A7 B7 經濟部中央標率局員工消費合作社印製 407435 五、發明説明（1 ) 本發明之技術範園 本發明乃關於通訊ΐ周故gn > miPA〇f 题1揭路及關於在目_路上供語音通迅 之語言處理。 相關技術之敘述 一典型國際網路電話利用，-聲板，—麥克風及二 個揚聲器/麥克風及揚聲器通常肢桌予上彼此相鄰。此 -構型會造成在接收機端有如回音之相當大之串音。此一 回音必須加以抑制以使國際網路可用。 曰 在GSM中，吾人均了解#行動電話用户通話或不通話時 使用VAD (聲音活㈣測）偵測。此一資訊用以在輸送聲音 時降低«。在不連續語音編碼中，根據ν〇χ原則（聲音 操作發射）、，-VAD單元負責债測所收到之聲順序是否代 表人類語言。此VAD單元可有二個狀態，第一 聲音順序爲人類聲音，另一狀態揚出聲音順序並 音。 如VAD單元指出某一聲音順序爲人類聲音，則糊將發 出第-狀態信號至語言編碼單元，其將聲音順序編在語言 框中。如某一聲音順序代表之聲音並非人類聲音，vad單 兀將發出第二個信號至SID (靜音敘述器）單元。該sid單元 將在每一 N:th框中送出一SID框。在其餘框中，可能之 機會發出框時，將不送出任何框β 一811)框中包含有關估 計之背景噪音及估計之送出侧之噪音頻譜。此方式可以節 省電池電力及無線電帶寬。 當SID單元自產生第一狀態信號改變爲產生第二狀態信號 "4- 各紙張尺度適用中國囤家標準（CNS ) Α4規格（210 X 297公釐 {請先閲請背面之泛意事項再填窝本頁jA7 B7 Printed by the Central Consumer Bureau of the Ministry of Economic Affairs, Consumer Cooperatives 407435 V. Description of the invention (1) The technical model of the present invention The present invention is about the communication week, so gn > miPA〇f Question 1 reveals the way and about the purpose_ Language processing for voice communication on the road. DESCRIPTION OF RELATED ART A typical international Internet telephone use,-soundboard,-microphone and two speakers / microphone and speaker are usually placed next to each other on the desk. This -configuration will cause considerable crosstalk on the receiver side as echo. This echo must be suppressed to make the international network available. In GSM, everyone knows that #mobile phone users use VAD (Voice Activity Detection) detection when they are talking or not. This information is used to reduce «when transmitting sound. In discontinuous speech coding, according to the vox principle (voice operation transmission), the -VAD unit is responsible for testing whether the sequence of sounds received represents human language. This VAD unit can have two states, the first order is human voice, and the other state is the order of sound. If the VAD unit indicates that a certain sound sequence is a human voice, the paste will send a first state signal to the language coding unit, which encodes the sound sequence in the language box. If the sound represented by a sound sequence is not human, the vad unit will send a second signal to the SID (silent narrator) unit. The sid unit will send a SID box in each N: th box. In the remaining boxes, when a chance is given, no box will be sent (β-811). The box contains the estimated background noise and the estimated noise spectrum of the sending side. This method can save battery power and radio bandwidth. When the SID unit changes from generating the first status signal to generating the second status signal " 4- Each paper size is subject to China Store Standard (CNS) Α4 specification (210 X 297 mm {Please read the general matters on the back before reading Filling this page j
經濟部中央標隼局員工消費合作社印掣 4J07435 -----________ B7 五、發明説明（2 ) 時’即自偵出語言改變至偵出非語言？因此，一時間期間 ’所明剩餘部分即加在其上，在此期間，語言编碼單元將 繼續發出語言框，就像收到之聲音順序爲人類語言-樣。 如果，在剩餘部分過後，VAD單位仍偵出非語言時，則產 生SID框。此—程序之理由爲人類語言中字與字之暫停不 應解釋爲非語言’而語言框產生器仍然爲活動的。 本發明之略述 本發明揭示—方法及一裝置以降低由串音引起之回音。 本發明之目的爲降低由串音引起之回音。 上述I問題，即如何降低由_音而引起之回音得以由狀 態機器所控制之麥克風及揚聲器.引進之開關而解決，該狀 態機器自麥克風接納信號之信號能作爲輸入，自麥克風之 信號之一 VAD旗，信號之信號能送至揚聲器及信號之VAD 旗送至揚聲器。 本發明之一優點爲由_音引起之回音得以大部分降低而 不需太多之計算功率。 以下之詳細說明將對一個精於此技藝人士對於其他優點 將更爲瞭解。 以下之詳細説明亦可更明顯了解本發明可以應用於更廣 泛之範圍。但應了解本發明之較佳具體實例僅以説明方式 Φε出，因此此技藝人士將可自詳細之敘述中作不同之改變 及修正仍可在本發明範圍内進行。 圖説之簡略説明 圖1爲本發明一具體實例之方塊圖。 -5- 本紙張尺度適用中國國家榡準{ CNS ) Α4規格（210Χ：297公釐) "" (請先聞讀背面之注意事項再填寫本頁) 訂 經濟部中央標準局貝X消費合作社印製 407435 A7 _____B7 五、發明説明（3 ) 圖2為一有限狀態圖解。 較佳具體實例之詳細說明 圖1中，一麥克風101連接至一 GSM編碼器102。信號到達 GSM編碼器1〇2之前，已依照已知科技予以數位化及個以抽 樣，但未7F於圖1中。自GSM編碼器丨02，編碼後信號被首 先通.過一可使能發射及不發射之開關1〇3發射至一接收機 (未在圖中示出）。自GSM編碼器1 〇2，一 acFe (自動相關 係數）傳送至VAD單元104。一長期預測遲後值&自GSM 框發射至VAD單元105。自VAD單元1〇4 , —代表信號之 能之值PE送至有限狀態機器1〇5。VAD單元1〇4亦計算旗 FE的指示VAD單元104是否偵出人類語言β旗匕傳送至 有限狀態機器105。旗FE在偵出人類聲音時為真。 圖1尚有一接收自發送者（未示出）之編碼聲音信號，並被 送至GSM解碼器106。自GSM解碼器1〇6，此解碼、抽樣 之聲音信號被首先送至一可使聲音信號能及不能到達揚聲 器107之開關108»為使揚聲器工作正常，根據已知科技， 需要一個D/A轉換，但未示出圖中。