| { |
| "paper_id": "O11-2002", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T08:05:32.312957Z" |
| }, |
| "title": "", |
| "authors": [], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "", |
| "pdf_parse": { |
| "paper_id": "O11-2002", |
| "_pdf_hash": "", |
| "abstract": [], |
| "body_text": [ |
| { |
| "text": "EQUATION", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [ |
| { |
| "start": 0, |
| "end": 8, |
| "text": "EQUATION", |
| "ref_id": "EQREF", |
| "raw_str": "+ f / 700)\uff0c\u800c\u983b\u5e36\u7684\u5bec\u5ea6\u70ba\u5169\u76f8\u9130\u6885\u723e\u523b\u5ea6\u5dee\u3002\u63a5\u8457\u8a08\u7b97\u5404\u983b\u5e36\u7684\u80fd\u91cf\u503c 2 , , b b u t b t j jl SG x \uf03d \uf03d \uf0e5 \uff0c \u5176\u4e2d l b \u70ba\u7b2c b \u500b\u983b\u5e36\u4e2d\u6700\u4f4e\u7684 FFT \u983b\u7387\u523b\u5ea6\uff0cu b \u70ba\u7b2c b \u500b\u983b\u5e36\u4e2d\u6700\u9ad8\u7684 FFT \u983b\u7387\u523b\u5ea6\uff0c \u7136\u5f8c\u518d\u5c07\u6240\u6709\u983b\u5e36\u7684\u5c0d\u6578\u80fd\u91cf\u503c\u9032\u884c\u96e2\u6563\u9918\u65cb\u8f49\u63db(Discrete Cosine Transform, DCT)[4]\uff0c \u4ee5\u53d6\u5f97\u5012\u983b\u8b5c\u4fc2\u6578(cepstral coefficients)[4]\uff0c\u8a73\u7d30\u8a08\u7b97\u65b9\u5f0f\u5982\u4e0b\uff1a \uf028 \uf029 \uf028 \uf029 2 , 1 1 log cos 0.5 b b u B t t j b b j l i x T j b BB \uf070 \uf03d\uf03d \uf0e6\uf0f6 \uf0e6\uf0f6 \uf03d\uf02d \uf0e7\uf0f7 \uf0e7\uf0f7 \uf0e7\uf0f7 \uf0e8\uf0f8 \uf0e8\uf0f8 \uf0e5\uf0e5 X (1) \u5176\u4e2d B \u70ba\u983b\u5e36\u7e3d\u6578\uff0cT b (j)\u70ba\u7b2c b \u500b\u983b\u5e36\u7684\u4e09\u89d2\u6ffe\u6ce2\u5668\u3002 3.2 \u9ad8\u65af\u6df7\u5408\u6a21\u578b(GMM) \u70ba\u4e86\u51dd\u805a\u540c\u7a2e\u9ce5\u4e4b\u4e0d\u540c\u53eb\u8072\u7684\u5171\u6709\u97f3\u8272\u7279\u5fb5\uff0c\u6211\u5011\u5229\u7528\u300c\u9ad8\u65af\u6df7\u5408\u6a21\u578b\u300d[4][5]\u4f86 \u9032\u884c MFCC \u53c3\u6578\u7d71\u8a08\u3002\u9ad8\u65af\u6df7\u5408\u6a21\u578b\u662f\u8072\u5b78\u8a0a\u865f\u5206\u985e\u4e2d\u6700\u5e38\u898b\u4e14\u6700\u6210\u529f\u7684\u6a21\u578b\u4e4b\u4e00\u3002\u4e00 \u500b\u9ad8\u65af\u6df7\u5408\u6a21\u578b\u5305\u542b\u82e5\u5e72\u9ad8\u65af\u6a5f\u7387\u5bc6\u5ea6\u51fd\u5f0f\uff0c\u6bcf\u4e00\u9ad8\u65af\u6a5f\u7387\u5bc6\u5ea6\u51fd\u5f0f\u4ee5\u671f\u671b\u503c\uf06d g \u8207\u8b8a\u7570 \u91cf\uf053 g \u6240\u63cf\u8ff0\uff0c\u5176\u4e2d g \u4ee3\u8868 G \u500b\u9ad8\u65af\u6a5f\u7387\u5bc6\u5ea6\u51fd\u5f0f\u4e2d\u4e4b\u7b2c g \u500b\uff0c\u53e6\u5916\u542b\u4e00\u52a0\u6b0a\u6578 w g \u5c07\u5404 \u9ad8\u65af\u6a5f\u7387\u5bc6\u5ea6\u51fd\u5f0f\u52a0\u7e3d\u6210\u70ba\u4e00\u6a5f\u7387\u5bc6\u5ea6\u3002\u6211\u5011\u5c07\u6a21\u578b\u53c3\u6578\u8a18\u70ba\uff1a () \u03bb T k = {w g , \uf06d g ,\uf053 g | 1 \uf0a3 g \uf0a3 G}\u3002\u9019\u4e9b\u53c3\u6578\u53ef\u7d93\u7531\u6700\u5927\u5316\u671f\u671b\u503c\u6cd5(Expectation-Maximization, EM)[4][5]\u4f30\u7b97\u51fa\uff0c\u6b64 \u5373\u8a13\u7df4\u968e\u6bb5\u6240\u9808\u57f7\u884c\u4e4b\u5de5\u4f5c\u3002\u82e5\u8cc7\u6599\u5eab\u4e2d\u6709 K \u96bb\u9ce5\uff0c\u5247\u6211\u5011\u7522\u751f K \u500b\u6a21\u578b ( ) ( ) ( ) 12 \u03bb , \u03bb , ..., \u03bb T T T K , \u4ee3\u8868\u9019\u4e9b\u9ce5\u9cf4\u8072\u7684\u97f3\u8272\u6a21\u578b\u3002 \u5728\u6e2c\u8a74\u968e\u6bb5\uff0c\u82e5\u6709\u4e00\u672a\u77e5\u9ce5\u9cf4\u8072\u7247\u6bb5\u4e4b\u7279\u5fb5\u5411\u91cf\u5e8f\u5217\u70ba X=X 1 ,X 2 ,\u2026,X T \uff0c\u5176\u4e2d\u6bcf\u4e00 \u5411\u91cf\u7dad\u5ea6\u70ba D\uff0c\u5247\u6211\u5011\u5c0d\u5404\u6a21\u578b () \u03bb T k \uff0c1 \uf0a3 k \uf0a3 K \u5206\u5225\u8a08\u7b97\u4f3c\u7136\u7387\uff1a \uf028 \uf029 ( ) ( ) 1 X|\u03bb (X |\u03bb ) K TT k t k k PP \uf03d \uf03d \uf0d5 (2) \u5176\u4e2d \uf028 \uf029 \uf028 \uf029 \uf028 \uf029 \uf028 \uf029 ( ) 1 \\ 1 11 X \u03bb exp X \u03bc \u03a3 X \u03bc 2 2 \u03a3 G T t k g t g g t g D g g Pw \uf070 \uf02d \uf03d \uf0a2 \uf0e6\uf0f6 \uf03d \uf02d \uf02d \uf02d \uf0e7\uf0f7 \uf0e8\uf0f8 \uf0e5 (3) \u5247\u6839\u64da\u6700\u5927\u4f3c\u7136\u7387\u6c7a\u7b56\u6cd5(maximum likelihood decision)[4]\uff0cX \u61c9\u5224\u65b7\u70ba () arg max (X | \u03bb ) T k k SP \uf03d", |
| "eq_num": "(4)" |
| } |
| ], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "\uf03d (8) 4.2 \u9ce5\u9cf4\u8072\u4e4b\u6a21\u578b\u5efa\u7acb \u70ba\u4e86\u6355\u6349\u97f3\u7b26\u9593\u76f8\u9023\u7684\u52d5\u614b\u8cc7\u8a0a\uff0c\u6211\u5011\u63a1\u7528\u96d9\u9023\u6587\u6a21\u578b(Bi-gram Model)\u6280\u8853\u4f86\u8a13 \u7df4\u5c6c\u65bc\u9ce5\u9cf4\u8072\u7684\u6a21\u578b\u3002 \u7576\u7d66\u5b9a\u4e00\u4e32\u97f3\u7b26\u5e8f\u5217 o 1 ,o 2 ,\u2026o t ,\u2026\uff0c\u96d9\u9023\u6587\u6a21\u578b\u53ef\u7528\u4ee5\u63cf\u8ff0\u8a72\u5e8f\u5217\u4e2d\u5404\u97f3\u7b26\u524d\u5f8c\u76f8 \u9023\u63a5\u7684\u95dc\u4fc2\uff0c\u5176\u505a\u6cd5\u662f\u7d71\u8a08\u6240\u6709\u53ef\u80fd\u4e4b\u5169\u97f3\u7b26\u7d44\u5408\uff0c\u4f8b\u5982 w 1 \u8207 w 2 \u7684\u767c\u751f\u983b\u7387\u6216\u6a5f\u7387\uff1a P(o t = w 1 | o t-1 = w 2 )\u3002\u82e5\u5169\u500b\u97f3\u7b26 w 1 \u8207 w 2 \u5e38\u51fa\u73fe\u65bc\u67d0\u4e00\u7a2e\u9ce5\u7684\u9cf4\u8072\u4e2d\uff0c\u5247\u6a5f\u7387 P(o t = w 1 | o t-1 = w 2 )\u503c\u6703\u8f03\u5927\uff1b\u53cd\u4e4b\uff0c\u82e5\u5169\u500b\u97f3\u7b26\u5e7e\u4e4e\u4e0d\u6703\u51fa\u73fe\u65bc\u67d0\u4e00\u7a2e\u9ce5\u7684\u9cf4\u8072\u4e2d\uff0c\u5247\u6a5f\u7387 P(o t = w 1 | o t-1 = w 2 )\u503c\u6703\u8f03\u5c0f\u3002 \u5728\u6e2c\u8a74\u968e\u6bb5\uff0c\u82e5\u6709\u4e00\u672a\u77e5\u9ce5\u9cf4\u8072\u7247\u6bb5\u4e4b\u97f3\u7b26\u5e8f\u5217\u70ba O=O 1 ,O 2 ,\u2026,O T \uff0c\u5247\u6211\u5011\u5c0d\u5404\u6a21 \u578b\uf06c k (P) \uff0c1 \uf0a3 k \uf0a3 K \u5206\u5225\u8a08\u7b97\u4f3c\u7136\u7387\uff1a \uf028 \uf029 ( ) (", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Template-based automatic recognition of birdsong syllables from continuous recordings", |
| "authors": [ |
| { |
| "first": "S", |
| "middle": [ |
| "E" |
| ], |
| "last": "Anderson", |
| "suffix": "" |
| }, |
| { |
| "first": "A", |
| "middle": [ |
| "S" |
| ], |
| "last": "Dave", |
| "suffix": "" |
| }, |
| { |
| "first": "D", |
| "middle": [], |
| "last": "Margoliash", |
| "suffix": "" |
| } |
| ], |
| "year": 1996, |
| "venue": "J. Acoust. Soc. Amer", |
| "volume": "100", |
| "issue": "2", |
| "pages": "1209--1219", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "S. E. Anderson, A. S. Dave, and D. Margoliash, -Template-based automatic recognition of birdsong syllables from continuous recordings,\u2016 J. Acoust. Soc. Amer., vol. 100, no. 2, pp. 1209-1219, Aug. 1996.", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Automated recognition of bird song elements from continuous recordings using dynamic time warping and hidden Markov models: A comparative study", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Kogan", |
| "suffix": "" |
| }, |
| { |
| "first": "D", |
| "middle": [], |
| "last": "Margoliash", |
| "suffix": "" |
| } |
| ], |
| "year": 1998, |
| "venue": "\u2016 J. Acoust. Soc. Amer", |
| "volume": "103", |
| "issue": "4", |
| "pages": "2187--2196", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "J. Kogan and D. Margoliash, -Automated recognition of bird song elements from continuous recordings using dynamic time warping and hidden Markov models: A comparative study,\u2016 J. Acoust. Soc. Amer., vol. 103, no. 4, pp. 2187-2196, Apr. 1998.", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "Comparison of Parametric Representations for Monosyllabic Word Recognition in Continuously Spoken Sentences", |
| "authors": [ |
| { |
| "first": "S", |
| "middle": [ |
| "B" |
| ], |
| "last": "Davis", |
| "suffix": "" |
| }, |
| { |
| "first": "P", |
| "middle": [], |
| "last": "Mermelstein", |
| "suffix": "" |
| } |
| ], |
| "year": 1980, |
| "venue": "\u2016 IEEE Trans. on Acoustic, Speech and Signal Processing", |
| "volume": "28", |
| "issue": "4", |
| "pages": "357--366", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "S. B. Davis and P. Mermelstein, -Comparison of Parametric Representations for Monosyllabic Word Recognition in Continuously Spoken Sentences,\u2016 IEEE Trans. on Acoustic, Speech and Signal Processing., vol. 28, no. 4, pp. 357-366, Aug. 1980.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Spoken Language Processing", |
| "authors": [ |
| { |
| "first": "X", |
| "middle": [], |
| "last": "Huang", |
| "suffix": "" |
| }, |
| { |
| "first": "A", |
| "middle": [], |
| "last": "Acero", |
| "suffix": "" |
| }, |
| { |
| "first": "H", |
| "middle": [ |
| "W" |
| ], |
| "last": "Hon", |
| "suffix": "" |
| } |
| ], |
| "year": 2001, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "X. Huang, A. Acero, and H. W. Hon, Spoken Language Processing, Prentice Hall, 2001.", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "Gaussian Mixture Modeling of Short-Time Fourier Transform Features for Audio Fingerprinting,\u2016 IEEE Transactions on Information Forensics and Security", |
| "authors": [ |
| { |
| "first": "A", |
| "middle": [], |
| "last": "Ramalingam", |
| "suffix": "" |
| }, |
| { |
| "first": "S", |
| "middle": [], |
| "last": "Krishnan", |
| "suffix": "" |
| } |
| ], |
