200845644 九、發明說明： 【發明所屬之技術領域】 網二:及經由面向分組的 5【先前技術】 (VoIP):面：二組的網路的即時通信中，例如IP語音 内到達2機所有的分組均在所需要的時間 '經由網際網:)二tr當以面向分組的方式(例如 路的不同路_且^ 同的分組採取經由資料網 的發射機八ft同的分組可以採用的、從該分組 於當前的網路;接收機的、經由資料網路的路徑取決 C, 序也报有組的順序來發送分組’到達接收機的順 15上，，較早日士 ^恭發現了有利路徑的分組甚至可能“追 的分組。守3★运但卻採取從發射機至接收機的較長路徑 將簡衝器在分組序列中的所有分組到達之前 允許相對大的==要巧者 之前以正料=對===在㈣分㈣傳遞分組 然而，選擇的緩衝器越小，或者 λ 帽輸時可-有的可允許延輪==: 5 200845644 在分組確實6錢時會導致分組錢，而以八 失、—機到達接收機需要過長的時間時也 二 4 土 ί貝枓損壞時，即分組實際上產 這種對延遲要求苛 〜曰外。 中出現，在這種應用中，Γ了=.:路中電:(， 至接收機的分組必須滿足的延遲需求在 k發射機 ίο 15 :佈it:、、:在發射機側佈置了音頻編‘並 時，也即並且在接收機處不再有要編碼的資料 潰”時中解;1器由於缺少資料呈現而：‘崩 【發明内容】 月的目的是通過一種依據申請專利範 1 用9=:賁料分組序列的設備、-種依據申請專利範圍第 J、、於傳輸資料分組序列的方法、一種依據申& 於對資料分組序列進行心 碼的;3=;::用_組序列進行解 來實現的。/ 種依據申叫專利範圍第25項的電腦程式 聽見ί=免解碼器側的崩潰和/或避免解碼器側的用戶可 "、’、、、，根據本發明，執行對分組序列中的分組是否 6 20 200845644 丟失或錯誤的檢測。如果確定了這種分組絲或分組差 錯則將提供替代分組，該替代分組是關於分組語法的有 ^刀且：、、丨而，其中該替代分組的音頻内容具有特定的内 ^特性。根據本㈣_於傳輸㈣分組相的設備輸出 k亂的刀、、且序列’然而’其中錯誤分組或未接收到的分 ，已被替代分組所替代，因此該傳輸設備所的輸出序列包 ίο 15 ti 20 中^接收到的分組以及一個替代分組。在一個實現 ’替代分組巾的分組或義内容和/或内容特性獨立於音 ㈣取決於在先或後續的分組或巾貞。然而，如果 /土二，取差錯_藏措施’則合成的音頻内容將取決於 先或後續的巾貞，即將科為預定的·立於信號。、 是麸Π亥替2分組在有效載荷區域中提供了對該分組 由：太二、且Ϊ事實的指示，其中該指示是可忽略的或者可 有i八=器來解釋的，以使得根據預定的内容特性如同 可:二::ί=該替代分組進行解碼，並且其中該指示 具有用釋，該擴展解碼器與該基本解碼器相比 功能，替代分組的内容的差錯隱藏措施的擴展 性料2具有與該預定的内容特性不同的内容特 效载;以是純有效载荷分組或者可以购 分，^優選地存在於有效載荷抑^不存在於報頭部 資料ΙΓ分組序列的接收機接收未擾il的資料分組流，該 對資料語法。該接收機將能夠容易地 刀組序列進行解碼。在該接收機作為基本接收機 7 200845644200845644 IX, invention description: [Technical field of invention] Network 2: and via packet-oriented 5 [previous technology] (VoIP): face: two groups of networks in instant communication, such as IP voice within 2 machines The packets are all at the required time 'via the Internet:), when the packet is used in a packet-oriented manner (for example, the different paths of the roads and the same packets are taken by the same packet of the transmitter of the data network, From the group to the current network; the path of the receiver via the data network depends on C, and the sequence also reports the order of the group to send the packet 'on the receiver's -15, and earlier, the priest found a favorable The grouping of paths may even be "chasing the packet. But the longer path from the transmitter to the receiver takes the simpler to allow a relatively large number of packets before the packet sequence arrives. Exactly ===== In the (fourth) minute (four) pass the packet However, the smaller the selected buffer, or the λ cap can be used - some allow the extension wheel ==: 5 200845644 When the group does 6 money, it will lead to grouping money And with eight lost, the machine arrives When the machine takes too long time, it is also 2 when the soil is damaged, that is, the group actually produces such a delay to the demanding delay. In the application, in this application, Γ =.: Road power: (, the delay to the receiver must meet the delay requirement in the k transmitter ίο 15 : cloth it:,,: when the audio side is arranged on the transmitter side, and that is, there is no more coding at the receiver The data collapses in the mid-solution; the 1 device is lacking in the presentation of the data: 'Break the