自收到之抽樣，編碼 聲音信號中，演繹出長期預測器滯後值ND並送至一VAD單 元 109。 由於GSM框之解碼不必使用VAD單元，GSM解碼器 缺少必要之參數以計算ACF »為了能計算ACF，一自動相關 單元110接收自GSM解碼器106之資料及計算ACFD，ACFD並 被送至GSM編碼器。自動相關器單元丨丨〇為GSM編碼器 之一部分如在標準中所述者。值Pd指示，至揚聲器之聲音 信號之能源自VAD單元1 〇9傳送到有限狀態機器1〇5。自 (请先閲讀背面之注意事項再填寫本頁) 裝 I訂 線 本紙張尺度適用中國國家標準（CNS ) A4規格（21〇χ297公爱） 4074J>d a? ______ B7 五、發明説明（4 ) VAD單元109，旗FD被送至該有限狀態機器，指出VAD單 元已偵出人類聲音。 有限狀態機器106包括設定開關103及109之功能，而與 輸入至有限狀態機器之値有關。 圖2爲圖1中之有限狀態機器之狀態及可能之轉換。 在狀態間之變換依照下列定義而爲之： •FE :編碼時VAD旗 :解碼時VAD旗 •PE :編碼時信號能源 •PD :解碼時信號能源 •剩餘時間：自決定至交換方向之時間，直到交換亦作定 。此一時間必須夠長以補償室内回音。 201. FE=1，及FD=0或FE=1及PE>PD，剩餘時間=〇 202. FE=0 ’ 剩餘時間=600 ms 203. FD=0’ FE=0或 FD=1 及 PD>PE，剩餘時間=〇 204. FD=0 ’ 剩餘時間=600 ms 205_FD=1，PD>PE，剩餘時間=6〇〇 206*fb=1 ’ PE>PD，剩餘時間=600 ms 在狀態發射207 ’爲自致動麥克風開關控制聲信號之發射 ，及開關控制聲音信號之發射至中止之揚聲器。在狀態接 收208中，爲自中止之麥克風開關控制聲音信號之發射，及 開關控制發射至致動揚聲器。在IDLE狀態2〇9，開關均被 禁止。 本發明已如上所述，非常明顯其可以用許多方式加以修 "7 ~ 本紙張尺度顧f麵家標準（CNS ) A4%^_ ( 210X297公楚） A7 B7 經濟部中央標準局員工消費合作社印掣 407^35 五、發明説明（5 ) 改。而該種修改不能認爲是悖離本發明之精神與範圍，所 有對此技藝人士非常明顯之修改均將包括在以下申請專利 範園中。 -8- 本紙張尺度適用中國國家標準{ CNS ) Α4規格（210Χ297公釐） (請先閣讀背面之'注意事項再填寫本頁)4J07435 -----________ B7 of the Central Bureau of Standards of the Ministry of Economic Affairs of the Consumers' Cooperatives B7 V. Description of the Invention (2) When ‘is the detection of non-verbal language? Therefore, the remainder of the time period is added to it, during which the language coding unit will continue to send out speech boxes, as if the sequence of sounds received was human language. If, after the remainder, the VAD unit still detects non-verbal, a SID box is generated. The reason for this procedure is that the pause of words and words in human language should not be interpreted as non-verbal 'and the speech box generator is still active. SUMMARY OF THE INVENTION The present invention discloses a method and a device for reducing echoes caused by crosstalk. The object of the present invention is to reduce the echo caused by crosstalk. The above-mentioned I problem, namely how to reduce the echo caused by the sound can be controlled by the state machine's microphone and speaker. The introduction of the switch solves the state machine's signal received from the microphone can be used as an input, and one of the signals from the microphone VAD flag, the signal of the signal can be sent to the speaker and the VAD flag of the signal can be sent to the speaker. An advantage of the present invention is that the echo caused by the tone can be largely reduced without requiring much computing power. The following detailed description will give a person skilled in the art more understanding of other advantages. The following detailed description also makes it clear that the present invention can be applied to a wider range. However, it should be understood that the preferred specific examples of the present invention are only described in an illustrative manner Φε, so those skilled in the art will be able to make different changes and modifications from the detailed description while still being within the scope of the present invention. Brief Description of the Drawings Figure 1 is a block diagram of a specific example of the present invention. -5- This paper size applies to the Chinese National Standard {CNS) A4 specification (210 ×: 297 mm) " " (Please read the precautions on the back before filling this page) Order the Central Bureau of Standards, Ministry of Economic Affairs, Shell X Consumption Printed by the cooperative 407435 A7 _____B7 V. Description of the invention (3) Figure 2 is a finite state diagram. Detailed description of a preferred specific example In FIG. 1, a microphone 101 is connected to a GSM encoder 102. Before the signal reaches the GSM encoder 102, it has been digitized and sampled according to known technology, but it is not shown in Figure 1 in 7F. From the GSM encoder 02, the signal is first passed after encoding. After a switch enabling transmission