1.CVC a me DSP ka leo hoemi:
Ke kūʻai nā mea kūʻai aku i nā poʻo Bluetooth, e lohe mau lākou i ka CVC a me DSP noise reduction functions i loaʻa i nā mea kālepa i ka hoʻolaha ʻana i nā headphone. No ka nui o nā mea hoʻohana i lohe i ka wehewehe ʻana, ʻaʻole maopopo ka nui o nā mea kūʻai aku i ka ʻokoʻa ma waena o nā mea ʻelua. ʻO ka ʻokoʻa, no kēlā pilikia ʻenehana, hele mākou i ka ʻepekema o nā mea ʻelua ma lalo o ke kumu hana a me ka ʻokoʻa.
ʻO DSP kahi pōkole no ka hana hōʻailona kikohoʻe. ʻO kāna kumu hana: hōʻiliʻili ka microphone i ka walaʻau kaiapuni o waho, a laila ma o ka hana hoʻohaʻahaʻa ʻana o ka walaʻau i loko o ka earphone, hoʻopiʻi ʻo ia e hana i kahi hawewe kani hoʻohuli e like me ka leo ambient, kahi e kāpae ai i ka walaʻau a pēlā e hoʻokō ai. ʻO ka hopena hoʻohaʻahaʻa leo maikaʻi.
He pōkole ʻo CVC no Clear Voice Capture. He polokalamu hoʻohaʻahaʻa noise ʻenehana. ʻO ke kumu o ka hoʻopaʻa ʻana i nā ʻano leo reverberation ma o ka polokalamu hoʻopau leo i kūkulu ʻia a me ka microphone.
Penei ka ʻokoʻa:
a. ʻOkoʻa ka mea, ʻo ka ʻenehana CVC ka mea nui no ka leo i hana ʻia i ka wā o ke kāhea ʻana, ʻo DSP ka mea nui no ka leo kiʻekiʻe a haʻahaʻa i ka leo o waho.
b. ʻO nā mea pōmaikaʻi ʻē aʻe, ʻo ka ʻenehana DSP ka mea e loaʻa ai ka loaʻa kālā pilikino i nā mea hoʻohana headset, a ʻo CVC ka mea e pōmaikaʻi ai i ka ʻaoʻao ʻē aʻe.
I ka hōʻuluʻulu manaʻo, hiki i nā headphone hoʻohana DSP a me CVC noise reduction technology ke hoʻemi pono i ka walaʻau o ke kaiapuni waho o ke kelepona, a hoʻomaikaʻi nui i ka maikaʻi o ke kelepona a me ke kani o nā headphone.
2.ANC ka leo hoemi:
E pili ana ka ANC i ka Active Noise Control, ka mea e hoʻemi ikaika i ka walaʻau. ʻO ke kumu kumu, ʻo ka ʻōnaehana hoʻohaʻahaʻa walaʻau e hana i nā hawewe kani hoʻohuli e like me ka walaʻau o waho, e hoʻokaʻawale i ka walaʻau. He kiʻi kiʻi kiʻi kiʻi kiʻi kiʻi 1 o ka leo kani hoʻopau ʻana i ka leo. Hoʻokomo ʻia ka chip ANC i loko o ka pepeiao pepeiao. ʻOhi ʻo Ref mic (reference microphone) i ka leo ambient ma nā pepeiao. Hapa mic (Error Microphone) E hōʻiliʻili i ke koena leo ma hope o ka emi ʻana o ka walaʻau ma ka pepeiao. Hoʻokani ʻo Speaker i ka anti-noise ma hope o ka hana ANC.
He kiʻi schematic o ka ʻōnaehana ANC ka Kiʻi 2, me ʻekolu papa, i hoʻokaʻawale ʻia e nā laina kaha. ʻO ke ala kumu kiʻekiʻe loa ʻo ia ke kahawai acoustic mai ka ref mic a i ka mic error, ʻo ka hana pane e hōʻike ʻia e P(z)P(z); ʻO ka papa waena ke kahawai analog, kahi o ke ala lua ke ala mai ka hoʻopuka kānana adaptive i ke koena hoʻihoʻi. E komo pū me DAC, kānana hana hou, hoʻonui mana, pāʻani leo, loaʻa hou, pre-amplifier, kānana anti-aliasing, ADC; ʻO ka papa lalo ke ala kikohoʻe, kahi e hoʻoponopono mau ai ka kānana adaptive i ka coefficient kaumaha kānana e hōʻemi i ke koena a hiki i ka convergence. ʻO ka hopena maʻamau ka hoʻokō ʻana i kahi kānana adaptive me ka hoʻohana ʻana i kahi kānana FIR i hui pū me ka LMS algorithm. E hoʻomaʻamaʻa i ke Kiʻi 2 a loaʻa i ka Figure 3.
