200307874 玖、發明說明： 技術領域 本發明一般係關於音頻識別。較具體言之，本發明係關 了攜式裝置配置用來識別一音頻磁軌。 先前技術 人類耳朵聽得到的聲音，即具有頻率每秒20至20,000振動 (20-20,〇〇〇 Hz)，稱之為音頻。音頻的例子包括音頻、音樂或 其他。另外，音頻一般聽自三個音源之一，即是現場演奏、 1己錄或廣播。一般地，記錄及廣播為類比或數位。類比記 錄包括磁帶記錄友唱片，而數位記錄包括小型磁碟（CD)、迷 你磁碟、各種檔案格式（如MpEG音頻層3 (Mp3)檔案）或其他。 類比廣播包括聲音複製，如經由立體音響及類比無線電廣 播。另一方面，數位廣播包括數位無線電廣播，如由幻^衛 星無線電及SIRIUS衛星無線電，及經網際網路流廣播，如實 際音頻、WINDOWS媒體或MP3流。 通系’音頻的聽眾需要識別收聽中的音頻磁執。音頻磁 軌為任何有限長度的音頻成分如歌曲、談話或其他。音頻 磁軌的識別甚-為重要其理由如，讓使用者能識別一條歌曲 以便購買該歌曲；了解作者；知道作者的詳細；能在將來 識別該青頻；確認版稅付予何人；標示一未識別音頻磁軌 表；或其他。 一般，由一些方法確立音頻磁軌的身份，如由聽者確認 該晉頻磁執，讀取相關寫入識別該音頻磁軌，或相信該音 頻磁軌的身份聲明。例如，一位聽者確認他所知的歌曲或 200307874 作者、，讀取音樂㈣叫以決定歌曲身份，或收聽無線播 青員罝佈一歌曲名稱及作者。 不過，在某些狀況中使用者不能使用上述方法識別—音 頻磁軌。確實，各音頻源具有其音頻識別的缺點。例如， 廣播的缺點為無線電廣播員不報告音頻磁軌的身份；他們 等候太久而不能報告，而聽者不能等到歌曲播完才聽報告； 而一般也不方便寫下歌曲的名稱等。記錄缺點的例子為， 向來纪錄不能通知聽者該音頻磁軌的身份。 不過’近年來數位記錄已在記錄音頻資料内介紹音頻識 別資料。這種，音頻資料另外稱為媒介資料並與許多數位音 頻樓案形式有關。-媒介資料的例子為關係Mp3音頻樓案的 ID3標籤。廷種媒介資料一般包括音頻檔案有關的基本資訊 如歌曲名稱、作者、磁軌長度等。同樣地，數位流廣播有 時候也連結媒介資料至其數位音頻流。 用於識別數位記錄的另外方法由GRACEN〇TE(以前為cddb of Berkdey，加州）提供及揭露於美國專利案第6,33〇,593、 6,240,459、6,23G，2G7、6,23G，192、6，161，132、6，154,773、6，㈤及 5,987,525號。GRACENOTE使用一小型磁碟資料庫（CDDB)以識 別由預先記錄CD產生的音樂。CDDB使用從CD目錄表，如磁 轨CD表及相關的磁軌時間，找到的專用識別符號以識別 上的歌曲。CDDB服務結合各種電腦軟體媒體播放機以識別 晋頻磁軌。這些媒體播放機使用cDDB移入檔案名稱&CD編 碼各歌曲的媒介資料。 CDDB技術的另外應用容許獨立CD播放機（未連接電腦或 200307874 網際網路）顯示歌曲名稱及作者資訊。為此，裝置必須局部 儲存GRACENOTE·資料庫及局部在裝置上執行上述相同的技 術。 CDDB技術的一項缺點為需要呈現全部預先記錄(：1)以便能 識別CD的個別晋頻磁軌。所以，本技術聽者不能用來識別 從記錄CD以外的音頻源所聽到的各別音頻磁軌。 另外形式的識別晉頻裝置使用一時間打印技術以識別音 頻磁軌。兩種已知裝置應用這種時間打印技術為s〇ny e_ MARKER及XENOTEI-TAG。這些裝置為非常簡單的鍵鏈裝置， 如果裝置上的按鈕被按下便簡單記錄資料及時間。使用中， 如果一位聽者從無線電聽到一歌曲想要識別，便按下裝置 上的按鈕及裝置便記錄按鈕按下的日期及時間。然後，如 果裝置與一桌上電腦同步，有關聽者裝置的使用者專用識 別符號及記錄的日期及時間資訊便經網際網路傳送至一伺 服=。一般而言，顯示一網頁以顯示聽者先前已識別作為 是最常聽的各無線電台的歌曲播放機。該裝置本身不儲存 任何使用者在選擇時收聽電台的有關資訊。呈現識別歌曲 的網頁，it常也提供有關購買含有選擇歌曲等的cd的選擇。 使用時間打印技術裝置的-項缺點為而不能完全自動處 里4別歌曲貝訊因為使用者需要記憶何時起動裝置及收聽 那個包口。另外，使用者必須操作電腦以獲得音頻磁軌識 別。特別而言，使用者必須識別他最常聽的無線電台。另 外相互作用需要經網際網路，結果，正常缺 網路的潛伏期，了 土、 、 ^枯，、，罔際 可罪度及速度。明顯不同，其相互作用一 200307874 般要比使用非網際網路為準的的音頻磁軌識別方 多。 又 另外，因為茲種裝置只記錄起動時間及日期，該種裝置 限於無線電廣播使用。另外，該種裝置要求服務供應商維 持一資料庫包含該服務供應商願意支援的各市場的電台的 完全播放表及播放時間。因為無線電台不提供這種資訊， 收集播放表及附屬播放時間一般由第三者執行。第三者以 以手動方式識別及將播放表及附屬播放時間輸入資料庫， 或自動識別這些播放表及附屬播放時間及由一電腦存入資 料庫。任何-種情況，識別及儲存都是複雜，需要重大努 力、成本高及一般限於最受歡迎的電台，因而無法包括許 多地區及市場。 另外扁别音頻磁軌的先前技術的方法係使用音頻指紋識 別。一般骨頻指紋識別使用軟體以識別一歌曲係藉由比較 一音頻樣本的專用音頻符號或指紋（以下稱「指紋」）與已知 晉頻樣本相關的已知「指紋」的資料庫。 一些服務供應商及/或軟體應用利用數位指紋識別技術以 識別音頻磁軌。例如，CLANG〇，一種由加州AumBLE magic 公司製造的軟體產品，使用數位指紋識別以識別流音頻廣 播’該廣播不提供相關音頻磁軌媒介資料。由AUDIBLE magic 公司執行的指紋揭露於美國專利案第5,918,223號。 另外相似青頻指紋識別技術的供應商為AUDITUDE，其軟 體產品Π)3ΜΑΝ的目標在於具有大量不正確或不完全的識別 貝料數位骨頻檔案的使用者。經由技術組合，包括音頻指 200307874 紋識別，ID3MAN識別音頻檔案及然後改正該檔案有關的識 別資料。 這些指紋識別裝覃或服務的缺點為使用者不能收聽來自 桌上電腦的音樂（CDDB支援的CD播放機除外，因為該裝置 需要儲存超大的GRACENOTE資料庫，及具有如上述的缺 點）。 另外識別音頻的方法使用一行動電話網路，如果收聽該 音頻後使用者想要識別，使用者呼叫一指定號碼以識別音 頻。至少有兩種方法用來提供這種服務。 第一方法，由名為BUZZHITS (現已停業）所提供，容許使用 者呼叫號碼及輸入識別呼叫者的使用者的識別符號（及呼叫 者的地理市場）及提及使用者廣播識別音頻的無線電台的廣 播頻率。一旦供應該廣播頻率，提供使用者樣本音頻夾以 便使用者從中選擇一樣本夾以獲得該音頻磁軌的身份。這 種資訊也用電子郵寄至該使用者。 雖然這種電話服務解決一些上述缺點，仍要求使用者手 動與該裝置相互作用及迫使使用㈣時在聽到音頻的時候 相互作用，該操作通常不方便。 另外，-旦音頻磁軌已經識別，上述音頻識別方法不 2執行額外動作。由於經新增的訂戶服務能下載的音 2，>肖費者希望有選擇以購買及下載聽自各種不同源 二要，用上述產品/服務完成—處理需要多步驟手動處 :要與網際網路及桌上電腦，或行動電話，相互作用。 &據上述，避免上述缺點需要—音頻識別裝置及方法 200307874 不但方便及容易使用，並具有低成本的準確識別。 - 發明内容 、本發明的一具體貫施例包括一可攜式裝置能從傳聲器、 音頻播放機及/或音頻接收器記錄。為了識別一音頻磁軌， 本可攜式裝置如果起動，記錄播放機播放中音頻磁軌或無 線電接收器接收中的一音頻樣本。如果可攜式裝置目前不 播放音頻或接收一無線電廣播，便可記錄經傳聲器的音頻 樣本。這種記錄音頻樣本儲存在該可攜式裝置的内部儲存 器，及以後，如果連接一客戶電腦，便上載至該客戶電腦。 客戶電腦處理·該音頻樣本以產生一音頻樣本的「指紋」然 後比較客戶電腦上或經網際網路耦合客戶電腦的識別伺服 态（ID伺服器）上的指紋資料庫。一旦指紋已經識別，名稱及 作者資訊便回到客戶電腦及最後在可攜式裝置上顯示。 另外，可攜式裝置本身處理記錄音頻樣本的指紋，及產 生該指紋。這具有減少儲存音頻樣本所需的儲存空間量的 優點，因為只有指紋儲存於該可攜式裝置。不過，本具體 貫施例要求該可攜式裝置具有充足處理器功率以執行指紋 識別。 — 退回作者及標題資訊以外，一旦音頻樣本或磁軌已經識 別，該裝置也執行額外動作。額外動作的例子包括下載訂 戶服務的識別音頻磁軌，推薦更多類似識別音頻磁軌的音 頻磁軌，從網際網路音樂商獲得識別音頻磁軌的價格。較 理想地’藉由從播放機顯示器選擇一功能表項目而選擇這 些額外動作，及也可從第三者服務供應商定購及下载。 -11 - 200307874 根據本發明提供一種用於可攜式裝置以識別音頻的方 法。於一可攜式装置上記錄一音頻磁軌的一音頻樣本。然 後將曰頻樣本儲存於該可攜式裝置的一快取記憶體内。傳 輸Μ頻樣本至_電腦裝置加以識別。由該電腦裝置接收 β曰頻樣本及在該音頻樣本上執行指紋識別以便獲得該音 f “本的一專用音頻指紋。然後搜尋指紋資料庫用於找出 與该指紋相配的一已知先前識別的音頻磁軌的一已知指 、、又找到相配及傳送先前識別的音頻磁軌的相關識別資料 土 4可祆式裝置。然後接收音頻樣本的識別及在該可攜式 裝置上顯示。_在另外具體實施例中，指紋識別在可攜式裝 置上執行。 在另外具體實施例中於可攜式裝置接收及播放—無線電 f播。接收—指令以識別—騎電廣播的音頻磁軌及無線 ^廣播的廣播頻率，及自動記錄的可攜式裝置接收識別該 骨頻磁軌的該指令的日期及時間。然後傳輸廣播頻率，曰 /月及時間及-專用裝置識別符號至—電腦裝置加以識別。 本㈣.由該電腦裝置接收。然後搜尋播放表資料庫以找出 入/廣播/頁率、日期及時間相配的一已知電台根據使用者 的地理位置由專用裝置識別符號決定的廣播頻率，及由該 電台廣播-音頻磁軌的時知曰期及時間。找出該音頻磁軌 ，關的廣播頻率、日期及時間及傳送至該可攜式裝置。該 可搞式裝置以後接收及顯示該識別音頻磁軌相關的資訊。 在另外具tf實施例中，指紋識別在可攜式裝置上執行。 根據本發明也提供一可攜式裝置、電腦裝置及識別飼服 -12- 200307874 器用於執行上述的方法。 所以、、、口 口本功此及一可攜式音樂播放機，識別方法便可 自動。另夕卜’本裝置的使用範圍便可擴大以涵蓋—較廣的 音樂“mm因而以小額外成本便可獲得額外功能。 同時’本裝置可協助編輯無線電台播放表的資料庫。該裝 置可（動4、外動作，而该額外動可根據使用者的喜愛加以 個人化處理。 實施方式 、圖1為根據本發明一具體實施例的一識別音頻系統謂的 方鬼圖系、、、充400包括至少一識別伺服器1〇2(以下稱「仍伺 服器」）及至少一客戶電腦1〇6經網路1〇4耦合另一電腦。id 伺服為102及客戶電腦1〇6為任何形式的電腦裝置。不過，在 具把實她例中客戶電腦1〇6為一桌上電腦及網路綱為網際 網路。 客戶電腦106耦合網路104經由任何適合通信鏈1〇8，如乙 太網路、共軸電纜、銅電話、線、光纖、無線、紅外線或 其他。-可攜4骨頻識別裝置U2(以下冑「可攜式裝置」） 耦合客戶電腦106。可攜式裝置112較理想為手掌型。可攜式 裝置112耦泛客戶電腦1〇6經由任何適合通信鏈l如通用 串聯匯流排(USB)、Firewire、乙太網路、共軸電纜、銅電話 線、光纖、無線、紅外線或其他。 在另外具體實施例中，客戶電腦1〇6為固定無線基地台隸 合一閉極/數據機連接網路辦。例如，客戶電腦為一侧（無 線真確度-IEEE 802.11b無線網路）基地台經由一數位訂戶線 -13- 200307874 (DSL)閘極（未#員示）耦合網路i〇4。在本具體實施例中，從可 攜式裝置112至客戶電腦1〇6的通信鏈為一斯朽無線通信鏈。 在另外具體貫施例中，並沒有客戶電腦1〇6及可攜式裝置 112直接通信ID伺服器1〇2。例如，可攜式裝置112包括行動 私老通佗電路系統經一行動電話網路（網路1〇4)通信伺服器 102。 在一另外具體實施例中，可攜式裝置112需要獨立識別音 頻。例如，可攜式裝置定期從網路1〇4下載更新播放表及/或 指紋識別資料庫，如以下詳細說明及參考圖6及9。 在一具體實施例中，一播放表供應商114及指紋供應商116 也耦合至網路104。播放表供應商114為一伺服器供應更新的 播放表至ID伺服器1〇2、客戶電腦1〇6，及/或可攜式裝置112， 而指紋供應商116為一伺服器供應新音頻磁軌的更新的指紋 資料至ID伺服器1〇2、客戶電腦1〇6及/或可攜式裝置112。 圖2為Π)伺服器102及/或圖示的客戶電腦ι〇6的方塊 圖。ID伺服器1〇2及/或客戶電腦1〇6顯示於一示意圖以避免 重複。不過’必須了解本發明的所有具體實施例中並不需 要出現下列的—ID伺服器1〇2及/或客戶電腦1〇6的所有元件而 只包括作為例子的元件。 較理想地，ω祠服器102及/或客戶電腦1〇6包括：至少一 資料處理器或中央處理單元（CPU)202; 一記憶體21〇;使用者 介面裝置206;如-監視器及鍵盤；通信電路系統綱用於通 信網路104(圖1)、ID伺服ϋ1〇·υ、客戶電腦(圖〇及/ 或可攜式裝置m(圖υ及至少一連接這些组件的匯流排 -14- 200307874 208。 記憶體210較理想包括操作系統212，如VXWORKS、LINUX 或WINDOWS具有指令用於處理、存取、儲存或搜尋資料等。 較理想記憶體210也包括通信程序214用於通信網路104(圖1) ，ID伺服器102(圖1)，客戶電腦106(圖1)及/或可攜式裝置 112(圖1);指紋識別程序216 ;搜尋程序218 ;指紋識別資料 庫220 ;無線電播放表資料庫224 ;地理識別符號234 ; —「無 識別」信息236 ;及一快取記憶體238用於暫時儲存資料。 指紋識別程序216係用來獲得一音頻磁軌的音頻樣本上的 一專用識別符i虎或指紋，詳細說明如以下圖4及5所示。指 紋識別程序216包括指令用於執行音頻樣本的指紋識別以獲 得一音頻樣本的專用音頻指紋。 搜尋程序218係用來搜尋指紋資料庫220以進行識別音頻， 詳細說明如以下圖4至6所示。指紋識別資料庫220包括許多 已知音頻樣本或音頻磁軌的指紋及其相關識別資料222(1)-(N) ，如歌曲名稱、作者或其他。 在一.另外具體實施例中，具有一無線電播放表資料庫 224。在本具髏實施例中，無線電播放表資料庫224包括許多 無線電頻率226(1)-(N)及各頻率226(1)-(N)相關的播放表228⑴-(N)。各播放表 228(1)-(N)包括一日期 230(1)·(Ν)及時間 232(1)-(N)， 及各音頻磁軌廣播在該日期及時間的身份232(1)-(N)。例如， 無緣電台KJAZ具有頻率98.7 FM，及一播放表包括Frank Sinatra 的「紐約，紐約」在2002/01/21，9:00 AM廣播。 由於散佈在世界上的許多無線電台共享相同頻率，提供 200307874 一地理識別符號234以識別該無線電台或在特定地區内的頻 率226(1)-(N)。本埤理識別符號234由任何適當的方法提供。 在一具體實施例中使用者提供該地理識別符號。在另外具 體實施例，該地理識別符號234係從使用者的專用網路位址 獲得。例如，客戶電腦觸及/或可攜式裝置ιΐ2的一網際網 路k走（IP)位址可用來概算使用者的地理區。在另外具體實 施例中，結合客戶電腦蘭及/或可攜式裝置112的全球定位 系、、·充（GPS)可用來决足使用者的地理區。如果仍飼服器⑽及/ 或客戶電腦106不能識別一音頻磁軌，使用「無識別」信息236 通知使用者沒有識別可作。或者，在接收無識別信息之前， 使用者可獲得一些「最近相配」的可能識別。 熟二本技術者必須了解這些裝置的某些元件仍伺服器⑺2 及/或客戶私知1〇6並不需要。例如，指紋識別程序us、搜 尋程序218及指紋資料庫224只有發生音頻磁軌指紋識別的裝 置需要。換言之，如果Π)伺服器102發生指紋識別，則上述 記憶體210元件只需在仍伺服器1〇2出現。同樣，在該具體實 施例中如果識別發生在可攜式裝置112(圖丨）上，則仍伺服器 102或客戶電腦· 1〇6都不需要上述記憶體21〇的元件。 圖3為圖1所不可攜式裝置112的方塊圖。熟悉本技術者必 須· 了解本發明的所有具體實施例中不需要出現下列可攜式 裝置112的所有元件而只包括作為例子的元件。 車乂理想，可攜式裝置112包括：至少一資料處理器或中央 處理單7C (CPU)302 ; —記憶體310 ;使用者介面裝置3〇8 ;如 按紐、一螢幕及一耳機；通信電路系統304用於通信網路 -16- 200307874 104(圖1)、ID伺服器102(圖1)及/或客戶電腦106(圖1); 一或 更多音頻播放機350，如一 CD或MP3播放機；一傳聲器352 ; 一無線電接收器354及天線356用於接收無線電廣播；及至少 一匯流排306連接這些組件。 記憶體310較理想包括操作系統312，如VXWORKS、LINUX 或WINDOWS具有指令用於處理、存取、儲存或搜尋資料等， 記憶體310較理想也包括通信程序314用於通信網路104(圖 1)、ID伺服器102(圖1)及/或客戶電腦106(圖1);指紋識別程 序316 ;搜尋程序318 ;指紋識別資料庫320 ;無線電播放表 資料庫324 ;地理識別符號334 ;地理識別程序336 ; —「無 識別」信息338 ;記錄程序340 ;播放機程序342 ;無線電程 序344 ; —快取記憶體346用於暫時儲存資料；頻率偵測程 序358 ;及一時脈360。 指紋識別程序316係用來獲得一音頻磁軌的音頻樣本上的 一專用識別符號或指紋，詳細說明如以下圖6所示。 同樣，在本具體實施例中搜尋程序318係用來搜尋指紋識 別資料庫320以進行識別音頻，詳細如以下所述。指紋識別 資料庫320包括許多已知音頻樣本或音頻磁軌的指紋及其相 關識別資料322(1)-(N)，如歌曲名稱、作者或其他。 ’在一另外具體實施例中，具有一無線電播放表資料庫 324。在本具體實施例中，無線電播放表資料庫324包括許多 無線電頻率326(1)-(N)及各頻率328(1)-(N)相關的播放表326(1)-(N)。各播放表 328(1)-(N)包括一日期 330(1)-(N)及時間 332(1)_(N)， 及各音頻磁軌廣播在該日期及時間的身份332(1)-(N)。 200307874 同樣用於上述另外具體實施例，提供一地理識別符號334 以協助識別無線-電台或一特別地區的頻率326(1)-⑼。例如， 根據識別地理區從一組儲存在該裝置的頻率中選擇該地理 識別符號334。本地理識別符號334由任何適當的方法提供。 在一具體實施例中使用者提供該地理識別符號334。在另外 具體貫施例中地理識別符號334係由地理識別程序336獲得。 如上述’這可從使用者專用網路位址決定。例如，可攜式 裝置112的一網際網路協定（ip)位址可用來概算使用者的地理 區。在另外具體實施例中，結合可攜式裝置112的全球定位 系統（GPS)可用·來決定使用者的地理區。 在所有具體實施例中，「無識別」信息236係用來通知使 用者無識別可作，如果可攜式裝置112不能識別音頻磁軌。 在本發明的具體實施例中，如果音頻磁軌的指紋識別係 用來識別音頻，記錄程序34〇記錄一儲存在快取記憶體348的 音頻樣本348。從音頻播放機350、傳聲器352及/或無線電接 收器354記錄該音頻樣本。 在本發明的具體實施例中，如果日期，時間，及無線電 台或廣播頻率、係用來識別來自無線電台播放表的音頻， 圮錄程序係用來記錄日期、時間及記錄一儲存在快取記憶 體348的廣播或無線電台349頻率。 較理想提供播放機程序342以播放音頻在音頻播放機35q。 特別需要播放機程序342用於播放數位音頻，wMP3音頻磁 軌或其他。 較理想提供無線電程序344以播放由天線356接收經無線電 -18- 200307874 接收器354的無線電。不過，必須了解本發明的所有具體實 施例中並不需要-出現上述記憶體31〇的所有元件而只包括作 為例子的元件。 頻率程序358係用來偵測無線電台廣播的頻率，及時脈36〇 係用來保持日期及時間。頻率程序358及時脈36〇的詳細說明 參考圖7至9。 圖4為根據本發明具體實施例由一 ID伺服器ι〇2執行識別的 一種識別音頻方法的流程圖。在一具體實施例中，音頻播 放機350(圖3)及/或可攜式裝置112的播放機程序342(圖3)在步 驟402經使用者·介面裝置3〇8(圖3)播放音頻。例如，一内裝]vq>3 播放機播放音頻經一耳機至使用者。在另外具體實施例中， 無線電接收器345(圖3)及/或無線電程序344(圖3)接收及播放 音頻廣播經可攜式裝置的耳機。 然後在步騾404接收指令以識別音頻。較理想，這些指令 來自使用者，如使用者按下可攜式裝置上的一「現在識別」 鈕，或其他。在一另外具體實施例中，自動接收一記錄指 令。例如，每2分鐘自動記錄一音頻樣本。播放及接收指令 的步驟402及404非本發明所必要及在某些具體實施例中並不 需要發生。 在步騾406由記錄程序342(圖3)記錄一音頻樣本及儲存一 音頻樣本348(圖3)於快取記憶體346(圖3)内。在一具體實施 例中，連續記錄音頻及自動分割成音頻樣本具有充分長度 以進行指紋識別。例如，連續記錄音頻及自動分割成30秒 音頻樣本其連續傳送至ID伺服器102作識別。 -19- 200307874 如果音頻播放機350(圖3)及/或可攜式裝置112的播放機程 序342(圖3)不播硃音頻，便在步騾4〇6經傳聲器352(圖3)記錄 音頻。 在具體實施例中，如果可攜式裝置112經客戶電腦1〇6耦合 網路104，在步騾408通信程序314(圖3)傳輸該音頻樣本至客 戶電腦106。如前述，本通信經通信鏈11〇發生，如一串聯埠 連接無線連接，或其他。在步騾41〇由客戶電腦接收的音頻 樣本及傳送至ID伺服器1〇2。 必須了解可攜式裝置不具有一永久的客戶電腦1〇6及/或仍 伺服器102的通信鏈，所以音頻樣本儲存在快取記憶體346(圖 3)直到可攜式裝置112及客戶電腦1〇6及/或m伺服器1〇2之間 建立連接。 在另外具體實施例中，並沒有客戶電腦，所以可攜式裝 置112經一行動電話網路通信ID伺服器1〇2，在步驟4〇8音頻 樣本直接傳送至ID伺服器1〇2。 在步驟412 ID伺服器102接收該音頻樣本。然後，步驟414 指紋識別程序216(圖2)在10伺服器1〇2上執行音頻樣本的指紋 識別以根據該-音頻樣本的特性或聲學特徵決定樣本的一專 用識別符號或指紋。