PESQ uye STOI Kutaura Kwemhando Metrics
PESQ uye STOI ndiwo akajairwa chinangwa metrics anoratidza kunaka kwekutaura kwakagadziriswa uye kuti zvinonzwisisika sei, pasina kuda vateereri vevanhu.
Pfupiso
They let engineers benchmark codecs, noise reducers, and speech-enhancement models automatically.
Kudzika Kwakadzika
PESQ (Perceptual Evaluation of Speech Quality), yakamisikidzwa seITU-T P.862, inofanotaura hunoonekwa kunaka kwekutaura, kunyanya parunhare uye kuongororwa kwekodeki. Inoenzanisa chiratidzo chereferensi chakachena kune chakaderedzwa uye chinoburitsa zvibodzwa pachiyero chakafanana neMOS (chinenge -0.5 kusvika 4.5), kutevedzera maonero ekunzwa kwevanhu. STOI (Short-Time Objective Intelligibility), yakaunzwa muna 2010, panzvimbo pacho inofanotaura kunzwisiswa: mangani mazwi anonzwisiswa nemuteereri. Inowiriranisa mahamvuropu enguva pfupi yekutaura kwakachena uye kwakagadziridzwa pamabhendi efrequency, ichiburitsa zvibodzwa kubva pa0 kusvika 1. Ose ari intrusive (reference-based) metrics. PESQ inopindura kuti 'zvinonzwika zvakanaka here?' nepo STOI ichipindura kuti 'unogona kuzvinzwisisa here?' Pamwe chete ndiwo maturusi ekuongorora akasarudzika ekusimudzira kutaura, denoising, uye derverberation masisitimu.
Technical Insight
Mametric ese ari maviri anopindirana: anoenzanisa referensi yakachena neiyo yakaderedzwa chiratidzo asati arova. PESQ inomepu masaini ese ari maviri pachiyero che psychoacoustic loudness (Bark bands), inokokorodza kukanganisa kwekuona nekufamba kwenguva, uye inodzoreredza kune kukosha kwakafanana neMOS. STOI inopatsanura kutaura kuita rimwe-chetatu-octave mabhendi, inotora mapfupi ~400 ms maemvuropu masegment, zvimedu nekuagadzirisa, yobva yacomta kuwirirana kuri pakati pereferensi nemahamvuropu akashatiswa. Avhareji yekubatana ikoko kunopa iyo 0-ku-1 yekunzwisisa mamaki.
Strategic Impact
Svika uye svika
Inonatsiridza kusvikika kuburikidza nekunyora, kurondedzera, uye mazwi ekubatanidza.
Mutengo uye bhajeti
Zvikwata zveMedia zvinogona kutumira odhiyo yakakwenenzverwa nekukurumidza nemabhajeti madiki.
Kumhanya uye chiyero
Masisitimu anotarisana nevatengi anogona kugadzirisa kutaurirana kwekutaura pamwero mukuru.
Ramangwana rePESQ uye STOI Kutaura Kwemhando Metrics
Nekuti PESQ neSTOI zvinoda referensi yakachena, tsvakiridzo iri kuchinjika ichienda kusiri-inopindira, referensi-yemahara metrics seDNSMOS neNISQA iyo inokora mhando kubva kune yakaderedzwa chiratidzo chete uchishandisa neural network. Mitsva mitsva yekudzidza-yakadzika inodzidziswawo kufanotaura MOS yemunhu zvakananga. Zvakadaro, PESQ neSTOI zvinoramba zvakadzika mabenchmarks, uye chimiro chakakosha chiri kuita kuti zvisiyaneke kuitira kuti zvigone kushandiswa zvakananga sekudzidzira kurasikirwa mabasa ematanho ekusimudzira kutaura kwete kungoongorora mushure meiyo-yechokwadi.
Real-World Implementation
Benchmarking yekusimudzira kutaura uye ruzha-kudzvinyirira modhi pane akajairwa bvunzo seti
Kuenzanisa foni uye VoIP codec mhando panguva network engineering
Tuning kunzwa-yamuro uye cochlear-implant kugadziridza kwepamusoro kunzwisisika
Kusimbisa deverberation algorithms mumusangano uye mapaipi anobatsira izwi
Njodzi & Guardrails
Kushandisa izwi zvisizvo uye njodzi dzekuedzesera dzinowedzera kana chibvumirano chisipo.
Kururama kunogona kudonha mumitauro, mataurirwo, kana nharaunda dzine ruzha.
Synthetic audio inogona kukanganisa kutaura kwechokwadi isina mavara akajeka.
Implementation Roadmap
Wana mvumo yakajeka yekutora inzwi, kugadzira, uye kushandisa zvakare.
Yedza mhando pavatauri vakasiyana uye mamiriro ekumashure.
Tsanangura apo munhu anofanira kuongorora kana kubvumidza zvabuda.
Label synthetic odhiyo uye chengetedza marekodhi ekuzvidavirira.
Ramba Uchiongorora
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 PESQ and STOI Speech Quality Metrics quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Gaidhi rinotevera
Matauriro Synthesis Hunhu
Mibvunzo inowanzo bvunzwa
What is PESQ and STOI Speech Quality Metrics?
PESQ uye STOI ndiwo akajairwa chinangwa metrics anoratidza kunaka kwekutaura kwakagadziriswa uye kuti zvinonzwisisika sei, pasina kuda vateereri vevanhu. Ivo vanorega mainjiniya benchmark macodecs, zvinoderedza ruzha, uye modhi-yekusimudzira kutaura.
PESQ inonyanya kuyera chii?
PESQ (ITU-T P.862) inofanotaura hunhu hwekutaura hunoonekwa pachiyero chakafanana neMOS.
Chii chinonzi STOI chinonyanya kufanotaura?
STOI inofungidzira kunzwisisika, kureva kuti kutaura kunonzwisisika sei, pachiyero che 0-ku-1.
Ose PESQ uye STOI anotsanangurwa se 'intrusive' metrics. Zvinorevei izvo?
Intrusive metrics inofananidza iyo yakaderedzwa siginecha inopesana neinongedzo yakachena kuverengera zvibodzwa.
Ndeupi huwandu hwezvibodzwa zveSTOI?
STOI inoburitsa kukosha pakati pe0 ne1, uko kukwirira kunoreva kunzwisisika.
PESQ modhi yekunzwa kwevanhu vachishandisa rudzii rwekunzwisisa frequency chiyero?
PESQ mepu masaini pane psychoacoustic ruzha inomiririra uchishandisa Bark-bhendi kugadzirisa.