contents of the invention】 The purpose of the month is to use a device based on the application of patent specification 1 9:: packet sequence, based on the scope of the patent application J, in the method of transmitting the data packet sequence, one based on the application of the heart code of the data packet sequence; 3 =;:: using the _ group sequence to solve the solution. / According to the patent application scope 25 The computer program of the item hears ί=free decoder side crash and/or avoids the decoder side user can", ',,, according to the invention, whether the packet in the packet sequence is lost or wrong 6 20 200845644 Detection. If such a grouping or grouping is determined Wrong will provide an alternative packet, which is related to the packet syntax, and the audio content of the substitute packet has a specific internal characteristic. According to the device (4) Outputting a chaotic knife, and the sequence 'however' of the erroneously grouped or unreceived points has been replaced by the substitute packet, so the output sequence of the transmission device is ί 15 15 ti 20 Substitute grouping. In one implementation, the grouping or meaning content and/or content characteristics of the alternative packet towel are independent of the sound (4) depending on the prior or subsequent grouping or frame. However, if / soil 2, take the error _ hiding measures' The synthesized audio content will depend on the first or subsequent frame, which is the predetermined signal. , the bran happine 2 packet provides an indication of the fact that the packet is: too two, and Ϊ facts in the payload area, wherein the indication is negligible or may be interpreted by an i == device, such that The predetermined content characteristics are as follows: two:: ί=the substitute packet is decoded, and wherein the indication has a release, the extended decoder functions as the base decoder, and the expansion of the error concealment measure of the content of the packet is replaced. Material 2 has a content payload that is different from the predetermined content characteristics; it is a pure payload packet or can be purchased, and preferably exists in the payload of the receiver that does not exist in the header data packet sequence. The data packet stream of the disturbance il, the pair of data syntax. The receiver will be able to easily decode the block sequence. At the receiver as a basic receiver 7 200845644
ίοΊο
時’該接收機㈣此容㈣對替代分—行料並再現預 定的音頻内容u，由於該音_容是敢的，因此這 將導致品質損失，並因此將不能很好地適合在先的分組或 幀以及後續的分組或幀。然而，與由於不存在分組以及因 此中斷了整個通信連接而導致的解碼器完全崩潰的情況相 比，這種短時品質損失並不成問題。 相反地，擴展解碼器能夠使用替代分組中的、指出了 該分組是替代分組而非普通分組的指示來識別上^趣， ，不是簡單地處理替代分組’而是在接收到替代分二夺 發起差錯隱藏措施。 二除幀/分組重複之外，差錯隱藏措施示例性地包括已存 於記憶體巾的先前分組或後較組在先前分組和後續分組 之間的外插。這種外插或_具體包括在差錯隱藏情況中 對音頻信號的短時頻譜進行分頻段能量測量和合成，該音 頻乜號的頻譜值是以大概隨機的方式產生的，然而，其中 該，機產生的音頻信號的分頻段能量取決於已=通過差錯 隱藏措施而產生的、先前和/或後續的正確接收 多分組的能量。 在一個實施例中，替代分組所具有的預定内容特性是 零頻，。這裏的結果是對就的替代分組進行解碼的基本 解碼益執行“雜訊抑制”。備選地，該内容特性可以是其 音頻内容和/或頻譜值與絕對聽音臨界值相關的音頻信 亚且是以該音頻内容小於該絕對聽音臨界值的兩倍的方式 來示例性地定義的，因此，在所有頻段中包括特^但少量 8 20 200845644 ，錶讯，該雜訊在特定情況下可能比簡單的“雜訊抑制” 在主觀上聽起來更好。 