該特性或聲學特徵包括音頻樣本的類 比波土曰里、節距、允度、頻寬、Mel頻率倒譜係數（MFCC) 或其他 種適合的指紋識別骨頻的方法揭露於美國專利 案第5,918,223號，該文以提及方式併入本文。 然後在步·驟416搜尋指紋資料庫22〇(圖2)以找出與記錄在 可攜式裝置的該音頻樣本的指紋相配或部份相配的已知先 200307874 前識別的音頻磁執的指紋。如果未找到相配（418-No)，則在 方驟430傳送一「無識別」信息236(圖2)至可攜式裝置112。 在本具體實施例中，可攜式裝置112經客戶電腦耦合至Π)伺 服器102 ’客戶電腦接收該「無識別」信息及在步騾432傳送 該信息至該可攜式裝置。 該可攜式裝置在步驟434接收該「無識別」信息，及在步 驟436顯示予使用者通知使用者該音頻不能識別。較理想該 信息顯示在該可攜式裝置的螢幕上。 如果找到該指紋與已知先前識別的音頻磁軌相配（418_Yes) ’則在步驟42Q傳送一先前識別的音頻磁軌相關的識別資料 至可攜式裝置112。在本具體實施例中，可攜式裝置n2經客 戶電腦耦合至ID伺服器1〇2，客戶電腦接收該音頻樣本的識 別及在步驟422傳送該信息至該可攜式裝置。 在步驟424及426該可攜式裝置分別接收及顯示該音頻樣本 的識別。例如，顯示一作者及歌曲名稱。在一較佳具體實 施例中，步驟428執行额外動作。這些額外動作只在已知音 頻樣本的身份後執行及例如包括客戶電腦1〇6及/或可攜式裝 置112自動顯示該識別的作者的網頁、傳記、磁片類別；自 動顯示一網頁銷售作者的歌曲或磁片集；從訂戶服務下載 認音頻磁執；推薦一相似音頻；從網路音樂商獲得該識別 磁軌的價才各，或其他。較理想地，藉由從可攜式裝置的顯 示器選擇-功能表項目而選擇這些額外動作，及也可從第 三者服務供應商訂購及下載。可攜式裝置也可恢復有關的 廣告的資訊或談㈣錄。另料錢外料在所有音頻樣 -21 - 200307874 本上執行不論音頻身份為已知或未知。例如使用者下载 數位音頻樓案及—具有全部或部份歌曲的身份，仍須傳送歌 曲用於識別以接收音頻磁軌的額外資訊或在該音頻磁軌執 行某些動作。 圖5為根據本發明另外具體實施例由客戶電腦ι〇6執行識 別的另外識別音頻方法的流程圖。在一具體實施例中，音 頻播放機350(圖3)及/或可攜式裝置112的播放機程序3幻（圖^ 在步騾502經使用者介面裝置308(圖3)播放音頻。例如，一 内裝MP3播放機播放音頻經一耳機至使用者。 然後在步驟504接收指令以識別音頻。較理想地，這些指 令來自使用者，如使用者按下可攜式裝置上的一「現在識 別」鈕。在一另外具體實施例中，自動接收一記錄指令。 例如，每2分鐘自動記錄一音頻樣本。必須說明播放音頻及 接收指令以識別音頻的步驟5〇2及5〇4非本發明所必要及在某 些具體實施例中並不需要發生。 在步驟506由記錄程序342(圖3)記錄一音頻樣本及在快取 記憶體349(圖3)内儲存一音頻樣本348(圖3)。在一具體實施 例中，連續記錄音頻及自動分割成音頻樣本具有充分長度 以進行指紋識別。例如，連續記錄音頻及自動分割成3〇秒 音頻樣本及由客戶電腦106連續識別。 如果音頻播放機350(圖3)及/或可攜式裝置112的播放機程 序342(圖3)不播放音頻，便在步騾5〇6經傳聲器352(圖3)記錄 音頻。 在具體實施例中’如果可攜式裝置U2經無線連接客戶電 -22- 200307874 腦1〇6,在步驟5〇8通信程序314(圖3)連續傳輸該音頻樣本至 客戶電m〇6直到無線連接中斷。在步驟510由客戶電腦1〇6 接收的音頻樣本。 必須了解可攜式裝置不具有-永久的客戶電腦106的通信 鏈，所以音頻樣本儲存在快取記憶體直到可攜式裝置112及 客戶電腦106之間建立連接時。 然後，步驟512指紋識別程序216(圖2)在客戶電腦1〇6上執 行骨頻樣本的指紋識別以根據該音頻樣本的特性或聲學特 徵決定樣本的一專用識別符號或指紋，如上述。 在另外具體·實施例中，指紋識別程序316(圖3)在可攜式裝 置112上執行晉頻樣本的指紋識別以根據該音頻樣本的特性 或聲學特徵決定樣本的一專用識別符號或指紋。然後傳送 该指紋至客戶電腦1〇6，該電腦在步驟514搜尋該指紋資料 庫。 然後在方驟514搜尋指紋資料庫220(圖2)以找出相配或部 份相配的記錄在可攜式裝置的該音頻樣本的指紋。如果找 不到相配（516-No)，則在步驟526傳送「無識別」信息236(圖2) 至可攜式裝置—112，在步驟528接收該「無識別」信息及在步 驟530顯示該信息通知使用者該音頻不能識別。較理想該信 息顯示在該可攜式裝置的螢幕上。 如果找到相配（516-Yes)，步驟518傳送該音頻樣本的識別至 可攜式裝置112。步騾520及522可攜式裝置接收及顯示音頻 樣本的識別，較理想在可攜式裝置的螢幕上。例如，顯示 一作者及歌曲名稱。在一較佳具體實施例中，步驟524執行 -23- 200307874 額外動作，如上述。 圖6為根據本發明另外具體實施例由可攜式裝置112執行 識別的另外識別音頻方法的流程圖。在一具體實施例中， 音頻播放機360(圖3)及/或可攜式裝置112的播放機程序 342(圖3)在步騾602經使用者介面裝置3〇8(圖3)播放音頻。例 如，一内裝MP3播放機播放音頻經一耳機至使用者。然後在 步·驟604接收一指令以識別音頻。在另外具體實施例中，自 動接收記錄指令。必須說明播放音頻及接收指令以識別音 頻的步驟602及404非本發明所必要及在某些具體實施例中並 不需要發生。- 在步騾606由記錄程序342(圖3)記錄一音頻樣本及在快取 1己憶體349 (圖3)内儲存一音頻樣本348(圖3)。在一具體實施 例中，連續記錄音頻及自動分割成具有充分長度的音頻樣 本以進行指紋識別。音頻樣本348(圖3)較理想暫時儲存在快 取記憶體346内（圖3)。 如果音頻播放機350(圖3)及/或可攜式裝置112的播放機程 序342(圖3)不播放音頻，便在步驟606經傳聲器352(圖3)記錄 音頻。 _ 步驟608指紋識別程序316(圖3)在可攜式112裝置上執行音 頻樣本的指紋識別以根據該音頻樣本的特性或聲學特徵決 定樣本的一專用識別符號或指紋。該特性或聲學特徵包括 音頻樣本的類比波型、音量、節距、亮度、頻寬、Mel頻率 倒譜係數（MFCC)或其他。一種適合的指紋識別音頻的方法 揭露於美國專利案第6,918,223號，該文以提及方式併入本 -24- 200307874 文0 然後在步驟610搜尋指紋資料庫320(圖3)以找出相配或部 份相配的記錄在可攜式裝置的該音頻樣本的指紋。如果找 不到相配（612-No)，則在步騾614顯示「無識別」信息340(圖3) 通知使用者該音頻不能識別。較理想該信息顯示在該可攜 式裝置的螢幕上。 如果找到相配（612-Yes)，在步驟616顯示該音頻樣本的識 別’較理想在可摇式裝置的勞幕上。例如，顯示一作者及 歌曲名稱。在一較佳具體實施例中，步騾618執行額外動作， 如上述。. 在可攜式裝置上執行指紋識別的一優點為節省可攜式裝 置的記憶體，因為指紋基本上低於音頻樣本。另外，本發 明的另外具體實施例只使用由連接網路的可攜式裝置組成 的同層間網路識別音頻。例如，如果該裝置具有無線網路 能力，及指紋識別在該可攜式裝置上執行，則藉由搜尋其 他網路連接可Μ裝置上的肢產生音頻樣本的識別。〃 在可攜式裝置上執行指紋識別的另外具體實施例具有中 央服務站，如-記錄儲存器以執行指紋識別。本具體實施例 減輕在服務站的負冑，因為各站不產生指紋及需要轉移及 維’持的資料較少。 圖7為根據本發明具體實施例由一識別伺服器㈤伺服 ’㈣行識別的-種識別無線電廣播音頻方法的流程圖。 無線電接收器354(圖3)及/或可攜式裝置m的無線電程序 344 (圖3)在步驟702接收及播放一無線電廣播。然後在步驟7〇4 -25- 200307874 接收一指令以識別該無線電廣播的音頻磁軌。較理想地， 廷些指令來自使用者，如使用者按下可攜式裝置上的一「現 在識別」鈕或其他。在一另外具體實施例中，自動接收一 識別音頻磁軌指令。例如，每2分鐘產生自動嘗試識別音頻 磁軌。 在步驟706記錄程序342 (圖3 )記錄及儲存無線電台的廣播 頻率，及可攜式裝置112接收識別音頻磁軌的指令的日期及 時間。例如，記錄一廣播頻率為957kHz，日期為2洲2年2月Μ 日，及時間113H00。記錄程序340(圖3)從頻率偵測程序358(圖 3)獲得頻率及從時脈36_ 3)獲得日期及時間。在其最簡單 =式中，頻㈣測程序3_ 3)只是察明使用者選擇的無線 包頻率’即是，讀取無線電接收器調整的值。或者，頻率 偵測程序358(圖3)偵測廣播遺率，—般廣播與音頻訊號一 起。這使用-般傳輸實際無線電識別的無線電資料服務 (娜）。咖-般實際無線電識別，這比較可靠因為電台識 別與地理無關。RDS實際傳輸有關無線電台所有人的資訊， =資訊清楚定義搜尋的播放表。頻率_程序说（圖3)偵測 η播頻率，即偵測無線電接收器調整的頻率。必須了解廣 ^的頻率係自動決^，即是使用者不供應無線電頻率至頻 率偵測程序358(圖3)。 時脈360(圖3)的曰期及時間最初該由使用者設定，或可攜 式裝置能自動設定時脈使用已知技 ^ 技術遙控時脈與一可靠時 =I )同步。料遙控時脈同步化的技術揭露於美 國專利4柳，3默娜，職，㈣皆以提及方式併入本 -26- 200307874 文0 記錄日期，時間及頻枭r 乂、早（以下稱時脈/頻率資料）349(圖3)儲 存在快取記憶體346(圖3)。 ^ _ ^ }货騾708，可攜式裝置112傳輸時 脈/頻率資料349(圖3)。 在本具體實施例中，如果可攜式裝置112經客戶電腦職 口 ID伺服《1〇2 ’在步驟7〇8通信程序似⑴）傳輸該時脈/頻 率貝料349(圖3)至客戶電腦1G6。如先前所述，本通信發生200307874 (ii) Description of the invention: TECHNICAL FIELD The present invention relates generally to audio recognition. More specifically, the present invention relates to a portable device configured to identify an audio track. Prior art The sound audible to the human ear, that is, having a frequency of 20 to 20,000 vibrations per second (20-20,000 Hz), is called audio. Examples of audio include audio, music or others. In addition, audio is usually heard from one of three sources, namely live performance, recording or broadcasting. Generally, recording and broadcasting are analog or digital. Analog records include magnetic tape recorders, while digital records include compact disks (CDs), mini disks, various file formats (such as MpEG Audio Layer 3 (Mp3) files), or others. Analog broadcasting includes sound reproduction, such as via stereo and analog radio broadcasting. Digital broadcasts, on the other hand, include digital radio broadcasts, such as those performed by Phantom Radio and SIRIUS satellite radio, and broadcasts over Internet streams, such as real audio, Windows media, or MP3 streams. Listeners of the general audio system need to identify the audio magnetics that are being listened to. An audio track is any finite-length audio component such as a song, conversation, or other. The identification of audio tracks is very important for reasons such as allowing users to identify a song to purchase the song; to know the author; to know the author's details; to identify the young band in the future; to confirm who the royalties are paid to; Identify audio track tables; or other. Generally, the identity of the audio track is established by some methods, such as the listener confirming the audio frequency track, reading the relevant write to identify the audio track, or trusting the identity statement of the audio track. For example, a listener confirms the song he knows or the author of 200307874, reads the music howl to determine the identity of the song, or listens to a radio broadcaster to publish a song name and author. However, in some cases the user cannot use the above method to identify-audio tracks. Indeed, each audio source has the disadvantage of its audio recognition. For example, the disadvantages of broadcasting are that the radio announcer does not report the identity of the audio track; they wait too long to report, and the listener cannot wait until the song is finished to listen to the report; and generally it is not convenient to write the name of the song, etc. An example of a recording disadvantage is that the recording has always failed to inform the listener of the identity of the audio track. However, 'in recent years digital recording has introduced audio identification information in recorded audio material. This type of audio material is otherwise referred to as media material and is related to many digital audio building formats. -An example of media material is the ID3 tag associated with the Mp3 audio building case. These media materials generally include basic information about audio files, such as song title, author, and track length. Similarly, digital broadcasts sometimes link media data to their digital audio streams. Another method for identifying digital records is provided by GRACENOTE (formerly cddb of Berkdey, California) and disclosed in U.S. Patent Nos. 6,33,593, 6,240,459, 6,23G, 2G7, 6,23G, 192, Nos. 6,161,132, 6,154,773, 6, and 5,987,525. GRACENOTE uses a small disk database (CDDB) to identify music produced from pre-recorded CDs. The CDDB uses a special identification symbol found from the CD directory table, such as the track CD table and related track times, to identify the song on it. The CDDB service combines various computer software media players to identify Jin frequency tracks. These media players use cDDB to move in media files with file names & CD-encoded songs. Another application of CDDB technology allows standalone CD players (not connected to a computer or the 200307874 Internet) to display song names and author information. To do this, the device must locally store GRACENOTE · databases and perform the same techniques described above locally on the device. One disadvantage of CDDB technology is the need to present all pre-records (: 1) in order to be able to identify individual Jin frequency tracks of the CD. Therefore, the listener of the present technique cannot be used to identify individual audio tracks heard from audio sources other than the recording CD. Another form of frequency identification device uses a time printing technique to identify audio tracks. Two known devices apply this time printing technique as sony MARKER and XENOTEI-TAG. These devices are very simple keychain devices, which simply record data and time if a button on the device is pressed. In use, if a listener hears a song from the radio and wants to recognize it, he presses the button on the device and the device records the date and time when the button was pressed. Then, if the device is synchronized with a desktop computer, the user-specific identification symbol of the listener device and the recorded date and time information are transmitted to a server via the Internet. In general, a web page is displayed to display the song player of each radio station that