and non-transmission, the switch 103 is transmitted to a receiver (not shown in the figure). From the GSM encoder 102, an acFe (Auto Correlation Coefficient) is transmitted to the VAD unit 104. A long-term prediction delay value & is transmitted from the GSM frame to the VAD unit 105. From the VAD unit 104, the value PE representing the energy of the signal is sent to the finite state machine 105. The VAD unit 104 also calculates the flag FE to indicate whether the VAD unit 104 detects the human language β flag and transmits it to the finite state machine 105. The flag FE is true when human voices are detected. FIG. 1 also has an encoded sound signal received from a sender (not shown) and sent to the GSM decoder 106. From the GSM decoder 106, the decoded and sampled sound signal is first sent to a switch 108 that enables the sound signal to reach the speaker 107. To make the speaker work properly, a D / A is required according to known technology Conversion, but not shown in the figure. From the received samples and encoded audio signals, the long-term predictor hysteresis value ND is deduced and sent to a VAD unit 109. Since the VAD unit is not necessary for decoding the GSM frame, the GSM decoder lacks the necessary parameters to calculate the ACF Device. The autocorrelator unit is part of the GSM encoder as described in the standard. The value Pd indicates that the energy of the sound signal to the speaker is transmitted from the VAD unit 109 to the finite state machine 105. Since (please read the precautions on the back before filling this page) Binding line The paper size is applicable to Chinese National Standard (CNS) A4 specification (21〇χ297 public love) 4074J > da? ______ B7 V. Description of the invention (4) The VAD unit 109 and the flag FD are sent to the finite state machine, indicating that the VAD unit has detected human voice. The finite state machine 106 includes functions of the setting switches 103 and 109, and is related to the input to the finite state machine. Figure 2 shows the state and possible transitions of the finite state machine in Figure 1. The transition between states is based on the following definitions: • FE: VAD flag during encoding: VAD flag during decoding • PE: signal energy during encoding • PD: signal energy during decoding • Remaining time: the time from the decision to the exchange direction, Until the exchange is decided. This time must be long enough to compensate for room echoes. 201. FE = 1, and FD = 0 or FE = 1 and PE &PD; PD, remaining time = 〇202. FE = 0 'remaining time = 600 ms 203. FD = 0' FE = 0 or FD = 1 and PD > PE, remaining time = 〇204. FD = 0 'remaining time = 600 ms 205_FD = 1, PD > PE, remaining time = 6〇206 * fb = 1' PE > PD, remaining time = 600 ms in status transmission 207 'It is a self-actuated microphone switch to control the emission of acoustic signals, and a switch to control the emission of acoustic signals to the suspended speakers. In the status receiving 208, the transmission of the sound signal is controlled by the suspended microphone switch, and the switch control is transmitted to the actuating speaker. In IDLE state 209, the switches are disabled. The present invention has been described above, and it is very obvious that it can be repaired in many ways. "7 ~ This paper size is standard (CNS) A4% ^ _ (210X297 Gongchu) A7 B7 Employee Consumer Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs Seal 407 ^ 35 5. Description of the invention (5) Change. Such a modification cannot be considered as deviating from the spirit and scope of the present invention, and all modifications obvious to those skilled in the art will be included in the following patent application parks. -8- This paper size applies to the Chinese National Standard {CNS) Α4 size (210 × 297 mm) (please read the “Precautions on the back side before filling out this page)