E ʻae mai iaʻu e kamaʻilio pōkole e pili ana i nā loina o ka kānana adaptive a me LMS (Least mean square) algorithm, a laila Figure 3. E like me ka mea i hōʻike ʻia ma ka Figure 4, i hāʻawi ʻia i ka hoʻokomo xx a me ka dd i makemake ʻia, hoʻonui ka kānana adaptive i nā coefficients i kēlā me kēia hoʻololi. liʻiliʻi a liʻiliʻi ka ʻokoʻa ma waena o ka pukana yy a me dd a hiki i ka kokoke ʻana o ke koena i ka ʻole a hui pū. ʻO LMS kahi algorithm hou no nā kānana adaptive. ʻO ka hana pahuhopu o LMS ka huinahā o ka hewa hikiwawe e2(n)=(d(n)−y(n))2e2(n)=(d(n)−y(n))2, i mea e hoemi ai. ka hana pahuhopu, ʻO ka hoʻohana ʻana i ka iho gradient e hāʻawi i ke ʻano hou o ka algorithm. (ʻO ka manaʻo algorithmic o ka hoʻohana ʻana i ka gradient descent e hōʻemi i kahi pahuhopu a loaʻa i ke ʻano hou o ka parameter e ʻimi ʻia he mea maʻamau, e like me ka hoʻihoʻi laina. ) =w(n)+μe(n)x(n)w(n+1)=w(n)+μe(n)x(n), kahi ʻo μμ ka nui ʻanuʻu. Inā hoʻololi ʻia ka nui o ka μμ me ka hoʻololi ʻana, ʻo ia kahi algorithm LMS step-by-step.
E kamaʻilio e pili ana i ke Kiʻi 3. Eia ka hoʻopuka ʻana o ka kānana adaptive ma hope o S(z)S(z) e hoʻohālikelike me ka hoʻopuka makemake. S(z)S(z) ke kumu o ka pa'a ole. Ma ka palapala, "ʻaʻole pololei 'like' ka hōʻailona kuhi I ka manawa me ka hōʻailona kuhikuhi", ua haki ka convergence o ka LMS. (ʻAʻole au i ʻike i ke ʻano o ia mea T__T) ʻO kahi ala kūpono ʻo FXLMS (Filtered-X LMS), e hiki ai iā x (n) ke hoʻokomo i ka module LMS ma o Sˆ(z)S^(z), Sˆ( z S^(z) kahi koho o S(z)S(z). Pahuhopu o FXLMS:
E2(n)=(d(n)−s(n)∗[wT(n)x(n)])2,
E2(n)=(d(n)−s(n)∗[wT(n)x(n)])2,
No laila, gradient=−2e(n)s(n)∗x(n)−2e(n)s(n)∗x(n), kahi i ʻike ʻole ʻia ai s(n)s(n), me kona hoʻohālikelike ʻana, no laila ʻO ke kumu hoʻohou FXLMS
w(n+1)=w(n)+μe(n)x'(n),
w(n+1)=w(n)+μe(n)x'(n),
Kahi x'(n)=sˆ(n)∗x(n)x'(n)=s^(n)∗x(n).
Ke hui pū ka kānana hoʻololi, E(z)=X(z)P(z)−X(z)W(z)S(z)≈0E(z)=X(z)P(z)−X(z ) W(z)S(z) ≈ 0, no laila W(z) ≈ P(z) / S(z) W(z) ≈ P(z) / S(z). ʻO ia hoʻi, ua hoʻoholo ʻia ka coefficient kaumaha o ka kānana adaptive e ke ala mua a me ke ala lua o nā headphone. ʻO ke ala mua a me ke ala lua o ke poʻo poʻo he paʻa, no laila ua paʻa ka helu kaumaha o ka kānana adaptive. No laila, no ka maʻalahi, ua hoʻoholo ʻia nā koena kaumaha o nā headphone ANC o nā mea hana ma ka hale hana. ʻOiaʻiʻo, ʻaʻole maikaʻi ka ʻike hoʻolohe o kēia leo ANC e like me ka ANC earphone me ka manaʻo adaptive maoli, no ka mea i nā kūlana maoli, ʻo ka walaʻau o waho e pili ana i ke kuhikuhi o ka pepeiao, ʻokoʻa ka wela a me nā mea like. ka pane ala o ka pepeiao.