the listener has previously identified as the most frequently listened. The device itself does not store any information about the user's listening to the station when selecting. A web page identifying songs is presented, and it often also provides the option to purchase a CD containing the selected song, etc. The disadvantage of using the time printing technology device is that it cannot be completely automatically stored because the user needs to remember when to start the device and listen to that bag mouth. In addition, the user must operate the computer for audio track identification. In particular, the user must identify the radio station he listens to most often. In addition, the interaction needs to go through the Internet. As a result, the incubation period of a normal Internet lacks the degree, speed, and guilt of the Internet. Obviously, the interaction is generally more than 200307874 than the non-Internet-based audio track identification method. In addition, since these devices only record the start time and date, they are limited to radio broadcast use. In addition, such a device requires the service provider to maintain a database containing the complete playlists and play times of the stations in each market that the service provider is willing to support. Because radio stations do not provide such information, the collection of playlists and ancillary play times are generally performed by a third party. The third party manually identifies and enters the playlists and auxiliary playback times into the database, or automatically recognizes these playlists and auxiliary playback times and stores them in a database by a computer. In any case, identification and storage is complex, requires significant effort, is costly, and is generally limited to the most popular stations, and therefore cannot cover many regions and markets. In addition, the prior art method of flattening audio tracks uses audio fingerprinting. Generally, bone frequency fingerprint identification uses software to identify a database of known "fingerprints" associated with a known audio frequency sample by comparing a dedicated audio symbol or fingerprint (hereinafter referred to as a "fingerprint") of an audio sample with a known Jin frequency sample. Some service providers and / or software applications use digital fingerprinting technology to identify audio tracks. For example, CLANG〇, a software product manufactured by the California AumBLE magic company, uses digital fingerprinting to identify streaming audio broadcasts. The broadcast does not provide relevant audio track media information. The fingerprint performed by AUDIBLE magic is disclosed in US Patent No. 5,918,223. In addition, the supplier of similar green frequency fingerprint recognition technology is AUDITUDE, and its software product Π) 3ΜΑΝ is aimed at users who have a large number of incorrect or incomplete identifications of digital bone frequency files. Through a combination of technologies, including audio finger 200307874 pattern recognition, ID3MAN identifies the audio file and then corrects the identification information associated with the file. The disadvantage of these fingerprint identification devices or services is that users cannot listen to music from desktop computers (except for CD players supported by CDDB, because the device needs to store a large GRACENOTE database and has the disadvantages described above). Another method for identifying audio uses a mobile phone network. If the user wants to identify after listening to the audio, the user calls a designated number to identify the audio. There are at least two ways to provide this service. The first method, provided by the name BUZZHITS (now closed), allows users to call numbers and enter user identifiers that identify callers (and the caller ’s geographic market) and radios that refer to users broadcasting identification audio Station broadcast frequency. Once the broadcast frequency is supplied, a user sample audio clip is provided so that the user can select a sample clip from it to obtain the identity of the audio track. This information is also emailed to the user. Although this telephone service addresses some of the aforementioned disadvantages, it still requires the user to manually interact with the device and force the user to interact with the audio while listening, which is often inconvenient. In addition, once the audio track has been identified, the above audio recognition method does not perform additional actions. As the new subscriber service can download the tone 2, Shao Fei hopes to have the option to purchase and download from a variety of different sources. The second is to complete with the above products / services-processing requires multiple steps. Manual processing: to communicate with the Internet Internet and desktop computers, or mobile phones, interact. & According to the above, the need to avoid the above disadvantages-audio recognition device and method 200307874 is not only convenient and easy to use, but also has accurate recognition at low cost. -Summary of the Invention A specific embodiment of the present invention includes a portable device capable of recording from a microphone, an audio player and / or an audio receiver. In order to identify an audio track, if the portable device is activated, it records an audio sample currently being played by the audio track or received by the radio receiver. If the portable device is not currently playing audio or receiving a radio broadcast, the microphone can record audio samples. This recorded audio sample is stored in the internal memory of the portable device, and later, if a client computer is connected, it is uploaded to the client computer. The client computer processes the audio sample to generate an "fingerprint" of the audio sample and then compares the fingerprint database on the client computer or on the identification server (ID server) of the client computer via the Internet. Once the fingerprint has been identified, the name and author information are returned to the client computer and finally displayed on the portable device. In addition, the portable device itself processes the fingerprint of the recorded audio samples and generates the fingerprint. This has the advantage of reducing the amount of storage space required to store audio samples, as only fingerprints are stored on the portable device. However, this specific embodiment requires that the portable device has sufficient processor power to perform fingerprint recognition. — In addition to returning author and title information, the device performs additional actions once audio samples or tracks have been identified. Examples of additional actions include downloading identification audio tracks for subscriber services, recommending more audio tracks similar to identification audio tracks, and obtaining prices for identifying audio tracks from Internet music dealers. Ideally, these additional actions are selected by selecting a menu item from the player display, and can also be ordered and downloaded from a third party service provider. -11-200307874 According to the present invention, a method for a portable device to recognize audio is provided. An audio sample of an audio track is recorded on a portable device. The frequency samples are then stored in a cache memory of the portable device. Pass the M-frequency sample to the computer device for identification. The computer device receives the beta frequency sample and performs fingerprint recognition on the audio sample to obtain a dedicated audio fingerprint of the tone f "book. The fingerprint database is then searched for a known previous recognition that matches the fingerprint A known finger of the audio track, and a matching device that finds and transmits the relevant identification data of the previously identified audio track is a portable device. Then it receives the identification of the audio sample and displays it on the portable device._ In another specific embodiment, fingerprint recognition is performed on a portable device. In another specific embodiment, the portable device receives and plays—radio broadcasts on the portable device. Receives—instructions to identify—the audio tracks of the riding radio and The radio frequency of the radio broadcast, and the date and time of the automatically recorded portable device receiving the instruction identifying the bone-frequency track. Then transmit the broadcast frequency, month / month and time, and the special device identification symbol to the computer device. Be identified. Received by the computer device. Then search the playlist database to find a known station that matches the incoming / broadcasting / page rate, date, and time. The broadcasting frequency determined by the special device identifier based on the user's geographic location, and the station's broadcast-audio track Know the date and time. Find the audio track, the relevant broadcast frequency, date and time, and send it to the portable device. The portable device later receives and displays information related to the identified audio track. In another tf embodiment, the fingerprint recognition is performed on a portable device. According to the present invention, a portable device, a computer device and an identification feeding device are also provided for performing the above method. Therefore, the recognition method can be automatic with a portable music player. In addition, the scope of use of this device can be expanded to cover a wider range of music "mm, so additional functions can be obtained at a small additional cost. At the same time, this device can help edit the database of radio playlists. (Action 4, external action, and the additional action can be personalized according to the user's preferences. Implementation, Figure 1 is a square ghost picture system that recognizes the audio system according to a specific embodiment of the present invention. 