBUG << "MATCHAligner: Tuning difference transform step size " << transform.getStepSize() << ", block size " << transform.getBlockSize() << endl;

        ModelTransformerFactory *mtf = ModelTransformerFactory::getInstance();

        QString message;
        m_tuningDiffOutputModel = mtf->transform(transform,
                                                 m_aggregateModel,
                                                 message);

        auto tuningDiffOutputModel =
            ModelById::getAs<SparseTimeValueModel>(m_tuningDiffOutputModel);
        if (!tuningDiffOutputModel) {
            SVCERR << "Align::alignModel: ERROR: Failed to create tuning-difference output model (no Tuning Difference plugin?)" << endl;
            ModelById::release(alignmentModel);
            emit failed(m_toAlign, message);
            return;
        }

        other->setAlignment(m_alignmentModel);
        m_document->addNonDerivedModel(m_alignmentModel);
    
        connect(tuningDiffOutputModel.get(),
                SIGNAL(completionChanged(ModelId)),
                this, SLOT(tuningDifferenceCompletionChanged(ModelId)));
    }
}

void
MATCHAligner::tuningDifferenceCompletionChanged(ModelId tuningDiffOutputModelId)
{
    if (m_tuningDiffOutputModel.isNone()) {
        // we're done, this is probably a spurious queued event
        return;
    }
        
    if (tuningDiffOutputModelId != m_tuningDiffOutputModel) {
        SVCERR << "WARNING: MATCHAligner::tuningDifferenceCompletionChanged: Model "
               << tuningDiffOutputModelId
               << " is not ours! (ours is "
               << m_tuningDiffOutputModel << ")" << endl;
        return;
    }

    auto tuningDiffOutputModel =
        ModelById::getAs<SparseTimeValueModel>(m_tuningDiffOutputModel);
    if (!tuningDiffOutputModel) {
        SVCERR << "WARNING: MATCHAligner::tuningDifferenceCompletionChanged: Model "
               << tuningDiffOutputModelId
               << " not known as SparseTimeValueModel" << endl;
        return;
    }

    auto alignmentModel = ModelById::getAs<AlignmentModel>(m_alignmentModel);
    if (!alignmentModel) {
        SVCERR << "WARNING: MATCHAligner::tuningDifferenceCompletionChanged:"
               << "alignment model has disappeared" << endl;
        return;
    }
    
    int completion = 0;
    bool done = tuningDiffOutputModel->isReady(&completion);

    SVDEBUG << "MATCHAligner::tuningDifferenceCompletionChanged: model "
            << m_tuningDiffOutputModel << ", completion = " << completion
            << ", done = " << done << endl;
    
    if (!done) {
        // This will be the completion the alignment model reports,
        // before the alignment actually begins
        alignmentModel->setCompletion(completion / 2);
        return;
    }

    m_tuningFrequency = 440.f;
    
    if (!tuningDiffOutputModel->isEmpty()) {
        m_tuningFrequency = tuningDiffOutputModel->getAllEvents()[0].getValue();
        SVCERR << "MATCHAligner::tuningDifferenceCompletionChanged: Reported tuning frequency = " << m_tuningFrequency << endl;
    } else {
        SVCERR << "MATCHAligner::tuningDifferenceCompletionChanged: No tuning frequency reported" << endl;
    }    
    
    ModelById::release(tuningDiffOutputModel);
    m_tuningDiffOutputModel = {};
    
    beginAlignmentPhase();
}

bool
MATCHAligner::beginAlignmentPhase()
{
    TransformId id = getAlignmentTransformName(m_subsequence);
    
    SVDEBUG << "MATCHAligner::beginAlignmentPhase: transform is "
            << id << endl;
    