| "year": 2006, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "A. Ramalingam and S. Krishnan, -Gaussian Mixture Modeling of Short-Time Fourier Transform Features for Audio Fingerprinting,\u2016 IEEE Transactions on Information Forensics and Security, 2006.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Speaker Verification Using Adapted Gaussian Mixture Models,\u2016 Digital Signal Processing", |
| "authors": [ |
| { |
| "first": "D", |
| "middle": [], |
| "last": "Reynolds", |
| "suffix": "" |
| }, |
| { |
| "first": "T", |
| "middle": [], |
| "last": "Quatieri", |
| "suffix": "" |
| } |
| ], |
| "year": 2000, |
| "venue": "", |
| "volume": "10", |
| "issue": "", |
| "pages": "19--41", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "D. Reynolds and T. Quatieri, -Speaker Verification Using Adapted Gaussian Mixture Models,\u2016 Digital Signal Processing 10, PP. 19-41, 2000.", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Predicting musical pitch from component frequency ratios\u2016", |
| "authors": [ |
| { |
| "first": "M", |
| "middle": [], |
| "last": "Piszczalski", |
| "suffix": "" |
| }, |
| { |
| "first": "B", |
| "middle": [ |
| "A" |
| ], |
| "last": "Galler", |
| "suffix": "" |
| } |
| ], |
| "year": 1979, |
| "venue": "Journal of the Acoustical Society of America", |
| "volume": "66", |
| "issue": "3", |
| "pages": "710--720", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "M. Piszczalski and B. A. Galler, -Predicting musical pitch from component frequency ratios\u2016, Journal of the Acoustical Society of America, 66(3), pp.710-720, 1979.", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "TABREF0": { |
| "type_str": "table", |
| "num": null, |
| "content": "<table><tr><td colspan=\"3\">\u672c\u8ad6\u6587\u7684\u7ae0\u7bc0\u7d44\u7e54\u5982\u4e0b\uff1a\u7b2c\u4e8c\u7ae0\u63cf\u8ff0\u8fa8\u8b58\u7cfb\u7d71\u67b6\u69cb\uff1b\u7b2c\u4e09\u7ae0\u8aaa\u660e\u97f3\u8272\u7279\u5fb5\u53c3\u6578\u64f7\u53d6</td></tr><tr><td colspan=\"3\">\u53ca\u7d71\u8a08\u6a21\u578b\uff1b\u7b2c\u56db\u7ae0\u63cf\u8ff0\u97f3\u9ad8\u7279\u5fb5\u53c3\u6578\u64f7\u53d6\u53ca\u7d71\u8a08\u6a21\u578b\uff1b\u7b2c\u4e94\u7ae0\u7d50\u5408\u97f3\u8272\u8207\u97f3\u9ad8\u7279\u5fb5\u4f86</td></tr><tr><td colspan=\"3\">\u8a2d\u8a08\u8fa8\u8b58\u7cfb\u7d71\uff1b\u7b2c\u516d\u7ae0\u5c07\u8aaa\u660e\u672c\u7814\u7a76\u6240\u4f7f\u7528\u7684\u9ce5\u9cf4\u8072\u8cc7\u6599\u5eab\u4e26\u5448\u73fe\u5be6\u9a57\u7d50\u679c\uff1b\u7b2c\u516d\u7ae0\u9032</td></tr><tr><td>\u884c\u672c\u7814\u7a76\u4e4b\u7e3d\u7d50\u3002</td><td/><td/></tr><tr><td>\u5ed6\u5049\u6069</td><td>\u9ece\u6b23\u6377</td><td>\u8521\u5049\u548c</td></tr><tr><td>\u570b\u7acb\u53f0\u5317\u79d1\u6280\u5927\u5b78 2. \u8fa8\u8b58\u7cfb\u7d71\u67b6\u69cb</td><td>\u570b\u7acb\u53f0\u5317\u79d1\u6280\u5927\u5b78</td><td>\u570b\u7acb\u53f0\u5317\u79d1\u6280\u5927\u5b78</td></tr><tr><td>\u96fb\u8166\u8207\u901a\u8a0a\u7814\u7a76\u6240</td><td>\u96fb\u8166\u8207\u901a\u8a0a\u7814\u7a76\u6240</td><td>\u96fb\u8166\u8207\u901a\u8a0a\u7814\u7a76\u6240</td></tr><tr><td colspan=\"3\">t8418044@ntut.edu.tw \u672c\u8ad6\u6587\u6240\u63d0\u51fa\u7684\u8fa8\u8b58\u7cfb\u7d71\u5982\u5716 1\u3002\u6211\u5011\u53ef\u5c07\u5176\u5206\u6210\u4e09\u500b\u55ae\u5143\uff0c\u5206\u5225\u662f\u300c\u97f3\u8272\u5206\u6790\u300d\u3001 t9419004@ntut.edu.tw whtsai@ntut.edu.tw</td></tr><tr><td colspan=\"3\">\u300c\u97f3\u9ad8\u5206\u6790\u300d\u3001 \u8207\u300c\u6574\u5408\u6c7a\u7b56\u300d \u3002\u5404\u55ae\u5143\u7686\u5305\u542b\u5169\u7a2e\u64cd\u4f5c\u6a21\u5f0f\uff0c\u4e00\u70ba\u8a13\u7df4\u3001\u53e6\u4e00\u70ba\u6e2c\u8a74\uff0c</td></tr><tr><td>\u7c21\u4ecb\u5982\u4e0b\uff1a</td><td>\u5716 2 \u97f3\u8272\u5206\u6790\u904e\u7a0b\u3002</td><td/></tr><tr><td colspan=\"3\">(1) \u97f3\u8272\u5206\u6790 3.1 \u7279\u5fb5\u53c3\u6578\u64f7\u53d6 \u6b64\u55ae\u5143\u76ee\u6a19\u662f\u64f7\u53d6\u5404\u7a2e\u9ce5\u4e4b\u9cf4\u8072\u7684\u97f3\u8272\u7279\u5fb5\uff0c\u4e26\u8868\u793a\u70ba\u7d71\u8a08\u6a21\u578b\uff0c\u4ee5\u4fbf\u8b58\u5225\u672a\u77e5\u9ce5 \u97f3\u8272\u4e4b\u5dee\u7570\u5982\u540c\u92fc\u7434\u8207\u5409\u4ed6\u90fd\u5f48\u594f\u76f8\u540c\u97f3\u7b26\uff0c\u537b\u807d\u8d77\u4f86\u6709\u4e0d\u540c\u7684\u8072\u97f3\u3002\u4e0d\u540c\u7684\u9ce5\u7a2e\uff0c \u9cf4\u8072\u3002 \u53ef\u8996\u70ba\u4e0d\u540c\u7684\u6a02\u5668\u4e00\u6a23\uff0c\u97f3\u8272\u4e5f\u4e0d\u76f8\u540c\u3002 (2) \u97f3\u9ad8\u5206\u6790 \u9996\u5148\uff0c\u6211\u5011\u5148\u5c07\u9ce5\u9cf4\u8072\u8a0a\u865f\u7d93\u7531\u77ed\u6642\u9593(short-term)\u5feb\u901f\u5085\u5229\u8449\u8f49\u63db(Fast