400 includes at least one identification server 102 (hereinafter referred to as "still server") and at least one client computer 106 coupled to another computer via the network 104. The id server is 102 and the client computer 106 is any In the example, the client computer 106 is a desktop computer and the network platform is the Internet. The client computer 106 is coupled to the network 104 via any suitable communication link 108, such as Ethernet, coaxial cable, copper telephone, wire, fiber optic, wireless, infrared or other.-Portable 4 bone-frequency identification device U2 (hereinafter "Portable Device") Coupling client computer 106. Portable device 112 ideal for palm type The portable device 112 is coupled to the client computer 106 via any suitable communication link such as a universal serial bus (USB), Firewire, Ethernet, coaxial cable, copper telephone line, optical fiber, wireless, infrared or other. In another specific embodiment, the client computer 106 is a fixed wireless base station connected to a closed pole / data machine connected to the network office. For example, the client computer is on one side (wireless accuracy-IEEE 802. 11b wireless network) The base station is coupled to the network i04 via a digital subscriber line -13- 200307874 (DSL) gate (not shown). In this specific embodiment, the communication link from the portable device 112 to the client computer 106 is a complete wireless communication link. In another specific embodiment, the client computer 106 and the portable device 112 do not directly communicate with the ID server 102. For example, the portable device 112 includes a mobile communication circuit 102 via a mobile telephone network (network 104). In another embodiment, the portable device 112 needs to independently identify audio. For example, the portable device periodically downloads and updates the playlist and / or fingerprint database from the network 104, as described in detail below and with reference to FIGS. 6 and 9. In a specific embodiment, a playlist supplier 114 and a fingerprint supplier 116 are also coupled to the network 104. The playlist supplier 114 supplies an updated playlist to a server to the ID server 102, the client computer 106, and / or the portable device 112, and the fingerprint supplier 116 supplies a new audio magnet to a server The updated fingerprint data of the track is sent to the ID server 102, the client computer 106, and / or the portable device 112. FIG. 2 is a block diagram of the ii) server 102 and / or the client computer io6 shown. ID server 102 and / or client computer 106 are shown in a schematic diagram to avoid duplication. However, it must be understood that all of the following elements need not be present in all the specific embodiments of the present invention-all components of the ID server 102 and / or the client computer 106 are included as examples only. Ideally, the ω temple server 102 and / or the client computer 106 includes: at least a data processor or a central processing unit (CPU) 202; a memory 21; a user interface device 206; such as a monitor and Keyboard; communication circuit system outline for the communication network 104 (Figure 1), ID servo ϋ 10 · υ, client computer (Figure 0 and / or portable device m (Figure υ and at least one bus connected to these components- 14- 200307874 208. The memory 210 preferably includes an operating system 212, such as VXWORKS, LINUX, or WINDOWS with instructions for processing, accessing, storing, or searching data, etc. The ideal memory 210 also includes a communication program 214 for a communication network. Road 104 (Figure 1), ID server 102 (Figure 1), client computer 106 (Figure 1) and / or portable device 112 (Figure 1); fingerprint identification program 216; search program 218; fingerprint identification database 220 ; Radio playlist database 224; geographic identification symbol 234; "no identification" information 236; and a cache memory 238 for temporary storage of data. The fingerprint recognition program 216 is used to obtain audio samples of an audio track A special identifier i tiger or fingerprint, detailed as follows Figures 4 and 5. The fingerprint recognition program 216 includes instructions for performing fingerprint recognition of audio samples to obtain a dedicated audio fingerprint of an audio sample. The search program 218 is used to search the fingerprint database 220 for identification audio. Figures 4 to 6 below. The fingerprint identification database 220 includes fingerprints of many known audio samples or audio tracks and their related identification information 222 (1)-(N), such as song title, author, or others. In another specific embodiment, a radio playlist database 224 is provided. In this embodiment, the radio playlist database 224 includes a number of radio frequencies 226 (1)-(N) and playlists 228 (i)-(N) associated with each frequency 226 (1)-(N). Each playlist 228 (1)-(N) includes a date 230 (1) · (N) and time 232 (1)-(N), and the identity of each audio track broadcast at that date and time 232 (1) -(N). For example, the radio station KJAZ has a frequency of 98. 7 FM, and a playlist including Frank Sinatra's "New York, New York" was broadcast on January 21, 2002, 9:00 AM. Since many radio stations scattered throughout the world share the same frequency, 200307874 is provided as a geographic identifier 234 to identify the radio station or the frequencies 226 (1)-(N) in a specific area. The logical identification 234 is provided by any suitable method. In a specific embodiment, the user provides the geographic identification symbol. In another specific embodiment, the geographic identifier 234 is obtained from a user's private network address. For example, an Internet access (IP) address of a client computer touching / or a portable device 2 may be used to estimate the geographic area of the user. In another specific embodiment, the global positioning system, GPS, and charger (GPS) combined with the client computer and / or the portable device 112 may be used to determine the geographic area of the user. If the feeder and / or the client computer 106 cannot recognize an audio track, a "no identification" message 236 is used to notify the user that no identification is available. Alternatively, before receiving the non-identifying information, the user may obtain some "recently matched" possible identities. Those skilled in the art must understand that certain components of these devices are still not required by the server 2 and / or the client's private knowledge 106. For example, the fingerprint identification program us, the search program 218, and the fingerprint database 224 are only required for the device that generates audio track fingerprint identification. In other words, if fingerprint recognition occurs on the server 102, the above-mentioned memory 210 components need only appear on the server 102. Similarly, in this embodiment, if the identification occurs on the portable device 112 (Figure 丨), neither the server 102 nor the client computer 106 need the components of the memory 21 above. FIG. 3 is a block diagram of the non-portable device 112 of FIG. 1. Those skilled in the art must understand that all the following elements of the portable device 112 need not be present in all the specific embodiments of the present invention, but include only the elements as examples. The car is ideal, and the portable device 112 includes: at least one data processor or central processing unit 7C (CPU) 302;-memory 310; user interface device 308; buttons, a screen and a headset; communication Circuit system 304 is used for communication network-16-200307874 104 (Figure 1), ID server 102 (Figure 1) and / or client computer 106 (Figure 1); one or more audio players 350, such as a CD or MP3 A player; a microphone 352; a radio receiver 354 and an antenna 356 for receiving radio broadcasts; and at least one bus 306 connected to these components. The memory 310 preferably includes an operating system 312, such as VXWORKS, LINUX, or WINDOWS with instructions for processing, accessing, storing, or searching data, etc. The memory 310 ideally also includes a communication program 314 for the communication network 104 (Figure 1 ), ID server 102 (Figure 1) and / or client computer 106 (Figure 1); fingerprint identification program 316; search program 318; fingerprint identification database 320; radio playlist database 324; geographical identification symbol 334; geographical identification Program 336;-"No identification" information 338; Recording program 340; Player program 342; Radio program 344;-Cache memory 346 for temporarily storing data; Frequency detection program 358; and Clock 360. The fingerprint identification program 316 is used to obtain a special identification symbol or fingerprint on an audio sample of an audio track. The detailed description is shown in FIG. 6 below. Similarly, in this embodiment, the search program 318 is used to search the fingerprint identification database 320 for identification audio, as described in detail below. The fingerprint identification database 320 includes fingerprints of many known audio samples or audio tracks and their associated identification information 322 (1)-(N), such as song title, author, or others. 