-
Notifications
You must be signed in to change notification settings - Fork 0
/
sound-data-parser.js
94 lines (73 loc) · 3.08 KB
/
sound-data-parser.js
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
'use strict';
function ParsedWave(buffer) {
const RIFF_HEADER_OFFSET = 0;
const FILE_SIZE_OFFSET = 4;
const RIFF_FORMAT_OFFSET = 8;
const SUBCHUNK1_ID_OFFSET = 12;
const AUDIO_FORMAT_OFFSET = 20;
const NUMBER_OF_CHANNELS_OFFSET = 22;
const SAMPLE_RATE_OFFSET = 24;
const BITS_PER_SAMPLE_OFFSET = 34;
const SUBCHUNK2_ID_OFFSET = 36;
const SUBCHUNK2_SIZE_OFFSET = 40;
const DATA_OFFSET = 44;
this.buffer = buffer;
this.riff = buffer.slice(RIFF_HEADER_OFFSET, RIFF_HEADER_OFFSET + 4).toString('utf8');
this.fileSize = buffer.readUInt32LE(FILE_SIZE_OFFSET);
this.riffType = buffer.slice(RIFF_FORMAT_OFFSET, RIFF_FORMAT_OFFSET + 4).toString('utf8');
this.subChunk1Id = buffer.slice(SUBCHUNK1_ID_OFFSET, SUBCHUNK1_ID_OFFSET + 4).toString('utf8');
this.audioFormat = buffer.readUInt16LE(AUDIO_FORMAT_OFFSET);
this.numberOfChannels = buffer.readUInt16LE(NUMBER_OF_CHANNELS_OFFSET);
this.sampleRate = buffer.readUInt32LE(SAMPLE_RATE_OFFSET);
this.bitsPerSample = buffer.readUInt16LE(BITS_PER_SAMPLE_OFFSET);
this.subChunk2Id = buffer.slice(SUBCHUNK2_ID_OFFSET, SUBCHUNK2_ID_OFFSET + 4).toString('utf8');
this.subChunk2Size = buffer.readUInt32LE(SUBCHUNK2_SIZE_OFFSET);
this.data = buffer.slice(DATA_OFFSET, this.subChunk2Size + DATA_OFFSET);
}
// Andrew - The bufferMapper function is going to accept a parsed wave-file and output
// an array of values corresponding to the data subchunk in a format which can
// be accepted as input to the neural network.
const bufferMapper = parsedWave => {
const SIXTEEN_BIT_ZERO = 32768;
const SIXTEEN_BIT_MAX = 65535;
parsedWave.neuralArray = [];
for (let i = 0; i < parsedWave.data.length; i += 2) {
const sample = parsedWave.data.readInt16LE(i);
const unsignedSample = sample + SIXTEEN_BIT_ZERO;
const sigmoidSample = unsignedSample / SIXTEEN_BIT_MAX;
parsedWave.neuralArray.push(sigmoidSample);
}
return parsedWave;
};
module.exports = data => {
const parsedWaveFile = new ParsedWave(data);
if (parsedWaveFile.riff !== 'RIFF') {
throw new TypeError('incorrect file type, must be RIFF format');
}
if (parsedWaveFile.fileSize > 10000000) {
throw new TypeError('file too large, please limit file size to less than 10MB');
}
if (parsedWaveFile.riffType !== 'WAVE') {
throw new TypeError('file must be a WAVE');
}
if (parsedWaveFile.subChunk1Id !== 'fmt ') {
throw new TypeError('the first subchunk must be fmt');
}
if (parsedWaveFile.audioFormat !== 1) {
throw new TypeError('wave file must be uncompressed linear PCM');
}
if (parsedWaveFile.numberOfChannels > 2) {
throw new TypeError('wave file must have 2 or less channels');
}
if (parsedWaveFile.sampleRate > 48000) {
throw new TypeError('wave file must have sample rate of less than 48k');
}
if (parsedWaveFile.bitsPerSample !== 16) {
throw new TypeError(`file's bit depth must be 16`);
}
if (parsedWaveFile.subChunk2Id !== 'data') {
throw new TypeError('subchunk 2 must be data');
}
const neuralMappedWaveFile = bufferMapper(parsedWaveFile);
return neuralMappedWaveFile;
};