Fourier \u6b64\u55ae\u5143\u76ee\u6a19\u662f\u64f7\u53d6\u5404\u7a2e\u9ce5\u4e4b\u9cf4\u8072\u7684\u97f3\u9ad8\u7279\u5fb5\uff0c\u4e26\u8868\u793a\u70ba\u7d71\u8a08\u6a21\u578b\uff0c\u4ee5\u4fbf\u8b58\u5225\u672a\u77e5\u9ce5 \u9cf4\u8072\u3002 Transform, FFT)\u6210\u70ba\u4e00\u4e32\u97f3\u6846\u983b\u8b5c\u5e8f\u5217\u3002\u4ee4 x t, j</td></tr><tr><td>(3) \u6574\u5408\u6c7a\u7b56</td><td/><td/></tr><tr><td colspan=\"3\">\u6b64\u55ae\u5143\u6574\u5408\u97f3\u8272\u8207\u97f3\u9ad8\u7684\u5224\u65b7\u8cc7\u8a0a\uff0c\u9032\u884c\u6700\u5f8c\u6c7a\u7b56\u4ee5\u544a\u77e5\u4f7f\u7528\u8005\u8fa8\u8b58\u7d50\u679c\u3002</td></tr><tr><td>1. \u524d\u8a00</td><td/><td/></tr><tr><td colspan=\"3\">\u76ee\u524d\u5168\u4e16\u754c\u5927\u7d04\u6709\u4e5d\u5343\u4e03\u767e\u591a\u7a2e\u9ce5\u985e\uff0c\u800c\u53f0\u7063\u9019\u6a23\u4e00\u500b\u5c0f\u5cf6\u5c31\u4f54\u4e86\u7d04\u4e8c\u5341\u5206\u4e4b\u4e00\u7684</td></tr><tr><td colspan=\"3\">\u7a2e\u985e\uff0c\u96d6\u7136\u6211\u5011\u5468\u906d\u4f4f\u6709\u8a31\u591a\u9019\u4e9b\u53ef\u611b\u7684\u9130\u5c45\uff0c\u4f46\u5f80\u5f80\u90fd\u53ea\u807d\u5230\u5b83\u5011\u7684\u53eb\u8072\uff0c\u537b\u4e0d\u77e5\u5b83</td></tr><tr><td colspan=\"3\">\u5011\u662f\u8ab0\u3002\u9ce5\u985e\u7684\u9cf4\u8072\u8c50\u5bcc\u4e14\u591a\u8b8a\uff0c\u6211\u5011\u671f\u671b\u85c9\u7531\u7269\u7a2e\u4e4b\u9593\u7684\u9cf4\u8072\u5dee\u7570\u6027\uff0c\u767c\u5c55\u51fa\u4e00\u5957\u9ce5</td></tr><tr><td colspan=\"3\">\u9cf4\u8072\u8fa8\u8b58\u7cfb\u7d71\uff0c\u8b93\u4e0d\u662f\u9ce5\u985e\u5c08\u5bb6\u7684\u4e00\u822c\u6c11\u773e\uff0c\u4e5f\u53ef\u4ee5\u5f9e\u81ea\u5df1\u96a8\u610f\u9304\u88fd\u7684\u4e00\u6bb5\u9ce5\u9cf4\u8072\u97f3\u6a94</td></tr><tr><td colspan=\"2\">\u4e2d\uff0c\u8b93\u7cfb\u7d71\u5224\u65b7\u6240\u5c6c\u9ce5\u7a2e\u4e26\u7372\u5f97\u4e4b\u76f8\u95dc\u8a0a\u606f\u3002</td><td/></tr><tr><td colspan=\"3\">\u76ee\u524d\u9ce5\u9cf4\u8072\u81ea\u52d5\u8fa8\u8b58\u7684\u76f8\u95dc\u7814\u7a76\u4ecd\u5341\u5206\u6709\u9650\u3002\u6587\u737b[1]\u4e2d\u4f7f\u7528\u52d5\u614b\u6642\u9593\u6821\u6b63(Dynamic</td></tr><tr><td/><td>\u5716 1 \u672c\u8ad6\u6587\u4e4b\u9ce5\u9cf4\u8072\u8fa8\u8b58\u7cfb\u7d71\u67b6\u69cb\u5716\u3002</td><td/></tr><tr><td>3. \u97f3\u8272\u5206\u6790</td><td/><td/></tr><tr><td colspan=\"3\">\u97f3\u8272\u5206\u6790\u904e\u7a0b\u5982\u5716 2 \u6240\u793a\u3002\u4e3b\u8981\u5305\u62ec\u9810\u8655\u7406\u3001\u7279\u5fb5\u53c3\u6578\u64f7\u53d6\u3001\u8207\u7d71\u8a08\u6a21\u578b\u5efa\u7acb\u53ca\u5339\u914d\u3002</td></tr><tr><td colspan=\"3\">\u81ea\u65bc\u97f3\u8272(timbre)\u7279\u5fb5\u3002\u672c\u8ad6\u6587\u6240\u63d0\u51fa\u4e4b\u8fa8\u8b58\u7cfb\u7d71\uff0c\u9664\u4e86\u8003\u616e\u97f3\u8272\u7279\u5fb5\u53c3\u6578\u5916\uff0c\u66f4\u52a0\u4e0a\u97f3</td></tr><tr><td colspan=\"3\">\u9ad8(pitch)\u7279\u5fb5\u53c3\u6578\u3002\u5be6\u9a57\u7d50\u679c\u767c\u73fe\u5c07\u9019\u5169\u7a2e\u7279\u5fb5\u53c3\u6578\u9032\u884c\u7d50\u5408\u5f8c\u53ef\u6709\u6548\u63d0\u5347\u9ce5\u9cf4\u8072\u8fa8\u8b58</td></tr><tr><td>\u6b63\u78ba\u7387\u3002</td><td/><td/></tr></table>", |
| "html": null, |
| "text": "Time Warping\uff0cDTW)\u6f14\u7b97\u6cd5\uff0c\u5c07\u6e2c\u8a74\u8072\u97f3\u6a94\u7684\u983b\u8b5c\u5716(spectrogram)\u8207\u4e8b\u5148\u7d93\u904e\u5c08\u5bb6\u6311 \u9078\u7684\u6a23\u677f\u97f3\u6a94\u4f5c\u6bd4\u5c0d\u3002\u6587\u737b[2]\u4e2d\u5206\u5225\u6bd4\u8f03 DTW \u548c\u96b1\u85cf\u5f0f\u99ac\u53ef\u592b\u6a21\u578b(Hidden Markov Model\uff0cHMM)\u5728\u8fa8\u8b58\u9ce5\u985e\u8072\u97f3\u4e0a\u7684\u6548\u80fd\uff0c\u5176\u4e2d\u4f7f\u7528 6 \u7a2e\u7279\u5fb5\u53c3\u6578\uff1a\u7dda\u6027\u9810\u4f30\u7de8\u78bc\u53c3\u6578 (Linear Predict Coding\uff0cLPC)\u3001\u7dda\u6027\u9810\u4f30\u5012\u983b\u8b5c\u4fc2\u6578\u3001LPC reflection\u3001\u6885\u723e\u5012\u983b\u8b5c\u4fc2\u6578 (Mel-frequency Cepstral Coefficients\uff0cMFCC) [3]\u3001log mel-filter bank channel \u548c linear mel-filter bank channel\u3002\u5be6\u9a57\u7d50\u679c\u986f\u793a\uff0c\u4f7f\u7528 DTW \u7684\u8fa8\u8b58\u6548\u80fd\u4e0d\u932f\uff0c\u4f46\u662f\u5c0d\u65bc\u96dc\u8a0a\u8f03\u5927 \u7684\u8f38\u5165\u8072\u97f3\u6a94\u6216\u662f\u5bb9\u6613\u6df7\u6dc6\u7684\u77ed\u4fc3\u9cf4\u53eb\u8072\uff0c\u5247\u9700\u8981\u6311\u9078\u66f4\u9069\u5408\u7684\u983b\u8b5c\u5716\u6a23\u672c\uff0c\u9019\u9053\u624b\u7e8c \u901a\u5e38\u9700\u8981\u7d93\u9a57\u8c50\u5bcc\u7684\u5c08\u5bb6\u53cd\u8986\u5730\u8a74\u9a57\u3002\u5c0d\u65bc\u4f7f\u7528 HMM\uff0c\u8fa8\u8b58\u6548\u679c\u5247\u53d6\u6c7a\u65bc\u8f38\u5165\u53c3\u6578\u7684 \u9451\u5225\u5ea6\uff0c\u4f46\u6548\u80fd\u4e0d\u4e00\u5b9a\u6bd4 DTW \u597d\u3002\u800c\u4e0d\u8ad6\u662f DTW \u6216 HMM\uff0c\u6240\u4f7f\u7528\u7684\u8fa8\u8b58\u7dda\u7d22\u7686\u4f86 \u70ba\u7b2c t \u500b\u97f3\u6846\u8a0a\u865f\u5728\u7b2c j \u500b FFT \u983b\u7387\u523b \u5ea6\u4e0a\u7684\u7d55\u5c0d\u632f\u5e45\u983b\u8b5c(magnitude spectrum) \uff0c\u5176\u4e2d 1 \uf0a3 j \uf0a3 J\uff0c\u800c J \u70ba\u983b\u7387\u523b\u5ea6\u7684\u7e3d\u6578\u3002 \u518d\u5c07\u5404\u97f3\u6846\u7684\u7d55\u5c0d\u632f\u5e45\u983b\u8b5c\u9001\u5165\u4e00\u500b\u300c\u4e09\u89d2\u983b\u5e36\u7d44\u300d (triangular filter banks) \uff0c\u5176\u4e2d \u983b\u5e36\u4e2d\u5fc3\u5373\u662f\u6885\u723e\u523b\u5ea6\u503c\uff0c\u8a72\u523b\u5ea6\u503c mel \u8207\u983b\u7387 f Hz \u7684\u8f49\u63db\u65b9\u5f0f\u70ba mel(f) = 2595 \uf0d7 log 10 (1" |
| }, |
| "TABREF1": { |
| "type_str": "table", |
| "num": null, |
| "content": "<table><tr><td colspan=\"6\">\u5716 4 \u97f3\u9ad8\u6c42\u53d6\u6d41\u7a0b\u5716</td><td/></tr><tr><td colspan=\"7\">\u5728\u5be6\u4f5c\u4e0a\uff0c\u8003\u616e\u9ce5\u9cf4\u8072\u8cc7\u6599\u5eab\u53ef\u80fd\u4e26\u7121\u5305\u542b\u5927\u91cf\u7684\u8072\u97f3\u6a23\u672c\u53ef\u4f9b\u7cbe\u78ba\u7684\u6a21\u578b\u8a13\u7df4\uff0c \u56e0\u6b64\u6211\u5011\u85c9\u7531\u6a21\u578b\u8abf\u9069\u6280\u8853[6]\u4f86\u7522\u751f\u500b\u500b\u9ce5\u7a2e\u7684\u6a21\u578b\u3002\u8a72\u6280\u8853\u5148\u6839\u64daEM\u6f14\u7b97\u6cd5\u5c07\u6240\u6709\u8a13 \u7df4\u7528\u4e4b\u9ce5\u9cf4\u8072\u898b\u7acb\u4e00\u500b\u9ad8\u65af\u6df7\u5408\u6a21\u578b\u7576\u4f5c\u901a\u7528\u6a21\u578b(Universal Model) \uff0c\u518d\u7d93\u7531\u6700\u5927\u4e8b\u5f8c \u6a5f\u7387(Maximum A Posterior, MAP)\u8abf\u9069\u6cd5\u9032\u884c\u901a\u7528\u6a21\u578b\u4e4b\u8abf\u6574\uff0c\u4ee5\u7522\u751f\u5404\u9ce5\u7a2e\u7684\u9ad8\u65af \u6df7\u5408\u6a21\u578b\u3002 \u4ee4 x t,j \uf028 \uf029 \uf028 \uf029 2 12* log 69.5 \uf0e8\uf0f8 440 freq j Nj \uf0e6\uf0f6 \uf03d\uf02b \uf0e7\uf0f7 (5)</td></tr><tr><td colspan=\"7\">4. \u97f3\u9ad8\u5206\u6790 \u97f3\u9ad8\u53c3\u6578\u8207\u8072\u97f3\u7684\u57fa\u983b(fundamental frequency, F0)\u6709\u95dc\uff0c\u4e00\u9023\u4e32\u57fa\u983b\u9ad8\u4f4e\u4e0d\u540c\u7684 , , , ( ) max n t n t j j N j e y x \uf022 \uf03d \uf03d (6)</td></tr><tr><td colspan=\"7\">\u8072\u97f3\u4e32\u5728\u4e00\u8d77\u5c31\u5982\u540c\u4e0d\u540c\u97f3\u7b26\u88ab\u6f14\u594f\u51fa\u4e00\u822c\u3002\u672c\u55ae\u5143\u7684\u57fa\u672c\u6982\u5ff5\u662f\u5047\u8a2d\u6bcf\u7a2e\u9ce5\u90fd\u6709\u5176\u5404</td></tr><tr><td colspan=\"7\">\u81ea\u7684\u6b4c\u8072\u6216\u6b4c\u5531\u8a9e\u8a00\uff0c\u50cf\u662f\u97f3\u7b26\u9ad8\u4f4e\u76f8\u9023\u6709\u5176\u7368\u7279\u7684\u898f\u5247\u3002\u82e5\u6211\u5011\u80fd\u6355\u6349\u6bcf\u7a2e\u9ce5\u7684\u97f3\u7b26</td></tr><tr><td colspan=\"7\">\u76f8\u9023\u63a5\u8cc7\u8a0a\uff0c\u5247\u53ef\u64da\u6b64\u8b58\u5225\u672a\u77e5\u7684\u9ce5\u9cf4\u8072\u6240\u5c6c\u9ce5\u7a2e\u985e\u3002\u5982\u57163\u6240\u793a\u70ba\u97f3\u9ad8\u5206\u6790\u904e\u7a0b\uff0c\u4e3b</td></tr><tr><td colspan=\"4\">\u8981\u5305\u62ec\u97f3\u9ad8\u7279\u5fb5\u64f7\u53d6\u53ca\u7d71\u8a08\u6a21\u578b\u5efa\u7acb\u8207\u5339\u914d\u3002</td><td/><td/><td/></tr><tr><td/><td/><td>C</td><td/><td/><td/><td/></tr><tr><td>z</td><td>, t n</td><td>0 \uf03d \uf0e5 c \uf03d</td><td>, t n h y \uf02b c</td><td>12</td><td>c</td><td>(7)</td></tr><tr><td colspan=\"6\">\u5176\u4e2d C \u662f\u6b32\u5217\u5165\u8003\u616e\u7684\u516b\u5ea6\u97f3\u7b26\u6578\uff0c\u800c h \u70ba\u5c0f\u65bc 1 \u7684\u6b0a\u91cd\u3002\u64da\u6b64\u5224\u5b9a\u6f14\u5531\u97f3\u7b26\u61c9\u70ba</td><td/></tr><tr><td/><td>t</td><td colspan=\"2\">arg max</td><td colspan=\"2\">, t n</td><td/></tr><tr><td/><td/><td>1</td><td/><td/><td/><td/></tr><tr><td colspan=\"6\">\u57163 \u97f3\u9ad8\u5206\u6790\u904e\u7a0b\u3002</td><td/></tr><tr><td>4.1 \u7279\u5fb5\u53c3\u6578\u64f7\u53d6</td><td/><td/><td/><td/><td/><td/></tr><tr><td colspan=\"7\">\u6211\u5011\u63a1\u7528\u6b21\u8ae7\u6ce2\u7e3d\u548c\u6cd5(Sub-Harmonic Summation, SHS)[7]\u9032\u884c\u97f3\u9ad8\u6c42\u53d6\u3002SHS</td></tr><tr><td colspan=\"7\">\u7684\u539f\u7406\u662f\u6839\u64da\u57fa\u983b\u9664\u4e86\u672c\u8eab\u7684\u80fd\u91cf\u8f03\u9ad8\u5916\uff0c\u5176\u500d\u983b\u8ae7\u6ce2 (Harmonic) \u7684\u80fd\u91cf\u4e5f\u901a\u5e38\u8f03\u9ad8\uff0c</td></tr><tr><td colspan=\"7\">\u82e5\u67d0\u4e00\u57fa\u983b\u8207\u5176\u500d\u983b\u8ae7\u6ce2\u7684\u80fd\u91cf\u7e3d\u548c\u660e\u986f\u8f03\u9ad8\u65bc\u5176\u4ed6\u983b\u7387\u6642\uff0c\u8a72\u983b\u7387\u6975\u6709\u53ef\u80fd\u70ba\u57fa\u983b\u3002</td></tr><tr><td>\u5176\u6c42\u53d6\u97f3\u9ad8\u6d41\u7a0b\u5982\u57164\u3002</td><td/><td/><td/><td/><td/><td/></tr></table>", |