'In a further embodiment, there is a radio playlist database 324. In this embodiment, the radio playlist database 324 includes a number of radio frequencies 326 (1)-(N) and playlists 326 (1)-(N) associated with each frequency 328 (1)-(N). Each playlist 328 (1)-(N) includes a date 330 (1)-(N) and time 332 (1) _ (N), and the identity of each audio track broadcast at that date and time 332 (1) -(N). 200307874 is also used in the other specific embodiments described above, providing a geographic identification symbol 334 to assist in identifying a radio-station or a particular area frequency 326 (1)-特别. For example, the geographic identifier 334 is selected from a set of frequencies stored in the device based on the identified geographic area. The present geographic identifier 334 is provided by any suitable method. In a specific embodiment, the user provides the geographic identification symbol 334. In another embodiment, the geographic identification symbol 334 is obtained by the geographic identification program 336. As described above, this can be determined from the user's private network address. For example, an Internet Protocol (IP) address of the portable device 112 may be used to estimate the geographic area of the user. In another embodiment, a global positioning system (GPS) combined with the portable device 112 can be used to determine the geographic area of the user. In all embodiments, the "no identification" information 236 is used to notify the user that no identification is available if the portable device 112 cannot identify the audio track. In a specific embodiment of the present invention, if the fingerprint identification system of the audio track is used to identify the audio, the recording program 340 records an audio sample 348 stored in the cache memory 348. The audio samples are recorded from the audio player 350, the microphone 352, and / or the radio receiver 354. In a specific embodiment of the present invention, if the date, time, and radio station or broadcast frequency are used to identify audio from a radio station playlist, the recording program is used to record the date, time, and record a stored in the cache Memory 348 radio or radio 349 frequency. It is desirable to provide a player program 342 to play audio on the audio player 35q. The player program 342 is particularly needed for playing digital audio, wMP3 audio tracks or other. A radio program 344 is preferably provided to play the radio received by the antenna 356 via the radio -18-200307874 receiver 354. However, it must be understood that all the specific embodiments of the present invention are not required-all the elements of the memory 31 described above are present and only the elements as examples are included. The frequency program 358 is used to detect the frequency broadcast by the radio station, and the clock 36 is used to maintain the date and time. For a detailed description of the frequency program 358 and the clock 36, refer to Figs. FIG. 4 is a flowchart of a method for identifying audio performed by an ID server ι02 according to a specific embodiment of the present invention. In a specific embodiment, the audio player 350 (FIG. 3) and / or the player program 342 (FIG. 3) of the portable device 112 plays audio through the user interface device 308 (FIG. 3) in step 402. . For example, a built-in] vq> 3 player plays audio to the user via a headset. In another embodiment, the radio receiver 345 (FIG. 3) and / or the radio program 344 (FIG. 3) receives and plays audio broadcasts via the headset of the portable device. An instruction is then received at step 404 to identify the audio. Ideally, these instructions come from the user, such as the user pressing an "identify now" button on the portable device, or other. In another specific embodiment, a recording instruction is automatically received. For example, an audio sample is automatically recorded every 2 minutes. Steps 402 and 404 of playing and receiving instructions are not necessary for the present invention and need not occur in some specific embodiments. At step 406, an audio sample is recorded by the recording program 342 (Fig. 3) and an audio sample 348 (Fig. 3) is stored in the cache memory 346 (Fig. 3). In a specific embodiment, continuous audio recording and automatic segmentation into audio samples have sufficient length for fingerprint identification. For example, the audio is continuously recorded and automatically divided into 30-second audio samples, which are continuously transmitted to the ID server 102 for identification. -19- 200307874 If the audio player 350 (Fig. 3) and / or the player program 342 (Fig. 3) of the portable device 112 does not play the audio, it is recorded in step 406 via the microphone 352 (Fig. 3) Audio. In a specific embodiment, if the portable device 112 is coupled to the network 104 via the client computer 106, the audio program is transmitted to the client computer 106 at step 408 by the communication program 314 (FIG. 3). As mentioned above, this communication occurs via the communication chain 110, such as a serial port connected wirelessly, or other. The audio sample received by the client computer at step 41 is transmitted to the ID server 102. It must be understood that the portable device does not have a permanent client computer 106 and / or the communication link of the server 102, so the audio samples are stored in the cache memory 346 (Figure 3) until the portable device 112 and the client computer A connection is established between 106 and / or m server 102. In another specific embodiment, there is no client computer, so the portable device 112 communicates with the ID server 102 via a mobile phone network, and the audio sample is directly transmitted to the ID server 102 at step 408. The ID server 102 receives the audio sample at step 412. Then, in step 414, the fingerprint identification program 216 (FIG. 2) performs fingerprint identification of the audio sample on the 10 server 102 to determine a specific identification symbol or fingerprint of the sample according to the characteristics or acoustic characteristics of the audio sample. This characteristic or acoustic feature includes the analogy of the audio sample, pitch, tolerance, bandwidth, Mel Frequency Cepstrum Coefficient (MFCC) or other suitable methods for identifying the bone frequency of the fingerprint, as disclosed in US Patent No. 5,918,223 No., which is incorporated herein by reference. Then in step 416, search the fingerprint database 22 (Fig. 2) to find the fingerprint of the audio magnetic ID that was identified before 200307874 that matches or partially matches the fingerprint of the audio sample recorded on the portable device. . If no match is found (418-No), a "no identification" message 236 (FIG. 2) is transmitted to the portable device 112 in step 430. In this specific embodiment, the portable device 112 is coupled to the ii) server 102 'via the client computer, and the client computer receives the "no identification" information and transmits the information to the portable device at step 432. The portable device receives the "no identification" message in step 434, and displays it to the user in step 436 to notify the user that the audio is not recognizable. Ideally, the information is displayed on the screen of the portable device. If the fingerprint is found to match the previously-recognized audio track (418_Yes) ', then at step 42Q, an identification data related to the previously-recognized audio track is transmitted to the portable device 112. In this embodiment, the portable device n2 is coupled to the ID server 102 via the client computer. The client computer receives the identification of the audio sample and transmits the information to the portable device in step 422. The portable device receives and displays the identification of the audio sample at steps 424 and 426, respectively. For example, display an author and song name. In a preferred embodiment, step 428 performs additional actions. These additional actions are performed only after the identity of the audio sample is known and include, for example, the client computer 106 and / or the portable device 112 automatically displaying the identified author's webpage, biography, and magnetic disc category; automatically displaying a webpage sales author A collection of songs or disks; download the audio track from the subscriber service; recommend a similar audio; get the identification track from an online music dealer for different prices, or other. Ideally, these additional actions are selected by selecting a menu item from the display of the portable device, and can also be ordered and downloaded from a third party service provider. The portable device can also retrieve information or conversations about the ad. In addition, all foreign materials are executed on all audio samples -21-200307874, regardless of whether the audio identity is known or unknown. For example, a user downloads a digital audio file and has the identity of all or part of a song, but must still send the song for identification to receive additional information about the audio track or perform certain actions on the audio track. FIG. 5 is a flowchart of another method for identifying audio performed by the client computer ιο according to another embodiment of the present invention. In a specific embodiment, the audio player 350 (FIG. 3) and / or the player program 3 of the portable device 112 (see FIG. ^ Plays the audio via the user interface device 308 (FIG. 3) at step 502. For example, A built-in MP3 player plays audio to the user via a headset. Then, in step 504, instructions are received to identify the audio. Ideally, these instructions come from the user. For example, if the user presses a "now "Identify" button. In another specific embodiment, a recording instruction is automatically received. For example, an audio sample is automatically recorded every 2 minutes. Steps 502 and 504 for playing audio and receiving instructions to identify the audio must be explained. It is necessary for the invention and does not need to occur in some specific embodiments. In step 506, an audio sample is recorded by the recording program 342 (FIG. 3) and an audio sample 348 (FIG. 3) is stored in the cache memory 349 (FIG. 3). 