【實施方式】 :面將參照附圖詳細地描述本發明的實施例。 固示出了一種用於傳輸表示音頻信號的資料分組 備包2 在第—圖中示例性地實現為基站1G的傳輸設 對分列中的分組的裝置11，分組語法是針 ㈣列」 匕外，基站10包括用於檢測分 接征了、刀組是否丟失或發生錯誤的裳置13。此外，還 =於，分組的褒置14，以替代錯誤分組卜或： 失的刀組。替代分組是關於分五 15 20 代分組的音頻内容具有預定的内容m ^組J而替 括用於輸出分組序列的裝置】5，該分 括:: 接收的分組以及-個替代分組二们 置15是搞合至天線16的用基站的不例’輸出裝 如NG聰規範）來將資料分組而範（例 二圖進行更詳細討論的移動單-序歹J傳輪至下面將參照第 第-圖中的裝置二17 15，以將接收到的普通（也、、、 耦合至輸出裝置 置15。此外，該用於提供替代二*錯的分組準時傳輸至裝 線路18與輸出裝置15相連。2的裝置14經由替代分組 的裝置14包括在其中儲存蚨地，用於提供替代分組 曰代/刀組的記憶體。只要識別到 9 200845644 了丟失的或錯誤的分組，裝置14都啟用記憶體訪問以從該 記憶體中取回替代分組，並經由線路18將該替代分組饋入 該輸出裝置15。 ' ίο 15 Ο 20 在一個實施例中，控制檢測裝置13以檢測分組丟失， 並在超過最大延遲的時間内未接收到分組序列中的分組 ，’啟用用於提供替代分組㈣置。在—個實施例中，該 最大延遲示例性地可經由控制線路19來控制。可以經由控 制線路丨9將在VoIP應用中示例性地包括最大延遲的⑽ (服務品質）請求饋人該檢測農置13。對於其他面向分組 的應用中，除VoIP之外’還可能存在經由控制線路19饋 入該檢測|置13的不同QoS請求。備選地，該檢測裝置 13還可以具有岐地設置的準則，根據該準則來檢測分組 差錯或分組丢失並啟用用於提供替代分組的裳置Μ。 署個實施例中，經由#代分組線路18提供給輸出裝 、、日代分組不僅具有預定的内容雜，且有對該 與二 ::::代分組產生與--容特二: 分了可示例性地位於移動單元2㈠的用於對 語法針對該解碼11包細於接收分組 I、私疋的分組序列的接收機21，該序列包括至少 10 200845644 個舅料分組 該分組語法的有 夕—個替代分組，該替代分組是關於 預定的内容护性"。刀奴，並且該替代分组的音頻内容包括 分組的事實的於_匕外，该替代分組包括對該分組是替代 碼器來解釋钟日Γ ^ ’只能通過第二51所示的擴展編 由檢測iT2ri而不能通過基本解碼器來解釋該指示。 被實現為檢該替代指示的解釋，該檢測器22 二圖中的擴展料I，含該指示並因此是替代分組。第 裝置如㈣包括差錯隱藏裝置Μ，該差錯隱藏 ίο 15 t 同的内容含在#代分財的敢内容特性不 碼器二=:ί:Γ容進行合成。此外，該擴展解 並非_八^衣124,該音頻呈現裝D4祕在分組 是替：分：ί日:呈現分組中的音頻内容，以及用於在分組 輸出單元25=㈣合成音頻内容。該音頻呈現裝置24與 換哭放大^ 該輸出單元25示例性地包括D/A轉 換^7、放大态和揚聲器。 路26具 音頻呈現裝置經由傳輪常規分組的分組線 秩代21相麵合。此外，該音頻呈現裝置24經由 與差錯隱藏裝置23 4目連，通過該替代分 呈_容從差錯隱藏裝置23傳輸至音頻 單置23可以以不同的方式來執行差錯隱藏。 减方法是簡單地重複先前的.貞和/或先前分組 %為幅h °備選地’可以實施該差錯隱藏方法以執 20 200845644 行外插或内插。兮々 執行。在頻譜值;;二=可以相對於頻譜顺 與先前巾貞相等的頻车可以基於-個或多個具有 段中示例性地由缓來執行差錯隱藏’在該多個頻 的方式產生為缝發生11產生或以大概確定When the receiver (4) accommodates (4) the alternative sub-line and reproduces the predetermined audio content u, since the tone is dare, this will result in loss of quality and therefore will not be well suited to the prior A packet or frame and subsequent packets or frames. However, such short-term quality loss is not a problem compared to the case where the decoder is completely collapsed due to the absence of packets and thus the entire communication connection is interrupted. Conversely, the extended decoder can use the indication in the alternate packet that indicates that the packet is a substitute packet instead of a normal packet, instead of simply processing the alternate packet' instead of receiving the substitute split two. Error concealment measures. In addition to the frame/packet repetition, the error concealment measure illustratively includes the extrapolation of the previous packet or the subsequent group that has been stored in the memory towel between the previous packet and the subsequent packet. Such extrapolation or _ specifically includes sub-band energy measurement and synthesis of the short-time spectrum of the audio signal in the case of error concealment, the spectral value of the audio nickname is generated in a roughly random manner, however, where The sub-band energy of the generated audio signal depends on the energy of the previous and/or subsequent correct reception of the multi-packet that has been generated by the error concealment