| "html": null, |
| "text": "\u70ba\u7b2c t \u500b\u97f3\u6846\u8a0a\u865f\u5728\u7b2c j \u500b FFT \u983b\u7387\u523b\u5ea6\u4e0a\u7684\u632f\u5e45\uff0c\u5176\u4e2d 1 \uf0a3 j \uf0a3 J\uff0c\u800c J \u70ba \u983b\u7387\u523b\u5ea6\u7e3d\u6578\u3002\u900f\u904e\u65b9\u7a0b\u5f0f 5 \u5c07 FFT \u523b\u5ea6\u8f49\u70ba MIDI \u97f3\u7b26\u523b\u5ea6 e 1 , e 2 ,\u2026, e N" |
| }, |
| "TABREF2": { |
| "type_str": "table", |
| "num": null, |
| "content": "<table><tr><td colspan=\"10\">\u516b\u4ee5\u5c71\u570b\u5bb6\u68ee\u6797\u904a\u6a02\u5340\u5e38\u898b\u9ce5\u985e\u9cf4\u8072 \u8868 4 \u57fa\u65bc\u97f3\u8272\u4e4b\u9ce5\u9cf4\u8072\u8fa8\u8b58\u6df7\u6dc6\u77e9\u9663\u3002 \u884c\u653f\u9662\u8fb2\u696d\u59d4\u54e1\u6703\u6797\u52d9\u5c40 \u7d05\u5634\u9ed1\u9d6f 0% 6.36% 0% 1.27% 9.55% 0% 65.6% 1.27% 15.92%</td><td>0%</td></tr><tr><td colspan=\"10\">) (| \u03bb ) PP 1 tT k t k t P O \uf03d \uf03d \uf0d5 \u884c\u653f\u9662\u8fb2\u696d\u59d4\u54e1\u6703\u6797\u52d9\u5c40 \u5c0f\u5377\u5c3e \u5c0f\u5544\u6728 \u5c0f\u5f4e\u5634 \u5c71\u7d05\u982d \u4e94\u8272\u9ce5 |\u03bb P O \u5927\u96ea\u5c71\u570b\u5bb6\u68ee\u6797\u904a\u6a02\u5340\u5e38\u898b\u9ce5\u985e\u9cf4\u8072 \u767d\u8033 \u756b\u7709 \u7d05\u5634 \u9ed1\u9d6f \u7d2b\u562f\u9d6f \u9ec3\u5634 \u7d2b\u562f\u9d6f 0% 0% 4% 0% 4% 12% 0% 80% 0% \u9ec3\u5634\u89d2\u9d1e 7.2% 21.62% 5.4% 0% 0% 1.8% 0% 0% 63.96% \u89d2\u9d1e \uf03d \u5247\u6839\u64da\u6700\u5927\u4f3c\u7136\u7387\u6c7a\u7b56\u6cd5(maximum likelihood decision)[4]\uff0cO \u61c9\u5224\u65b7\u70ba \u5c0f\u5377\u5c3e 64.93% 10.38% 0% 9.09% 0% 6.49% 6.49% 0% 2.59% \u6a39\u9d72 4.1% 0% 5.47% 0% 0% 0% 21.91% 0% 2.73% 65.75% (9) 0% \u6a39\u9d72 0% 0% \u8868 2 \u7db2\u8def\u4e0a\u6536\u96c6\u9ce5\u9cf4\u8072\u4e4b\u4f86\u6e90\u3002 \u7db2\u7ad9\u540d\u7a31 \u5c0f\u5544\u6728 9.8% 62.74% 14.7% 2.94% 3.92% 2.94% 2.94% 0% 0% 0% \u7db2\u7ad9\u4f4d\u7f6e \u53f0\u7063\u5927\u5b78\u52d5\u7269\u535a\u7269\u9928 \u5c71\u7d05\u982d 7.4% 3.7% 6.17% 70.37% 6.17% 3.7% 0% 2.46% 0% 0% http://archive.zo.ntu.edu.tw/ \u5c0f\u5f4e\u5634 0% 6.45% 82.58% 0% 0% 0% 0% 6.45% 0% 4.51% 7. \u7d50\u8ad6\u8207\u672a\u4f86\u5c55\u671b</td></tr><tr><td colspan=\"10\">() arg max (O | \u03bb ) P k k SP \uf03d 5. \u7d50\u5408\u97f3\u8272\u8207\u97f3\u9ad8\u5206\u6790\u4e4b\u8fa8\u8b58\u7cfb\u7d71 \u7d93\u5be6\u9a57\u767c\u73fe\uff0c\u4f7f\u7528\u97f3\u8272\u5206\u6790\u8207\u97f3\u9ad8\u5206\u6790\u6240\u7372\u5f97\u7684\u9ce5\u9cf4\u8072\u8fa8\u8b58\u7d50\u679c\u6709\u8a31\u591a\u5dee\u7570\u8207\u4e92\u88dc (10) \u4e4b\u8655\uff0c\u56e0\u6b64\u5617\u8a74\u7d50\u5408\u97f3\u8272\u8207\u97f3\u9ad8\u4e4b\u8fa8\u8b58\u7cfb\u7d71\uff0c\u5982\u5716 6 \u6240\u793a\u3002 \u570b\u7acb\u9cf3\u51f0\u8c37\u9ce5\u5712 http://www.fhk.gov.tw/ Bird Call Recordings http://www.geocities.com/RainForest/9003/birdcall.htm Macaulay Library http://macaulaylibrary.org/index.do \u4e94\u8272\u9ce5 0% 0.91% 0% 0% 74.88% 0% 13.69% 0% 6.84% 3.65% \u767d\u8033\u756b\u7709 3.48% 0% 8.13% 0% 0% 76.74% 5.81% 5.81% 0% \u672c\u8ad6\u6587\u63d0\u51fa\u7d50\u5408\u97f3\u8272\u8207\u97f3\u9ad8\u7dda\u7d22\u4f86\u8fa8\u8b58\u9ce5\u985e\u9cf4\u8072\u3002\u5be6\u9a57\u767c\u73fe\uff0c\u7576\u6211\u5011\u55ae\u7368\u7528\u97f3\u8272\u6216 0% \u7d05\u5634\u9ed1\u9d6f 0% 6.39% 0% 4.45% 8.28% 0% 68.78% 2.54% 15.92% \u97f3\u9ad8\u7dda\u7d22\u4f86\u8fa8\u8b58\u5341\u7a2e\u9ce5\u7684\u9cf4\u8072\u6642\uff0c\u50c5\u6709 1 \u81f3 2 \u7a2e\u9ce5\u7684\u9cf4\u8072\u8fa8\u8b58\u7387\u9ad8\u65bc 80%\uff1b\u4f46\u82e5\u5c07\u5169\u7a2e 0% \u7d2b\u562f\u9d6f 0% 0% 16% 0% 0% 8% 0% 76% 0% \u7279\u5fb5\u7d50\u5408\u5f8c\uff0c\u5c07\u6709 4 \u7a2e\u9ce5\u8fa8\u8b58\u7387\u8d85\u904e 80%\uff0c\u9a57\u8b49\u7d50\u5408\u4f7f\u7528\u97f3\u8272\u53ca\u97f3\u9ad8\u4e4b\u65b9\u6cd5\u80fd\u63d0\u5347\u9ce5\u9cf4 0% \u9ec3\u5634\u89d2\u9d1e 2.7% 13.51% 0% 9% 0% 7.2% 0% 0% 67.56% 0% \u8072\u8fa8\u8b58\u6548\u80fd\u3002 \u9019\u4e9b\u5546\u696d CD \u548c\u7db2\u7ad9\u5305\u542b\u5404\u7a2e\u4e0d\u540c\u9ce5\u985e\u7684\u9cf4\u8072\uff0c\u5728\u8003\u616e\u5230\u8a13\u7df4\u8a9e\u6599\u91cf\u7684\u591a\u5be1\u3001\u5730\u57df \u6a39\u9d72 2.73% 0% 10.95% 0% 0% 0% 21.91% 0% 8.21% 56.16% \u7531\u65bc\u53d7\u9650\u65bc\u9ce5\u9cf4\u8072\u97f3\u6a94\u7684\u6536\u96c6\u6578\u91cf\u6709\u9650\uff0c\u6240\u4ee5\u672c\u7bc7\u8ad6\u6587\u53ea\u6311\u9078\u5927\u53f0\u5317\u5730\u5340\u5e38\u898b\u7684\u5341 \u7684\u5408\u7406\u6027\u3001\u79d1\u76ee\u5dee\u7570\u548c\u9cf4\u8072\u985e\u578b\u4e0a\uff0c\u672c\u7bc7\u8ad6\u6587\u5171\u6311\u9078\u4e86 10 \u7a2e\u5927\u53f0\u5317\u5730\u5340\u5e38\u898b\u7684\u9ce5\u7a2e\u4f86 \u7a2e\u9ce5\u7a2e\uff0c\u5c0d\u65bc\u7cfb\u7d71\u7684\u5f37\u5065\u6027\u548c\u5be6\u7528\u6027\u9084\u9700\u8981\u52a0\u5165\u66f4\u591a\u9ce5\u7a2e\u4f86\u9a57\u8b49\u3002\u76f8\u8f03\u65bc\u76ee\u524d\u4eba\u8072\u8fa8\u8b58 \u9032\u884c\u5be6\u9a57\uff0c\u5206\u5225\u70ba\u5c0f\u5377\u5c3e\u3001\u5c0f\u5544\u6728\u3001\u5c0f\u5f4e\u5634\u756b\u7709\u3001\u5c71\u7d05\u982d\u3001\u4e94\u8272\u9ce5\u3001\u767d\u8033\u756b\u7709\u3001\u7d05\u5634\u9ed1 \u89c0\u5bdf\u73fe\u6709\u9ce5\u9cf4\u8072\u8cc7\u6599\u5eab\uff0c\u5176\u57fa\u983b\u7bc4\u570d\u4ecb\u65bc 366 Hz \u81f3 8591 Hz\uff0c\u5c0d\u61c9\u97f3\u7b26\u70ba 66-120\u3002 \u9d6f\u3001\u7d2b\u562f\u9d87\u3001\u9ec3\u5634\u89d2\u9d1e\u3001\u6a39\u9d72\u3002 6.2.2 \u57fa\u65bc\u97f3\u9ad8\u65b9\u6cd5\u5206\u6790\u4e4b\u9ce5\u8072\u8fa8\u8b58\u7d50\u679c \u7684\u6210\u679c\u4f86\u8aaa\uff0c\u9ce5\u9cf4\u8072\u8fa8\u8b58\u7cfb\u7d71\u9084\u6709\u6975\u5927\u7684\u6539\u9032\u7a7a\u9593\u3002</td></tr><tr><td colspan=\"10\">\u4e00\u6211\u5011\u5c07\u5546\u696d CD \u97f3\u6a94\u53ca\u7db2\u8def\u4e0a\u97f3\u6a94\u5169\u7a2e\u4f86\u6e90\u5408\u4f75\u5f8c\u5404\u53d6\u4e00\u534a\uff0c\u4f7f\u5f97\u8a13\u7df4\u97f3\u6a94\u8207\u6e2c \u5634\u756b\u7709\u53ca\u4e94\u8272\u9ce5\u7684\u8fa8\u8b58\u7387\u9ad8\u65bc 80%\uff0c\u5176\u9918\u7684\u9451\u5225\u5ea6\u90fd\u9817\u4f4e\u3002 \u8a74\u97f3\u6a94\u4e4b\u6a94\u6848\u6578\u4e0d\u6703\u5dee\u7570\u592a\u5927\u3002 \u57fa\u65bc\u97f3\u9ad8\u5206\u6790\u4e4b\u8fa8\u8b58\u7d50\u679c\u5982\u8868 5\uff0c\u7e3d\u8fa8\u8b58\u7387\u70ba 72.09%\u3002\u82e5\u50c5\u89c0\u5bdf\u8868\u4e2d\u9ce5\u9cf4\u8072\uff0c\u50c5\u6709\u5c0f\u5f4e \u53c3\u8003\u6587\u737b</td></tr><tr><td colspan=\"7\">\u8868 5 \u57fa\u65bc\u97f3\u9ad8\u4e4b\u9ce5\u9cf4\u8072\u8fa8\u8b58\u6df7\u6dc6\u77e9\u9663\u3002 \u5c0f\u5377\u5c3e \u5c0f\u5544\u6728 \u5c0f\u5f4e\u5634 \u5c71\u7d05\u982d \u4e94\u8272\u9ce5 \u767d\u8033 \u756b\u7709 \u7d05\u5634 \u9ed1\u9d6f 61.03% 19.48% 0% 10.38% 0% 5.19% 3.89% 2.94% 71.56% 12.74% 0% 1.96% 0% 3.92% \u4f7f\u7528\u5546\u696d 6.2 \u5be6\u9a57\u7d50\u679c \u5c0f\u5377\u5c3e \u5c0f\u5544\u6728 \u5c0f\u5f4e\u5634 0% 7.74% 82.58% 0% 0% 0% 0%</td><td>\u7d2b\u562f\u9d6f 0% 0% 7.74%</td><td>\u9ec3\u5634 \u89d2\u9d1e 0% 2.94% 0%</td><td>\u6a39\u9d72 0% 3.92% 1.93%</td></tr><tr><td colspan=\"10\">\u5716 6 \u6574\u5408\u97f3\u8272\u8207\u97f3\u9ad8\u4e4b\u8fa8\u8b58\u7cfb\u7d71\u67b6\u69cb\u5716\u3002 0% 7.4% 75.3% 2.46% 1.23% 0% \u57fa\u65bc\u97f3\u8272\u5206\u6790\u4e4b\u9ce5\u8072\u8fa8\u8b58\u7d50\u679c 1.23% 0% 1.36% 0% 1.82% 82.19% 0% 11.41% \u9996\u5148\u9032\u884c\u4e0d\u540c\u9ad8\u65af\u6df7\u5408\u6578\u7684\u9ce5\u9cf4\u8072\u8fa8\u8b58\u5be6\u9a57\uff0c\u7d50\u679c\u5217\u65bc\u8868 3\u3002\u5f9e\u8868\u4e2d\u5f97\u77e5\uff0c\u7576\u6df7\u5408 6.2.1 \u5c71\u7d05\u982d 0% 0% 0% \u4e94\u8272\u9ce5 0% 2.73% 0.45 % \u767d\u8033\u756b\u7709 0% 0% 11.62% 0% 0% 76.74% 5.81% 5.81% 0% 0% \u6578\u70ba 64 \u6642\uff0c\u5176\u8fa8\u8b58\u7387\u512a\u65bc\u5176\u4ed6\u6df7\u5408\u6578\u3002\u56e0\u6b64\u5728\u672c\u7bc7\u8ad6\u6587\u4e2d\uff0cGMM \u4e4b\u6df7\u5408\u6578\u7686\u70ba 64\u3002 \u7d05\u5634\u9ed1\u9d6f 0% 6.36% 0% 2.54% 9.55% 0% 63.05% 2.54% 15.92% 0%</td></tr><tr><td colspan=\"10\">\u5176\u8fa8\u8b58\u7d50\u679c\u5982\u8868 4\uff0c\u7e3d\u8fa8\u8b58\u7387\u70ba 71.08%\uff0c\u5176\u8fa8\u8b58\u7387\u4e4b\u8a08\u7b97\u65b9\u5f0f\u5982(12)\u3002\u82e5\u50c5\u89c0\u5bdf\u8868\u4e2d\u9ce5 \u7d2b\u562f\u9d6f 0% 0% 4% 0% 12% 28% 0% 56% 0% 0% \u5c07\u4e00\u672a\u77e5\u9ce5\u7a2e\u7684\u9cf4\u8072\u8a0a\u865f\uff0c\u7d93\u7531\u97f3\u8272\u5206\u6790\u8207\u97f3\u9ad8\u5206\u6790\u5f8c\u7522\u751f\u5169\u4f3c\u7136\u7387\uff0c\u518d\u5c07\u9019\u5169\u4f3c \u9cf4\u8072\uff0c\u50c5\u6709\u5c0f\u5f4e\u5634\u756b\u7709\u7684\u8fa8\u8b58\u7387\u8f03\u9ad8\uff0c\u5176\u9918\u7684\u9451\u5225\u5ea6\u90fd\u9817\u4f4e\u3002 \u9ec3\u5634\u89d2\u9d1e 7.2% 21.62% 5.4% 0% 0% 0.09% 0% 0% 64.86% 0% \u7136\u7387\u505a\u52a0\u6b0a\u7e3d\u548c\uff0c\u6700\u5f8c\u6311\u9078\u52a0\u7e3d\u5f8c\u7684\u6700\u5927\u4f3c\u7136\u7387\u3002\u64da\u6b64\u5224\u65b7\u8a72\u9ce5\u9cf4\u8072\u70ba\u4f55\u7a2e\u9ce5\uff0c\u5373\uff1a \u6a39\u9d72 5.4% 0% 8.21% 0% 0% 0% 24.65% 0% 2.73% 58.9%</td></tr><tr><td colspan=\"10\">^( 1 arg max[ P(X| kk ) ( ) )+ P(O| TP kK S \uf061 \uf06c \uf062 \uf06c \uf03d \uf0d7 \uf0d7 100% )] \uf0b4 \u88ab\u6b63\u78ba\u8fa8\u8b58\u51fa\u7684\u97f3\u6a94\u6578\u76ee \u8fa8\u8b58\u7387\u6b63\u78ba\u7387= \u7e3d\u6e2c\u8a66\u97f3\u6a94\u6578\u76ee \u5c07\u97f3\u8272\u53ca\u97f3\u9ad8\u65b9\u6cd5\u7d50\u5408\u5f8c\u4e4b\u9ce5\u9cf4\u8072\u8fa8\u8b58\u7d50\u679c \uf0a3\uf0a3 \u5982\u7b2c 5 \u7ae0\u4ecb\u7d39\uff0c\u6211\u5011\u5c07\u97f3\u8272\u53ca\u97f3\u9ad8\u5169\u7a2e\u65b9\u6cd5\u7d50\u5408\u3002\u5728(11)\u5f0f\u4e2d\uff0c\u6211\u5011\u5c07 GMM \u8a08\u7b97 (12) 6.2.3 (11)</td></tr><tr><td colspan=\"10\">\u5728\u672c\u7bc7\u8ad6\u6587\u4e2d\uff0c\u03b1 \u8207 \u03b2 \u5206\u5225\u8a2d\u5b9a\u6210 0.6 \u8207 0.4\u3002 6. \u9ce5\u9cf4\u8072\u8fa8\u8b58\u7cfb\u7d71 \u5f8c\u7684\u4f3c\u7136\u7387 \uf028 \uf029 () P X | T k \uf06c \u53ca Bigram \u8a08\u7b97\u51fa\u7684 \uf028 P O | P () k \uf06c \u8868 3 \u4e0d\u540c\u9ad8\u65af\u6df7\u5408\u6578\u4e4b\u8fa8\u8b58\u7387\u3002 \uf029 \u5206\u5225\u4e58\u4e0a \u03b1 \u8207 \u03b2 \u4e4b\u6b0a\u91cd\uff0c\u5728\u672c\u7bc7\u8ad6 \u9ad8\u65af\u6df7\u5408\u6578 4 8 16 32 64 \u6587\u88e1\uff0c\u03b1=0.6\u3001\u03b2=0.4\u3002 128 \u5c0f\u5377\u5c3e 55.84% 58.44% 58.44% 59.74% 64.93% \u89c0\u5bdf\u8868 6\uff0c\u7d50\u5408\u5f8c\u4e4b\u9cf4\u8072\u7e3d\u8fa8\u8b58\u7387\u9054 75.04%\u3002\u89c0\u5bdf\u8868\u4e2d\u9ce5\u9cf4\u8072\u4e4b\u8fa8\u8b58\u7387\uff0c\u5c0f\u5f4e\u5634\u756b 62.33% \u5c0f\u5544\u6728 59.8% 59.8% 60.78% 62.74% 62.74% 62.74% \u7709\u3001\u4e94\u8272\u9ce5\u3001\u767d\u8033\u756b\u7709\u53ca\u7d2b\u562f\u9d6f\u7684\u8fa8\u8b58\u7387\u7686\u9ad8\u65bc 80%\uff0c\u4e14\u6700\u4f4e\u7684\u8fa8\u8b58\u7387\u4e5f\u9ad8\u65bc 60%\u3002</td></tr><tr><td colspan=\"10\">\u5c0f\u5f4e\u5634 6.1 \u9ce5\u9cf4\u8072\u8cc7\u6599\u5eab 80% \u5c71\u7d05\u982d 69.13% \u8868 6 \u6574\u5408\u97f3\u8272\u3001\u97f3\u9ad8\u65b9\u6cd5\u5f8c\u4e4b\u9ce5\u9cf4\u8072\u8fa8\u8b58\u6df7\u6dc6\u77e9\u9663\u3002 81.93% 83.22% 83.22% 82.58% 69.13% 70.37% 71.6% 70.37% \u672c\u8ad6\u6587\u6240\u4f7f\u7528\u7684\u9ce5\u9cf4\u8072\u97f3\u6a94\u6709\u5169\u500b\u4f86\u6e90\uff0c\u5206\u5225\u662f\u5e02\u9762\u4e0a\u8ca9\u552e\u7684\u5546\u696d CD \u53ca\u7db2\u8def\u4e0a\u6536 81.93% 69.13% \u96c6\u4f86\u7684\u9ce5\u9cf4\u8072\u97f3\u6a94\uff0c\u6574\u7406\u5982\u8868 1\u30012 \u6240\u793a\uff1a \u4e94\u8272\u9ce5 72.14% 73.05% 73.97% 74.42% 74.88% 74.88% \u767d\u8033\u756b\u7709 74.41% 75.58% 77.9% 77.9% 76.744% 76.74% \u5c0f\u5377\u5c3e \u5c0f\u5544\u6728 \u5c0f\u5f4e\u5634 \u5c71\u7d05\u982d \u4e94\u8272\u9ce5 \u767d\u8033 \u756b\u7709 \u7d05\u5634 \u9ed1\u9d6f \u7d2b\u562f\u9d6f \u9ec3\u5634 \u89d2\u9d1e \u6a39\u9d72</td></tr><tr><td>\u7d05\u5634\u9ed1\u9d6f \u5c0f\u5377\u5c3e</td><td>62.42% 67.53% 12.98%</td><td>63.69% 0%</td><td>9.09%</td><td>64.96% 0%</td><td colspan=\"2\">66.24% 5.19% 2.59%</td><td>68.78% 0%</td><td>2.59%</td><td>67.51% 0%</td></tr><tr><td>\u7d2b\u562f\u9d6f \u5c0f\u5544\u6728</td><td>68% 2.94% 75.49%</td><td colspan=\"4\">\u8868 1 \u5546\u696d CD \u4e4b\u4f86\u6e90\u3002 72% 76% 76% 9.8% 0% 0.09% 0%</td><td>3.92%</td><td>76% 0%</td><td>2.94%</td><td>76% 3.92%</td></tr><tr><td>\u9ec3\u5634\u89d2\u9d1e \u5c0f\u5f4e\u5634 \u6a39\u9d72 \u5c71\u7d05\u982d \u7e3d\u8fa8\u8b58\u7387 \u4e94\u8272\u9ce5 \u767d\u8033\u756b\u7709</td><td colspan=\"2\">\u5c08\u8f2f\u540d\u7a31 63.96% 0% 5.8% \u9ce5-\u91ce\u9ce5\u9cf4\u5531\u5716\u9451 85.16% 63.96% 50.68% 52.05% 1.23% 0% 6.17% \u53f0\u5317\u9ce5\u8996\u754c 67.12% 68.23% 0% 1.36% 0% 0% 0% 10.46%</td><td>0% 75.3% 1.82% 0%</td><td>63.96% 0% 56.16% 2.46% 69.52% 83.1% 0%</td><td colspan=\"3\">63.96% 0% \u51fa\u7248\u5546 0% 57.53% 1.23% 0% \u98a8\u6f6e\u6709\u8072\u51fa\u7248\u6709\u9650\u516c\u53f8 67.56% 5.8% 56.16% 1.23% 70.25% 71.08% 0% 9.13% 0% \u53f0\u5317\u5e02\u653f\u5e9c\u65b0\u805e\u8655 80.23% 4.65% 3.48%</td><td colspan=\"2\">65.76% 3.2% 54.79% 0% 70.25% 3.19% 0% 0% 1.36% 0% 0%</td></tr></table>", |
| "html": null, |
| "text": "CD \u4e0a\u7684\u97f3\u6a94\uff0c\u539f\u59cb\u97f3\u6a94\u683c\u5f0f\u4e00\u5f8b\u70ba\u96d9\u8072\u9053 44.1KHz \u7684 PCM WAV \u6a94\u3002\u81f3 \u65bc\u7db2\u8def\u4e0a\u6536\u96c6\u800c\u4f86\u7684\u97f3\u6a94\uff0c\u539f\u59cb\u683c\u5f0f\u6709\u4e9b\u662f\u7d93\u904e MP3 \u58d3\u7e2e\u904e\u7684\u97f3\u6a94\uff0c\u4e5f\u6709 PCM \u683c\u5f0f\u7684 WAV \u6a94\uff0c\u800c\u53d6\u6a23\u983b\u7387\u5f9e 8KHz \u81f3 48KHz \u4e0d\u7b49\u3002\u70ba\u4e86\u5be6\u9a57\u7684\u4e00\u81f4\u6027\u53ca\u7bc0\u7701\u904b\u7b97\u6642\u9593\uff0c\u6211 \u5011\u628a\u5169\u8005\u4f86\u6e90\u7684\u97f3\u6a94\u7686\u8abf\u6574\u70ba 22.05KHz \u55ae\u8072\u9053\u7684 PCM WAV \u6a94\u3002" |
| } |
| } |
| } |
| } |