3). In a specific embodiment, continuous recording of audio and automatic segmentation into audio samples have sufficient length for fingerprint identification. For example, continuous recording of audio and automatic segmentation into 30-second audio samples and continuous operation by the client computer 106 No. If the audio player 350 (FIG. 3) and / or the player program 342 (FIG. 3) of the portable device 112 does not play audio, the audio is recorded via the microphone 352 (FIG. 3) at step 5506. In the specific embodiment, 'if the portable device U2 is wirelessly connected to the customer's electricity-22- 200307874 brain 106, the communication program 314 (Figure 3) in step 508 continuously transmits the audio sample to the customer's electricity m06 until wireless The connection was interrupted. The audio samples received by the client computer 106 in step 510. It must be understood that the portable device does not have a permanent communication link with the client computer 106, so the audio samples are stored in cache memory until the portable device 112 When the connection is established with the client computer 106. Then, in step 512, the fingerprint recognition program 216 (FIG. 2) performs fingerprint recognition of the bone frequency sample on the client computer 106 to determine one of the samples according to the characteristics or acoustic characteristics of the audio sample. The special identification symbol or fingerprint is as described above. In another specific embodiment, the fingerprint identification program 316 (FIG. 3) performs fingerprint identification of the Jin frequency sample on the portable device 112 to determine based on the characteristics or acoustic characteristics of the audio sample. kind A special identification symbol or fingerprint. Then the fingerprint is transmitted to the client computer 106, which searches for the fingerprint database at step 514. Then it searches the fingerprint database 220 (FIG. 2) at step 514 to find a matching or part. A matching fingerprint of the audio sample recorded on the portable device. If no matching (516-No) is found, then in step 526, a "no identification" message 236 (Figure 2) is transmitted to the portable device-112, The "no identification" information is received in step 528 and the information is displayed in step 530 to notify the user that the audio cannot be recognized. Ideally, the information is displayed on the screen of the portable device. If a match is found (516-Yes), step 518 transmits the identification of the audio sample to the portable device 112. Steps 520 and 522 portable devices receive and display audio samples for identification, which is ideal on the screens of portable devices. For example, an author and song name are displayed. In a preferred embodiment, step 524 performs additional actions of -23- 200307874, as described above. FIG. 6 is a flowchart of another method of identifying audio performed by the portable device 112 according to another embodiment of the present invention. In a specific embodiment, the audio player 360 (FIG. 3) and / or the player program 342 (FIG. 3) of the portable device 112 plays audio at step 602 via the user interface device 308 (FIG. 3). . For example, a built-in MP3 player plays audio to a user via a headset. An instruction is then received at step 604 to identify the audio. In another specific embodiment, the recording instruction is automatically received. It must be explained that steps 602 and 404 of playing audio and receiving instructions to identify audio are not necessary for the present invention and need not occur in some specific embodiments. -At step 606, an audio sample is recorded by the recording program 342 (Fig. 3) and an audio sample 348 (Fig. 3) is stored in the cache 1 memory 349 (Fig. 3). In a specific embodiment, the audio is continuously recorded and automatically divided into audio samples with sufficient length for fingerprint identification. Audio samples 348 (Figure 3) are ideally temporarily stored in cache memory 346 (Figure 3). If the audio player 350 (FIG. 3) and / or the player program 342 (FIG. 3) of the portable device 112 does not play audio, the audio is recorded via the microphone 352 (FIG. 3) in step 606. _ Step 608 The fingerprint identification program 316 (FIG. 3) performs fingerprint identification of the audio sample on the portable 112 device to determine a specific identification symbol or fingerprint of the sample based on the characteristics or acoustic characteristics of the audio sample. This characteristic or acoustic characteristic includes the analog waveform, volume, pitch, brightness, bandwidth, Mel Frequency Cepstral Coefficient (MFCC) of the audio sample, or others. A suitable method for fingerprint recognition audio is disclosed in U.S. Patent No. 6,918,223, which is incorporated by reference into this-24-200307874 Article 0 and then searches the fingerprint database 320 (Figure 3) at step 610 to find a matching or Partially matched fingerprints of the audio samples recorded on the portable device. If no match is found (612-No), a "No Recognition" message 340 (Fig. 3) is displayed at step 614 to notify the user that the audio cannot be recognized. Ideally, this information is displayed on the screen of the portable device. If a match is found (612-Yes), it is shown in step 616 that the identification of the audio sample ' is ideal on the screen of the rockable device. For example, show an author and song name. In a preferred embodiment, step 618 performs additional actions, as described above. . One advantage of performing fingerprint recognition on a portable device is to save the memory of the portable device because the fingerprint is substantially lower than the audio sample. In addition, another embodiment of the present invention uses only a peer-to-peer network composed of portable devices connected to the network to identify audio. For example, if the device has wireless network capabilities and fingerprint recognition is performed on the portable device, identification of audio samples is generated by searching for limbs on other network-connected devices.另外 Another specific embodiment for performing fingerprint identification on a portable device has a central service station, such as a record storage to perform fingerprint identification. This specific embodiment alleviates the burden on the service station because each station does not generate fingerprints and needs less data to transfer and maintain. Fig. 7 is a flowchart of a method for identifying radio broadcast audio by an identification server "servo '" according to a specific embodiment of the present invention. The radio receiver 354 (FIG. 3) and / or the radio program 344 (FIG. 3) of the portable device m receives and plays a radio broadcast at step 702. Then at step 704 -25- 200307874 a command is received to identify the audio track of the radio broadcast. Ideally, these instructions come from the user, such as the user pressing an "Now Identify" button on the portable device or other. In another specific embodiment, an audio track identification command is automatically received. For example, an automatic attempt to identify audio tracks occurs every 2 minutes. At step 706, the recording program 342 (Fig. 3) records and stores the radio frequency of the radio station, and the date and time when the portable device 112 received the instruction to identify the audio track. For example, record a broadcast frequency of 957kHz, a date of February 2nd, February 2nd, and a time of 113H00. The recording program 340 (Fig. 3) obtains the frequency from the frequency detection program 358 (Fig. 3) and the date and time from the clock 36_3). In its simplest form, the frequency estimation procedure 3_ 3) is just to find out the wireless packet frequency selected by the user, that is, to read the value adjusted by the radio receiver. Alternatively, the frequency detection program 358 (Fig. 3) detects the broadcasting residual rate, and the general broadcasting and audio signals are together. This uses a radio data service (Na) that transmits the actual radio identification. In general, radio identification is more reliable because radio identification is not related to geography. RDS actually transmits information about the owner of the radio station, = the information clearly defines the playlist to search. The frequency_program says (Figure 3) to detect the n-broadcast frequency, which is to detect the frequency adjusted by the radio receiver. It must be understood that the wide frequency is determined automatically, that is, the user does not supply radio frequency to the frequency detection program 358 (Figure 3). The date and time of the clock 360 (Fig. 3) should be set by the user at first, or the portable device can automatically set the clock using known techniques ^ Technology to remotely control the clock and synchronize with a reliable time = I). The technology of remote clock synchronization is disclosed in U.S. Patent No.4, No.3, No.3, No.3, No.2, and No.3, which are incorporated by reference in this article. 26-200307874 Article 0 Record date, time, and frequency Clock / frequency data) 349 (Figure 3) is stored in cache memory 346 (Figure 3). ^ _ ^} Cargo 708, portable device 112 transmits clock / frequency data 349 (Figure 3). In this specific embodiment, if the portable device 112 transmits the clock / frequency signal 349 (FIG. 3) to the client via the client computer ’s ID server “1 02 'the communication procedure at step 7 08 seems to be ⑴”. Computer 1G6. As mentioned earlier, this communication occurs