measure. In one embodiment, the substitute packet has a predetermined content characteristic of zero frequency. The result here is to perform "noise suppression" on the basic decoding benefit of decoding the alternate packet. Alternatively, the content characteristic may be an audio message whose audio content and/or spectral value is related to an absolute listening threshold and is exemplarily in a manner that the audio content is less than twice the absolute listening threshold Defined, therefore, in all frequency bands including special ^ but a small amount of 8 20 200845644, the message, in some cases may be subjectively sound better than the simple "noise suppression". [Embodiment]: Embodiments of the present invention will be described in detail with reference to the accompanying drawings. It is shown that a data packet backup packet 2 for transmitting an audio signal is exemplarily implemented as a device 11 in a transmission pair of columns of the base station 1G, and the packet syntax is a pin (four) column. In addition, the base station 10 includes a skirt 13 for detecting a tap, whether the knife group is missing, or an error has occurred. In addition, = is, the grouping of the device 14 to replace the wrong group or: the lost knife group. The alternative grouping is that the audio content of the sub-five 15 15 generation grouping has a predetermined content m^group J and the means for outputting the sequence of packets] 5, the sub-division:: the received packet and the -substitute packet 15 is a combination of the base station of the antenna 16 (the output is installed as NG Cong specification) to group the data (the second figure is discussed in more detail in the mobile single-sequence J transmission wheel to refer to the first - means 2 17 15 in the figure to receive the received normal (also, coupled to the output device 15). In addition, the packet for providing an alternate second error is transmitted to the line 18 and connected to the output device 15 on time. The device 14 of 2 replaces the packet via the device 14 for storing the memory for providing a substitute packet/knife set. As long as 9 200845644 is identified as a missing or erroneous packet, the device 14 enables the memory. The body accesses to retrieve the replacement packet from the memory and feeds the replacement packet to the output device 15 via line 18. ' ίο 15 Ο 20 In one embodiment, the detection device 13 is controlled to detect packet loss, and More than the most The packet in the sequence of packets is not received within the delay time, 'Enable to provide a substitute packet (4). In an embodiment, the maximum delay is exemplarily controllable via control line 19. It can be via control line 丨9 The (10) (Quality of Service) request, which typically includes the maximum delay in the VoIP application, is fed to the detection farm 13. For other packet-oriented applications, there may be a detection of this detection via the control line 19 in addition to VoIP. The different QoS requests are set 