Hōʻoia Matlab
E kākau i ka code Matlab, me ka hoʻohana ʻana i ka kānana adaptive o ka nui o ka nui LMS, hōʻike ʻia nā hopena simulation ma ke kiʻi 5. Ma ka laulā o 0 a 2 kHz, hoʻohana ʻia ka feedforward ANC e hoʻopau i ka leo keʻokeʻo Gaussian, a ʻo ka hoʻohaʻahaʻa ʻana i ka leo he 30 dB + ma ka awelika. ʻO ka FXLMS ma ka waihona Matlab kahi paʻa-paʻa, a ʻoi aku ka maikaʻi o ka hopena.
Q&A
a. No ke aha ka ANC no ka leo haʻahaʻa haʻahaʻa ma lalo o 2 kHz?
Ma ka lima hoʻokahi, hiki i ke kani kino kino o nā headphone (passive noise reduction) hiki ke hoʻopaʻa pono i ka walaʻau kiʻekiʻe, ʻaʻole pono e hoʻohana i ka ANC e hōʻemi i ka leo kiʻekiʻe. Ma kaʻaoʻao'ē aʻe,ʻo ka leo nui ka lōʻihi o ka lōʻihi lōʻihi a hiki ke hoʻopaʻaʻia i kahi manawa lōʻihi,ʻoiai ke nānā aku neiʻo ACA i ke kani kiʻekiʻe.
b. Ke ʻoi aku ka nui o ka lohi uila ma mua o ka lohi mua, pehea e hoʻemi nui ai ka hana o ka algorithm?
P(z) liʻiliʻi, S(z) lohi nui, e like me P(z)=z-1, S(z)=z-2, aia wale nō ke hiki iā W(z)=z ke hoʻokō i nā koi, ʻaʻole -kumu, hiki ole ke loaa.
c. He aha ka ʻokoʻa ma waena o Feedforward ANC, feedforward ANC narrow-band, a me ANC pane?
Loaʻa i ka ʻōnaehana Feedforwad kahi mic ref a me kahi mic hewa e hōʻiliʻili i ka walaʻau waho a me nā hōʻailona koena o loko. Hoʻokahi wale nō mic hala ka ʻōnaehana pane, a ua hana ʻia ka hōʻailona kuhikuhi e ka mic hewa a me ka hoʻopuka kānana adaptive.
ʻO ka Broad-band feedforward ka hale i wehewehe ʻia ma luna. I loko o ka ʻāpana haiki, hoʻopuka ke kumu walaʻau i kahi hōʻailona hōʻailona hōʻailona, a hoʻokumu ka mea hana hōʻailona i kahi hōʻailona kuhikuhi no ka kānana adaptive. Pili wale no ka hoopau ana i ka walaau.
Manaʻo Manaʻo Hoʻohana ʻo ANC i ka mic hala e hoʻihoʻi i ka hōʻailona i hōʻiliʻili ʻia e ka ref mic i ka hoʻolālā feedforward no ka mea he hewa wale nō ka mic. ʻAʻole ʻoluʻolu ke ala i ke kumu kumu, no laila ua hoʻopau ʻia nā ʻāpana walaʻau i ʻike ʻia, ʻo ia hoʻi ka walaʻau haiki. Pono e hoʻomaopopo ʻia inā ʻaʻole ʻoluʻolu ka feedforward i ke kumu kumu, ʻo ia hoʻi, ʻoi aku ka lōʻihi o ka lohi uila ma mua o ka lohi acoustic channel nui, hiki iā ia ke hoʻopau wale i ka narrowband periodic noise.
Aia kekahi ʻano Hybrid ANC e komo pū ana me ka hoʻokumu ʻana i ka feedforward a me ka manaʻo manaʻo. ʻO ka pōmaikaʻi nui hiki iā ʻoe ke mālama i ke kauoha o ka kānana adaptive.