經通信鏈11G(圖丨），如-串聯埠連接、無線連接或其他。在 步驟710由客戶電腦1〇6接收時脈/頻率資料項圖3)及傳送至 ID伺服器1〇2。 —必眉了解可攜式裝置不具有一永久的客戶電腦舰及/或id 伺服器102的通信鏈，所以時脈/頻率資料从％圖3)儲存在快 取屺憶體直到可攜式裝置112及客戶電腦1〇6及/或仍伺服器 102之間建立連接。 在另外具體實施例中，並沒有客戶電腦，所以可攜式裝 置112經一行動電話網路直接通信ID伺服器1〇2，在步驟7〇8 音頻樣本直接傳送至ID伺服器1〇2。 步驟712，ID伺服器1〇2接收時脈/頻率資料349(圖3)。在步 驟716無線電播放表資料庫224(圖2)搜尋記錄在可攜式裝置 112的時脈/頻率資料349(圖3)的相配。如果找不到相配（718_N〇) ’在步驟730傳送一「無識別」信息236(圖2)至可攜式裝置U2 。在本具體實施例中，可攜式裝置112經客戶電腦耦合ID伺 服器102 ’客戶電腦接收該「無識別」信息及在步驟732傳送 該信息至該可攜式裝置。 -27- 200307874 該可攜式裝置在步騾734接收該「無識別」信息，及在步 驟736顯示予使甩者通知使用者該音頻不能識別。較理想該 信息顯示在該可攜式裝置的螢幕上。 如果找到相配（718-Yes)，步驟72〇音頻磁軌的識別資料傳送 至茲可攜式裝置112。在本具體實施例中，可攜式裝置112經 客戶電腦耦合至ID伺服器1〇2，步驟722客戶電腦接收音頻樣 本的識別及傳送該識別至可攜式裝置。在步騾724及726該可 攜式裝置分別接收及顯示該晉頻樣本的識別，如作者及歌 曲名稱。在一較佳具體實施例中，步驟728執行額外動作， 如上述。 圖8為根據本發明另外具體實施例由客戶電腦1〇6執行識 別的另外識別無線電廣播音頻方法的流程圖。無線電接收 器354(圖3)及/或可攜式裝置1丨2的無線電程序344(圖在步驟 802接收及播放可攜式裝置112上的一無線電廣播。然後在步 騾804接收指令以識別無線電播放的音頻磁執。較理想地， 這些指令來自使用者，如使用者按下可攜式裝置上的一「現 在識別.」鈕或其他。在一另外具體實施例中，自動接收一 識別音頻磁軌指♦。例^，每2分鐘產生自自嘗試識別音頻 磁軌。 步驟806，記錄程序342(圖3)儲存時脈/頻率資料349(圖3)。 與上述情況相似，記錄程序340(圖3)從頻率偵測程序(圖3) 獲得時脈/頻率資料349(圖3)及從時脈36〇(圖3)獲得日期及時 間。 時脈/頻率資料339(圖3)儲存在快取記憶體346(圖3)，及步 -28- 200307874 驟808傳輸時脈/頻率資料349(圖3)至客戶電腦1〇6。如前述， 本通“經通信鏈_11〇(圖i)發生，如一串聯埠連接、無線連接 或其他。然後在步驟810由客戶電腦1〇6接收該音頻樣本。 必須了解可攜式裝置不具有一永久的客戶電腦106的通信 鍵’所以時脈/頻率資料349(圖3)儲存在快取記憶體直到可 攜式裝置112及客戶電腦1〇6之間建立連接時。 在步驟814搜尋無線電播放表資料庫224(圖2)以找出相配 的圮錄在可攜式裝置H2的時脈/頻率資料349(圖3)。如果找 不到相配（816-No)，步驟826傳送一「無識別」信息236(圖2) 至可攜式裝置112。在步騾828及830可攜式裝置接收及顯示 該無識別信息。 如果找到相配（816-Yes)，步騾818傳送該音頻磁軌的識別至 可攜式裝置112。步騾820及822可攜式裝置接收及顯示音頻 樣本的識別，如作者及歌曲名稱。在一較佳具體實施例中， 在步驟824執行额外動作，如上述。 圖9為根據本發明另外具體實施例由可攜式裝置112執行 識別的另外識別無線電廣播音頻方法的流程圖。無線電接 收裔354(圖3)发/或可攜式裝置η]的無線電程序344(圖3)在步 驗902接收及播放一可攜式裝置n2上的無線電廣播。在一上 述相似的情況下，步驟904接收一指令以識別該無線電廣播 的音頻磁軌。 在步驟906 ’圮錄程序342(圖3)儲存時脈/頻率資料349(圖 3)。在類似上述的情況下，記錄程序34〇(圖3)從頻率偵測程 序358(圖3)獲得時脈/頻率349(圖3)及從時脈360(圖3)獲得曰 -29-Via the communication chain 11G (Figure 丨), such as-serial port connection, wireless connection or other. In step 710, the client computer 106 receives the clock / frequency data item (Figure 3) and transmits it to the ID server 102. —Be sure to understand that the portable device does not have a permanent communication link between the client computer and / or the id server 102, so the clock / frequency data is stored from the cache memory to the portable device until the portable device A connection is established between 112 and the client computer 106 and / or the server 102. In another specific embodiment, there is no client computer, so the portable device 112 directly communicates with the ID server 102 via a mobile phone network, and in step 708, the audio sample is directly transmitted to the ID server 102. In step 712, the ID server 102 receives the clock / frequency data 349 (FIG. 3). In step 716, the radio playlist database 224 (Fig. 2) searches for a match of the clock / frequency data 349 (Fig. 3) recorded in the portable device 112. If no match is found (718_N0) 'in step 730, a "no identification" message 236 (Fig. 2) is transmitted to the portable device U2. In this specific embodiment, the portable device 112 receives the "no identification" information via the client computer coupled to the ID server 102 'and the client computer transmits the information to the portable device in step 732. -27- 200307874 The portable device receives the "No Recognition" message at step 734 and displays it at step 736 to cause the dumper to notify the user that the audio is not recognizable. Ideally, the information is displayed on the screen of the portable device. If a match is found (718-Yes), the identification data of the audio track is transmitted to the portable device 112 at step 72. In this specific embodiment, the portable device 112 is coupled to the ID server 102 via the client computer. In step 722, the client computer receives the identification of the audio sample and transmits the identification to the portable device. At steps 724 and 726, the portable device receives and displays the identification of the Jin frequency samples, such as the author and song name. In a preferred embodiment, step 728 performs additional actions, as described above. FIG. 8 is a flowchart of another method for identifying radio broadcast audio performed by the client computer 106 in accordance with another embodiment of the present invention. The radio receiver 354 (FIG. 3) and / or the radio program 344 of the portable device 1 2 (illustrated by receiving and playing a radio broadcast on the portable device 112 at step 802. Then receiving an instruction at step 804 to identify Audio magnetically played by radio. Ideally, these instructions come from the user, such as the user pressing an "identify now" button on the portable device or other. In another embodiment, an identification is automatically received. The audio track refers to ♦. For example, ^ is generated every 2 minutes from trying to identify the audio track. Step 806, the recording program 342 (Figure 3) stores the clock / frequency data 349 (Figure 3). Similar to the above situation, the recording program 340 (Figure 3) Obtain clock / frequency data from frequency detection program (Figure 3) 349 (Figure 3) and date and time from clock 36 (Figure 3). Clock / frequency data 339 (Figure 3) Stored in cache memory 346 (Figure 3), and steps -28- 200307874, step 808 to transmit clock / frequency data 349 (Figure 3) to the client computer 106. As mentioned above, this communication "via communication link_11〇 (Figure i) Occurs, such as a serial port connection, wireless connection, or other. Then in step 810, the client computer 1 6 to receive the audio sample. It must be understood that the portable device does not have a permanent communication key for the client computer 106, so the clock / frequency data 349 (Figure 3) is stored in cache memory until the portable device 112 and the client computer When a connection is established between 106. At step 814, the radio playlist database 224 (Figure 2) is searched to find the matching clock / frequency data 349 (Figure 3) recorded in the portable device H2. If you find If it does not match (816-No), step 826 sends an "unidentified" message 236 (Figure 2) to the portable device 112. At steps 828 and 830, the portable device receives and displays the unidentified information. If found Matching (816-Yes), Step 818 transmits the identification of the audio track to portable device 112. Steps 820 and 822 portable devices receive and display the identification of audio samples, such as the author and song name. In a preferred embodiment, additional actions are performed at step 824, as described above. Figure 9 is a flowchart of another method for identifying radio broadcast audio performed by the portable device 112 according to another embodiment of the present invention. The radio receiver 354 ( Figure 3) Hair and / or portable η] radio program 344 (Figure 3) receives and plays a radio broadcast on a portable device n2 at step 902. In a similar situation described above, step 904 receives an instruction to identify the audio track of the radio broadcast At step 906, the recording program 342 (Fig. 3) stores the clock / frequency data 349 (Fig. 3). In a case similar to the above, the recording program 34 (Fig. 3) starts from the frequency detection program 358 (Fig. 3) Obtain a clock / frequency of 349 (Figure 3) and -29 from Clock 360 (Figure 3)