13. Alternatively, the detecting means 13 may also have a set criterion for detecting packet errors or packet losses according to the criteria and enabling the presence of the alternative packets. In the example, the output packet is provided to the output device via the # generation packet line 18, and the day generation packet not only has a predetermined content, but also has a pair of the two:::: generation generations. a receiver 21 located at the mobile unit 2 (1) for arranging a packet sequence that is finer than the received packet I, private for the decoding 11 packet, the sequence comprising at least 10 200845644 packets of the packet The eve of the law is an alternative grouping, which is about the predetermined content protection " knife slave, and the audio content of the substitute packet includes the fact of the grouping, the substitute grouping includes Instead of the coder to explain the clock Γ ^ ' can only be detected by the extension shown in the second 51 by the detection of iT2ri and can not be explained by the basic decoder. Implemented as an explanation of the replacement indication, the detector 22 The extension I in the figure contains the indication and is therefore a substitute group. The device as (4) includes an error concealment device, and the error is hidden ίο 15 t The same content is included in the #代财财的敢性性不码器二= : ί: The content is synthesized. In addition, the extended solution is not _ 八 衣 124, the audio presentation D4 secret grouping is: :: the presentation of the audio content in the packet, and for synthesizing the audio content in the packet output unit 25 = (4). The audio presentation device 24 and the switching output unit 25 illustratively include a D/A conversion, an amplification state, and a speaker. The path 26 has an audio presentation device that is aligned via a packet line rank generation 21 that is conventionally grouped by a pass. Furthermore, the audio rendering device 24 is connected to the error concealing device 23, and the error concealment can be performed in a different manner by the alternative presentation/transmission from the error concealing device 23 to the audio unit 23. The subtraction method simply repeats the previous .贞 and/or the previous packet % is the frame h ° alternatively. The error concealment method can be implemented to perform 20 200845644 line extrapolation or interpolation.兮々 Execute. In the spectral value;; two = can be equal to the spectrum with the previous frame, the frequency can be based on - or more with segments in the example of the implementation of error concealment in the way of the multiple frequency generated as a seam 11 occurrences occur or are roughly determined
