Audio AI JAGORA

Mel-Frequency Cepstral Coefficients

Mel-Frequency Cepstral Coefficients (MFCCs) ƙaƙƙarfan jeri ne na lambobi waɗanda ke taƙaita sifar mitar sauti kamar yadda kunnuwan ɗan adam ke fahimce shi.

2 min karatuAn sabunta ta ƙarshe

Dubawa

For decades they were the workhorse feature for speech recognition, speaker identification, and music analysis.

Zurfafa nutsewa

MFCCs suna canza ɗan guntun odiyo zuwa lambobi kusan 13 waɗanda ke ɗaukar timbre. Bututun yana ɗaukar siginar igiyar ruwa, yana karya shi cikin firam ~ 25ms, yana ƙididdige bakan wutar lantarki ta hanyar canjin Fourier, sannan ya karkatar da axis a kan ma'aunin mel, wanda ke rataye makada kamar yadda cochlea ke yi: a ƙasa da 1kHz kuma a sama. Ƙarfin mel ɗin an matsar da su (suna kwaikwayon hasashe mai ƙarfi) kuma a ƙarshe sun wuce ta hanyar canza yanayin cosine, wanda ke ƙawata su kuma yana tattara bayanai cikin ƴan ƙididdiga na farko. Sakamakon yana da ƙarfi ga amo da sautin lasifika, wanda shine dalilin da ya sa na gargajiya Hidden Markov Model da Gaussian Mixture Model jawabai sun dogara da MMFCs kusan ko'ina a duniya kafin zurfin koyo.

Fahimtar Fasaha

Ma'auni na mel yana ƙayyadad da tsinkaye tare da mel = 2595 log10(1 + f/700), don haka daidaitattun matakan mel suna sauti daidai. Canjin cosine mai hankali na ƙarshe (DCT) shine matakin 'cepstral': yana ɗaukar bakan log-mel azaman sigina kuma yana raba nau'ikan nau'ikan muryoyin murya a hankali (ƙananan ƙididdigar cepstral, ɓangaren da muke kiyayewa) daga saurin daidaitawar sauti (masu ƙima, yawanci watsar), da keɓance ainihin sautin sauti daga farar lasifikar.

Dabarun Tasiri

Shiga ku isa

Yana inganta samun dama ta hanyar rubutu, ba da labari, da mu'amalar murya.

Kudin da kasafin kuɗi

Ƙungiyoyin kafofin watsa labaru na iya jigilar sauti mai gogewa cikin sauri tare da ƙaramin kasafin kuɗi.

Gudu da sikelin

Tsarin fuskantar abokin ciniki na iya aiwatar da hulɗar magana a mafi girman ma'auni.

Makomar Mel-Frequency Cepstral Coefficients

Ƙarshe-zuwa-ƙarshen zurfin cibiyoyin sadarwa suna ƙara koyan fasali kai tsaye daga raƙuman raƙuman raƙuman ruwa ko log-mel spectrograms, suna tsallake DCT, don haka MVCCs masu tsabta suna shuɗewa daga ASR na zamani. Duk da haka sun kasance shahararru don nauyi, kan na'ura, da ƙananan ayyuka masu ƙarancin bayanai: tsinkayar kalmomin mahimmanci, gano ayyukan murya, zanen yatsan sauti, da bioacoustics. Yi tsammanin MMFCs za su dage a matsayin ingantaccen tushe mai fa'ida kamar yadda koyan gaba-gaba suka mamaye manyan samfura.

Aiwatar da Gaskiyar Duniya

Fasalolin Acoustic don masu gane magana ta HMM-GMM kamar farkon Sphinx da tsarin HTK

Tabbatar da mai magana da diarization, tantance wanda ke magana akan kira

Rarraba nau'ikan kiɗa da buga yatsa na waƙa (Madaidaicin timbre irin na Shazam)

Gano kurakuran inji ko kiran dabba daga sauti a cikin masana'antu da saka idanu na bioacoustic

Hatsari & Tsare-tsare

Rashin amfani da murya da haɗarin kwaikwaya yana ƙaruwa lokacin da aka rasa izini.

Daidaituwa na iya faɗuwa cikin lafuzza, yaruka, ko mahalli masu hayaniya.

Ana iya kuskuren sauti na roba don ingantacciyar magana ba tare da bayyananniyar lakabi ba.

Taswirar Hanya

1

Sami tabbataccen izini don ɗaukar murya, cloning, da sake amfani.

2

Gwajin ingantattun masu magana daban-daban da yanayin baya.

3

Ƙayyade lokacin da dole ne ɗan adam ya duba ko ya amince da abubuwan da aka fitar.

4

Yi lakabin sauti na roba da kuma adana bayanan da aka tabbatar don yin lissafi.

Ci gaba da Bincike

Free newsletter

Get the daily AI briefing

Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.

One email each weekday. Unsubscribe in one click. We never sell or share your address.

Test yourself

Take the Mel-Frequency Cepstral Coefficients quiz

Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.

Fara tambayoyi

Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation

Tambayoyin da ake yawan yi

What is Mel-Frequency Cepstral Coefficients?

Mel-Frequency Cepstral Coefficients (MFCCs) ƙaƙƙarfan jeri ne na lambobi waɗanda ke taƙaita sifar mitar sauti kamar yadda kunnuwan ɗan adam ke fahimce shi. Shekaru da yawa sun kasance fasalin aikin dokin aiki don tantance magana, tantance lasifika, da nazarin kiɗa.

Menene 'mel' a cikin MCCC ke nufi?

Ma'aunin mel yana jujjuya mitar ta yadda matakan daidaita daidai suke su yi sauti daidai da nisa ga mutane, a sarari a sarari a ƙananan mitoci kuma da ƙarfi a manyan.

Wanne canji aka yi amfani da shi na ƙarshe don samar da ƙididdiga na cepsral?

Bayan an ƙididdige kuzarin log-mel, ƙwanƙwasa mai ƙwanƙwasa ta canza su ta ƙawata su kuma tana tattara bayanai cikin ƴan ƙididdiga na farko.

Me yasa MMFCs ke da ƙarfi ga murhun lasifika?

Ƙananan ƙananan ƙididdiga na cepstral suna ɗaukar ambulaf ɗin murya (abun ciki na sauti) yayin da manyan ke kama farar, don haka kiyaye ƙananan ƙananan yana rage girman farar.

Kusan adadin adadin MCCC nawa ake kiyayewa akan kowane firam?

Zaɓin gama gari shine kusan ƙididdiga 13, wani lokaci ana ƙara su tare da deltas ɗin su, waɗanda ke taƙaita timbre.

Me yasa ake amfani da log ɗin akan kuzarin mel-band?

Haushin sautin ɗan adam kusan logarithmic ne, don haka matsawar log ɗin yana sa fasalulluka sun fi dacewa da fahimtar daidaitawa kuma suna daidaita kewayo mai ƙarfi.