I 200307874 期及時間。 時脈/頻率資料349(圖3)儲存在快取記憶體346(圖3)。步騾 910搜尋無線電播放表資料庫324(圖3)儲存時脈/頻率資料 349(圖3)的相配。如果找不到相配（912-N〇)，則在步驟914傳 送「無識別」信息338(圖3)。如果找到相配（912_Yes)，在步 驟916顯示該晋頻磁軌的識別。在一較佳具體實施例中，及 在步驟918執行額外動作，如上述。 另外，ID伺服器102(圖1}、客戶電腦1〇6(圖1}或可攜式裝 置112(圖1)上的指紋識別資料庫22〇(圖2)及/或32〇(圖3)，較 理想地，定期_由指紋供應商116(圖丨）更新。同樣地，ID伺服 器1〇2(圖1)、客戶電腦1〇6(圖1}或可攜式裝置112(圖1}上的無 線電播放表資料庫224(圖2)及/或324(圖3)，較理想地，定期 由播放表供應商114(圖1)更新。 在本發明的一另外具體實施例中，可攜式裝置112將時脈/ 頻率附加在音頻樣本。本資料可供許多用途使用，如決定 使用者的收聽的習慣，如收聽目標、廣告或其他。 在本發明的一另外具體實施例中，如果一使用者收聽二 次裝置，如汽-車無線電，可使用可攜式裝置以識別二次裝 置凋整的廣播頻道及從該頻道記錄。作這種記錄，可攜式 裝置搜尋所有頻道直到找到訊號與經可攜式裝置傳聲器聽 到的空中音頻相配的電台。然後，使用上述技術之一識別 邊音頻。本具體實施例避免有關記錄空中離訊的缺點。另 外，記錄調整的無線電頻率，該頻率可用來增加指紋識別 處理及提供額外資訊至資料庫。 200307874 本發明的另外具體實施例利用一組預設的音調廣播。在 一具體實施例使用音調識別預定識別的音頻的開頭及結 尾。例如，在音頻磁軌之前及之後，無線電台傳輸一或更 多音1周。可攜式裝置112(圖丨）的配置係用來記錄由該音調包 封的骨頻磁軌。一旦記錄，如上述識別該音 在另外具體實施例中，音調本身包含識別資料。例如， 一串音_ (如3快嗶聲）表示一識別符號如一連續數。本識別 符號用於檢查可攜式裝置112(圖丨）或ID伺服器1〇2(圖丨）的資 料庫相關資訊。另外的例子，為一串的音調代表作者姓名、 歌曲名稱、無·線電台識別符號或其他。該一串的音調也可 以由人耳識別，但必須符合音調與正常音頻區別的規定。 例如必須使用預定音碉或符合規定長度的音調開始。該 具體實施例的一種用法，當地樂團將其音調識別符號登記 於貝料庫以便由系統加以識別。可攜式裝置同步化後恢復 ”料如口泝使用者那裡1己錄特別當地樂團或那裡購買含 ^當地_的歌曲的CD。另外，音調識別符號可以關連特 疋作者#、、泉私台φ及用於作市場推銷。音調識別符號可 傳輸至其他C用者經任何適#方法，如電子郵件，從一可 攜式二置發射波束至另一裝置或其他。另夕卜，音調及音調 識別付號可用I協助任何上述音頻磁軌識別別方法。 前述本發明之特定具體實施例的說明料為解說之用。 並非完全或用來限制本發明為所述㈣式。根據以上說明， 可看出本發明有許多修改及蠻承 、 卞夕1夕？文及夂更。例如，任何前述具體實 施例或方法，可以相互社八，特別Α 、 …口特別地，如果具體實施例或 -31 - 200307874 方法的結合有助於音頻磁軌的識別。所有具體實施例之選 擇及說明，係為對本發明及其實際應用的原理做出最好的 解釋，藉以使其他熟悉本技術者對本發明及具有各種修改 的各種具體實施例做最好的利用以適合所希望的特定用 途。另外，步驟的次序並不需要依照所配置的次序。本發 明之範圍由附件申請專利範圍及其相當的範圍加以定義。 圖式簡單說明 為了更了解本發明的性質及目#，須參考下列的詳細說 明及附圖，其中： 圖1為根據本發明的一具體實施例利用可攜式裝置識別音 頻系統的示意圖； 圖2為識別伺服器（ID伺服器）及/或圖丨所示的客戶電腦的 方塊圖； 圖3為圖1所示可攜式裝置的方塊圖； 圖4為根據本發明具體實施例由一仍伺服器執行識別的一 種識別音頻方法的流程圖； 圖5為根據本發明另外具體實施例由客戶電腦執行識別纪 另外識別音頻-方法的流程圖； 圖6為根據本發明另外具體實施例由可攜式裝置執行識及 的另外識別音頻方法的流程圖； 圖7為根據本發明具體實施例由_識別伺服器陶服器 執行識別的—種識別無線電廣播音頻方法的流程圖； 圖8為根據本發明另外具體實施例由客戶電腦執行識別的 另外識別無線電廣播音頻方法的流程圖；及 -32- 200307874 圖9為根據本發明另外具體實施例由可攜式裝置執行識別 的另外識別無線_電廣播音頻方法的流程圖。 各圖中相似參考號碼表示相對部件。 圖式代表符號說明 100 識別音頻系統 102 ID伺服器 104 網路 106 客戶電腦 108 通信鏈 110 通信鏈 112 可攜式裝置 114 播放表供應商 116 指紋供應商 202 中央處理單元 204 通信電路系統 206 使用者介面裝置 208 匯流排 210 記憶體 212 操作系統 214 通信程序 216 指紋識別程序 218' 搜尋程序 220 指紋識別資料庫 222 指紋I 200307874 Issue and time. Clock / frequency data 349 (Figure 3) is stored in cache memory 346 (Figure 3). Step 910 searches the radio playlist database 324 (Figure 3) to store the clock / frequency data 349 (Figure 3) for a match. If no match is found (912-No), a "no identification" message 338 is transmitted in step 914 (Figure 3). If a match is found (912_Yes), the identification of the Jin frequency track is displayed in step 916. In a preferred embodiment, and in step 918 additional actions are performed, as described above. In addition, the ID server 102 (FIG. 1), the client computer 106 (FIG. 1), or the fingerprint identification database 22 (FIG. 2) and / or 32 (FIG. 3) on the portable device 112 (FIG. 1) ), Ideally, periodically updated by fingerprint supplier 116 (Figure 丨). Similarly, ID server 102 (Figure 1), client computer 106 (Figure 1) or portable device 112 (Figure The radio playlist database 224 (Fig. 2) and / or 324 (Fig. 3) on 1} is ideally updated regularly by the playlist supplier 114 (Fig. 1). In another embodiment of the present invention, The portable device 112 adds the clock / frequency to the audio sample. This information can be used for many purposes, such as determining the user's listening habits, such as listening goals, advertisements, or others. In another specific embodiment of the present invention If a user listens to a secondary device, such as a car-vehicle radio, he can use a portable device to identify the broadcast channel of the secondary device and record from that channel. For this record, the portable device searches all Channel until you find a station with a signal that matches the aerial audio heard through the microphone of the portable device. Then use the above technique An identification side audio. This specific embodiment avoids the disadvantages related to recording aerial separation. In addition, the adjusted radio frequency is recorded. This frequency can be used to increase fingerprint recognition processing and provide additional information to the database. 200307874 Another specific embodiment of the present invention Broadcast using a preset set of tones. In one embodiment, tones are used to identify the beginning and end of a predetermined identified audio. For example, before and after an audio track, a radio station transmits one or more tones for 1 week. Portable The configuration of the device 112 (Figure 丨) is used to record the bone frequency magnetic track enclosed by the tone. Once recorded, the tone is identified as described above. In another specific embodiment, the tone itself contains identification information. For example, a string of _ (Such as 3 fast beeps) indicates an identification symbol such as a continuous number. This identification symbol is used to check the database related information of the portable device 112 (Figure 丨) or ID server 102 (Figure 丨). Another example , Is a string of tones representing the author's name, song name, radio station identification symbol, or other. The string of tones can also be recognized by the human ear, but must be Provisions for the difference between tones and normal audio. For example, it is necessary to start with a predetermined tone or a tone that meets the prescribed length. One use of this specific embodiment is that the local orchestra registers its tone identification symbols in the shell database for identification by the system. May Resume after the synchronization of the portable device ”, such as tracing the user ’s 1 local recorded orchestra or purchase a CD containing ^ local_ songs. In addition, the tone identification symbol can be related to the special author # 、、 泉 私 台 φ And used for marketing. Tone identification symbols can be transmitted to other users through any suitable method, such as e-mail, from a portable two-transmission beam to another device or other. In addition, tones and tones The identification number can assist with any of the above audio track identification methods. The foregoing description of the specific embodiments of the present invention is illustrative. It is not complete or used to limit the present invention to the aforementioned formula. According to the above description, it can be seen that the present invention has many modifications and strict inheritance. Text and change. For example, any of the foregoing specific embodiments or methods may be mutually exclusive, especially A, ... particularly if the combination of specific embodiments or -31-200307874 methods is helpful for audio track identification. The selection and description of all specific embodiments are to make the best explanation of the principles of the present invention and its practical application, so that those skilled in the art can make the best use of the present invention and various specific embodiments with various modifications. Suitable for the specific purpose desired. In addition, the order of the steps need not follow the configured order. The scope of the present invention is defined by the scope of the attached patent application and its equivalent scope. BRIEF DESCRIPTION OF THE DRAWINGS In order to better understand the nature and purpose of the present invention, the following detailed description and drawings must be referred to, in which: FIG. 1 is a schematic diagram of identifying an audio system using a portable device according to a specific embodiment of the present invention; 2 is a block diagram of an identification server (ID server) and / or a client computer shown in FIG. 丨; FIG. 3 is a block diagram of the portable device shown in FIG. 1; FIG. 4 is a block diagram of a portable device according to a specific embodiment of the present invention; FIG. 5 is a flowchart of a method for recognizing audio according to another embodiment of the present invention; FIG. 5 is a flowchart of a method for recognizing audio by a client computer according to another embodiment of the present invention; FIG. 7 is a flowchart of a method for identifying audio of a radio broadcast according to a specific embodiment of the present invention by performing recognition by a portable server; FIG. 8 is A flowchart of another method for identifying radio broadcast audio performed by a client computer according to another specific embodiment of the present invention; and -32- 200307874 FIG. 9 is a diagram according to the present invention Another embodiment is a flowchart of another method for identifying wireless radio broadcast audio by the portable device. Similar reference numbers in the figures indicate opposing parts. Explanation of Symbols for Graphical Representation 100 Identification Audio System 102 ID Server 104 Network 106 Client Computer 108 Communication Chain 110 Communication Chain 112 Portable Device 114 Playlist Supplier 116 Fingerprint Supplier 202 Central Processing Unit 204 Communication Circuit System 206 User Interface device 208 Bus 210 Memory 212 Operating system 214 Communication program 216 Fingerprint identification program 218 'Search program 220 Fingerprint identification database 222 Fingerprint
-33- 200307874 224 226 228 230 232 234 236 238 302 304 306 308 310 312 314 316 318 320 322 324 326 328, 330 無線電播放表資料庫 頻率 播放表 曰期 時間 地理識別符號 無識別信息 快取記憶體 中央處理單元 通信電路系統 匯流排 使用者介面裝置 記憶體 操作系統 通信程序 指紋識別程序 搜尋程序 指紋識別資料庫 指紋 無線電播放表資料庫 頻率 播放表 曰期 時間 332 地理識別符號 地琿ID程序 無識別信息 I己錄程序 播放機程序 無線電程序 快取記憶體 音頻樣本 時脈/頻率資料 音頻播放機 傳聲器 無線電接收器 天線 頻率偵測程序 時脈 -35--33- 200307874 224 226 228 230 232 234 236 238 302 304 306 308 310 312 314 316 318 320 322 324 326 328, 330 Processing unit communication circuit system bus user interface device memory operating system communication program fingerprint identification program search program fingerprint identification database fingerprint radio playlist database frequency playlist date time 332 geographic identification symbol location ID program no identification information I Recorded program player program radio program cache memory audio sample clock / frequency data audio player microphone radio receiver antenna frequency detection program clock -35-