C ίο 15 20 權以使得該頻譜值：m,然後對該頻譜值進行加 量源自-個或多個^二⑨里等於目標能量，該目標能 先前和/或後續的鳩可:或一個或多個後續的悄。該 發生分組錢並且 =f的接收分組，或者可以是在 組裝置所產生=::丢失多個連續分 器是基於=::===列r頻編碼 包括時頻轉換級30，通 土；又、的日頻編碼器 換為短時頻譜序列d a、級3g將時域音頻信號轉 器3i，量化器31由^^物請人執行量化的量化 化_量化雜t心主量 二=’视_33 二^:二:： 、馬益。_曰編碼器提供了與由量化器31以比例因數的形 =例it地提供的、並由熵編碼器33以所使㈣編碼表的 形式示例性地提供_助資訊有_位元㈣，並且該位 疋序列形成了要提供給在輪出側輸出f料分組序列的分組 打包器34的貧料。除了該分組打包器之外，如同示例性地 在關鍵字 MP3 ( MPEG-1 層 3 )或 AAC ( MpEG_4 )或 ac_3 12 200845644 於變換:二ί 7:::出示的::編碼器表㈣ 現為針對每個音頻日^•需求將分組打包器34實 者產生多於—_音產生—個分組，或 入單個分組中。曰A、卩，將多個編碼的短時頻譜弓| 第四圖示出了第二圖中立 表示，具體地是示出了音頻呈王現裝置24的更詳細的 的協作。在輸入側，音頻呈衣置1與差錯隱藏裝置23 ίο 15 Ο 20 將中分離出辅助資訊的分組拆包器40。 碼器。===短時頻譜)饋入網編 反量化哭42 ^仪了里化索引，將該量化索引饋入 的頻譜二然=4:::=提供量化後又反量化 生輸出音頻信號。肺中 轉換之後使㈣頻譜值產 助資訊來控制^炳^益41和反量化器42均可通過辅 量化器42接魏接收碼_，而該反 然後在分㈣h執麵反量化的比例因數。 在的檢測特性時t器4〇包括如第二圖的檢測器22令存 送替代分組指示，^组拆包器40能夠向隱藏裳置22發 組並非第三圖中的；1得該隱藏裝置23可以識別當前的分 代分組。在這種情二1器產生的分組，而是基站產生的替 元序列，或者向反〜μ ’该隱藏裝置將向綱解碼器提供位 轉換裝置提供頻譜^器提供量化索料列，或者向頻時 内容饋入解碼器功乾^列，以便在任何位置處將合成音頻 力_。優選地，在該鏈的束 13 200845644 頻時轉換級43處）饋入合成的頻譜音頻内容。 、，這些頻譜内容優選地取決於已正確接收的先前的頻 譜，或者取決於後續頻譜，該後續頻譜可能已存在，並且 可能包括與頻譜值、頻段或二者（即與頻譜值和頻段 5關的差錯隱藏裝置，以用於合成該音頻内容。 如同使用第-圖和第二圖已示出的，在基站和移動單 凡中利用本發明的實現，其中假設基站可以以不正確的順 序來接收分組，例如當基站與網際網路相耦合時，而移動 單元取決於以正確的順序來接收分組序列。例如，通 10 DECT標準定義基站與移動單元之_這種通信連接。 在經由面向分組的網路的即時通信（例如ν〇ιρ)中， 不能保證所有的分組都在所需要的時間内到達接收機。在 特定的（非常有限的）時間之後，必須將尚未到達的分板 歸類為去失的分組。在IP分組錢的情況下，產生替代立 15頻巾貞並通過基站進行傳輸。 /這種替代將由下-代（NG) DECT系統中的基站來執 行。從基站向移動單元傳輸該替代的幢，以代替原始的（而 非接收到的）悄。NG DECT規範特別地通過寬頻和超寬頻 音頻編解碼器和IP終端來表示當前的DECT規範的擴展 2〇 NG DECT站包括一個或多個從NG DECT基站接收呼 叫的热線電話。因此，可以使得v〇Ip呼叫直接經過如 DECT電話。在理想情況下，可以將ν〇ιρ語音分組從基站 傳輸至移動單元，而無需在基站中進行再次編石馬。 幻汝^使用其常規语法未提供用信號通知rjj貞丟失的 200845644 特殊方式的音頻編解碼哭 頻謝可由符合標準:解碼器進:::組:二種”音 例中，應當同日夺向擴展的 Ψ肖但在一個貫施 中m站卩自卩捭成的，以使得基站可以如同純 =ΓΠ貧料傳輸而不再進行解碼和編碼的站）- ίο 15 =)二基站應當僅執行非常少量的分組拆包(如 中的何處，因 容進行正確解碼對由°亥貝科分組序列所表示的音頻内 專用二的广用特定的用戶 提供信號通知===馬器(即擴展編碼器） 元流語法標準尚未提供這種信號=代巾貞’即使底層的位 載荷部分時，其中該有效 何都不會考慮 差錯因此優親將削旨示容納在該有效載荷部分中: 二刀組僅具有有效載荷部分而不具有報頭時，僅將該 :不二口，荷:分:從而能夠實現本實施例。優選 Λ S不各、，内在音頻貧料和/或音頻資料部分中。 20 200845644 個貫現中，在未提供用信號通知㈣失的顯式方 =丨—#代音頻巾貞和/或替代分組滿足該替 ，替代》組可由符合標準的料器進行解碼㈣—準則。 繼向擴展解碼器提供明確識別繼 *曰代巾貞的方式’以使得該擴展解㈣可以啟用差錯障藏。 ίο 15 八，果替代巾貞是根據駭的縣化的㈣流程語法和/或 法或龄法的有賴，則符合標準的傳統的解碼器 將月匕夠對該替代t貞進行解碼。在—個實現中，優選地通過 不具有音_容的巾貞來替代錢的巾貞，即，執行所謂的雜 鱗制。雜訊抑㈣味著整侧譜被設置為零。具體地， 通過，用AAC標準（MPEG.4·音頻），該變化優選地用於 產生令頻④，在零頻譜巾將用於傳輸頻譜值的最高的比例 因數頻段設置為零（麵」㈣）。備選地，還可以傳輸針 對，次為零的比例因數頻段的頻譜值。這原理上可以使用 所提供的不同碼本中的任何可用的Huffman碼本來實現， ，中當使用碼本“ZER0_HCB”（〜。祕咖碼表)時， 絲員然不必傳輸這些譜線。 注空f當指出，檢測錯誤幀或丟失幀的基站可以已執行差 二^藏扣鈿。然而，根據本發明，由於差錯隱藏措施在計 2面較為複雜，並且額外地需要估計先前的（以及可能 二的）9頻k號’因此基站優選地不執行這種差錯隱 曰^具體地，當DECT基站為多個移動單元“提供服 ，為了能夠執行複雜的差錯隱藏，其結果將是基站 人對所有音頻内容連續地進行解碼和編碼。除了對於較 16 200845644 高的處理器和記憶體資源的相關需求之外，特別是在使用 有損編碼器時’結果將是由於串列編碼效應而導致的附加 的品質惡化。此外，延遲將大大增加。 、C ίο 15 20 The weight is such that the spectral value: m, and then the amount of the spectral value is derived from - or more ^ 2 9 equals the target energy, the target can be previous and / or subsequent: or one Or multiple follow-ups. The packet that generates the packet money and =f, or may be generated by the group device =:: loss of a plurality of consecutive classifiers is based on =::=== column r-frequency coding including time-frequency conversion stage 30, commonplace; Moreover, the day frequency encoder is replaced by the short-time spectrum sequence da, the level 3g is the time-domain audio signal converter 3i, and the quantizer 31 is quantized by the object-to-person quantization _quantization miscellaneous t-heart quantity two=' _33 2^: 2::, Ma Yi. The _ 曰 encoder is provided