    TransformFactory *tf = TransformFactory::getInstance();

    auto aggregateModel =
        ModelById::getAs<AggregateWaveModel>(m_aggregateModel);
    auto alignmentModel =
        ModelById::getAs<AlignmentModel>(m_alignmentModel);

    if (!aggregateModel || !alignmentModel) {
        SVCERR << "MATCHAligner::alignModel: ERROR: One or other of the aggregate & alignment models has disappeared" << endl;
        return false;
    }
    
    Transform transform = tf->getDefaultTransformFor
        (id, aggregateModel->getSampleRate());

    transform.setStepSize(transform.getBlockSize()/2);
    transform.setParameter("serialise", 1);
    transform.setParameter("smooth", 0);
    transform.setParameter("zonewidth", 40);
    transform.setParameter("noise", true);
    transform.setParameter("minfreq", 500);

    int cents = 0;
    
    if (m_tuningFrequency != 0.f) {
        transform.setParameter("freq2", m_tuningFrequency);

        double centsOffset = 0.f;
        int pitch = Pitch::getPitchForFrequency(m_tuningFrequency,
                                                &centsOffset);
        cents = int(round((pitch - 69) * 100 + centsOffset));
        SVCERR << "MATCHAligner: frequency " << m_tuningFrequency
               << " yields cents offset " << centsOffset
               << " and pitch " << pitch << " -> cents " << cents << endl;
    }

    alignmentModel->setRelativePitch(cents);
    
    SVDEBUG << "MATCHAligner: Alignment transform step size "
            << transform.getStepSize() << ", block size "
            << transform.getBlockSize() << endl;

    ModelTransformerFactory *mtf = ModelTransformerFactory::getInstance();

    QString message;
    m_pathOutputModel = mtf->transform
        (transform, m_aggregateModel, message);

    if (m_pathOutputModel.isNone()) {
        transform.setStepSize(0);
        m_pathOutputModel = mtf->transform
            (transform, m_aggregateModel, message);
    }

    auto pathOutputModel =
        ModelById::getAs<SparseTimeValueModel>(m_pathOutputModel);

    //!!! callers will need to be updated to get error from
    //!!! alignment model after initial call
        
    if (!pathOutputModel) {
        SVCERR << "MATCHAligner: ERROR: Failed to create alignment path (no MATCH plugin?)" << endl;
        alignmentModel->setError(message);
        return false;
    }

    pathOutputModel->setCompletion(0);
    alignmentModel->setPathFrom(m_pathOutputModel);

    connect(pathOutputModel.get(), SIGNAL(completionChanged(ModelId)),
            this, SLOT(alignmentCompletionChanged(ModelId)));

    return true;
}

void
MATCHAligner::alignmentCompletionChanged(ModelId pathOutputModelId)
{
    if (pathOutputModelId != m_pathOutputModel) {
        SVCERR << "WARNING: MATCHAligner::alignmentCompletionChanged: Model "
               << pathOutputModelId
               << " is not ours! (ours is "
               << m_pathOutputModel << ")" << endl;
        return;
    }

    auto pathOutputModel =
        ModelById::getAs<SparseTimeValueModel>(m_pathOutputModel);
    if (!pathOutputModel) {
        SVCERR << "WARNING: MATCHAligner::alignmentCompletionChanged: Path output model "
               << m_pathOutputModel << " no longer exists" << endl;
        return;
    }
        
    int completion = 0;
    bool done = pathOutputModel->isReady(&completion);

    if (m_withTuningDifference) {
        if (auto alignmentModel =
            ModelById::getAs<AlignmentModel>(m_alignmentModel)) {
            if (!done) {
                int adjustedCompletion = 50 + completion/2;
                if (adjustedCompletion > 99) {
                    adjustedCompletion = 99;
                }
                alignmentModel->setCompletion(adjustedCompletion);
            } else {
                alignmentModel->setCompletion(100);
            }
        }
    }

    if (done) {
        m_incomplete = false;
        
        ModelById::release(m_pathOutputModel);
        m_pathOutputModel = {};

        emit complete(m_alignmentModel);
    }
}
} // end namespace sv