with the form of the scale factor by the quantizer 31 and is exemplarily provided by the entropy coder 33 in the form of the (four) code table. _ bit information (4), And the bit sequence forms a poor material to be supplied to the packet packer 34 that outputs the f packet sequence on the wheel side. In addition to the packet packer, as exemplarily shown in the keyword MP3 (MPEG-1 Layer 3) or AAC (MpEG_4) or ac_3 12 200845644: 2: 7::: Presented by: Encoder Table (4) The packet packer 34 is generated to generate more than -_tones for each audio day, or into a single packet.曰A, 卩, a plurality of coded short-time spectrum bows | The fourth figure shows a second diagram neutral representation, specifically showing a more detailed collaboration of the audio presentation device 24. On the input side, the audio dressing unit 1 and the error concealing device 23 ίο 15 Ο 20 separate the packet unpacker 40 for the auxiliary information. Code. ===Short-time spectrum) Feeding the network code The anti-quantization crying 42 ^ instrument has the inner index, the spectrum fed by the quantization index is second = 4:::= provides quantization and then inverse quantizes the output audio signal. After the conversion in the lung, (4) the spectral value production information is controlled to control both the Bingyiyi 41 and the inverse quantizer 42 by the auxiliary quantizer 42 to receive the Wei received code_, and the inverse then divides the proportional factor of the inverse quantization in the sub (4)h face . In the detection characteristic, the device 4 includes a detector 22 as shown in the second figure to store the replacement group indication, and the group unpacker 40 can send the group to the hidden skirt 22, which is not in the third figure; Device 23 can identify the current generation of packets. In this case, the packet generated by the device, but the base sequence generated by the base station, or the reverse ~μ' the hidden device will provide the bit decoder with the spectrum converter to provide the quantized data column, or The time-frequency content is fed into the decoder function column to synthesize the audio force _ at any position. Preferably, the synthesized spectral audio content is fed at the frequency band 13 200845644 at the frequency conversion stage 43). The spectral content preferably depends on the previous spectrum that has been correctly received, or on the subsequent spectrum, which may already exist and may include the spectral value, the frequency band or both (ie, with the spectral value and the frequency band 5) Error concealing means for synthesizing the audio content. As has been shown using the first and second figures, the implementation of the invention is utilized in base stations and mobile units, where it is assumed that the base stations can be in an incorrect order Receiving packets, for example when the base station is coupled to the Internet, and the mobile unit depends on receiving the sequence of packets in the correct order. For example, the communication protocol for the base station and the mobile unit is defined by the 10 DECT standard. In instant communication on the network (eg ν〇ιρ), there is no guarantee that all packets will arrive at the receiver within the required time. After a certain (very limited) time, the sub-boards that have not yet arrived must be classified. For the lost packet. In the case of IP packet money, an alternative 15 tone frame is generated and transmitted through the base station. / This replacement will be performed by the next generation. (NG) The base station in the DECT system performs the transmission. The replacement block is transmitted from the base station to the mobile unit instead of the original (not received). The NG DECT specification specifically passes the wideband and ultra-wideband audio codecs and The IP terminal to indicate the extension of the current DECT specification. The 2 NG NG DECT station includes one or more hotline calls that receive calls from the NG DECT base station. Therefore, the V〇Ip call can be made to pass directly through, for example, a DECT phone. Ideally, The ν〇ιρ voice packet can be transmitted from the base station to the mobile unit without having to re-arrange it in the base station. The illusion is not provided by the conventional syntax without signaling the rjj贞 lost 200845644 special mode audio codec crying frequency Xie can conform to the standard: decoder into::: group: two kinds of sound examples, should be the same day to capture the expansion of the ambiguity but in a consistent implementation of the m station, so that the base station can be as pure = Stations that transmit poorly and are no longer decoded and encoded) - ίο 15 =) The two base stations should only perform a very small number of packet unpacking (such as where, due to the correct decoding of the pair by Hibec The group of sequences represented by the audio is dedicated to the specific user-provided signal === horse (ie extended encoder). The stream stream grammar standard has not yet provided such a signal = 代 贞 ' even if the underlying bit load part , where the validity is not considered to be an error, so the superior is to be accommodated in the payload portion: When the second knife group only has the payload portion and does not have the header, only the: the two, the minute: This enables the present embodiment to be implemented. Preferably, Λ S is not in each, in the inner audio poor material and/or audio data portion. 20 200845644 In the past, in the absence of the signalling (4) missing explicit party = 丨 -# The sub-audio frames and/or the alternative groupings satisfy the replacement, and the alternative group can be decoded by the standard-compliant hopper (four)-criteria. The method of providing explicit identification to the extended decoder is then provided to enable the extended solution (4) to enable error blocking. Ίο 15 VIII, the fruit replacement 贞 is based on the 县 county's (four) process grammar and / or law or age law depends on the standard, the traditional decoder will be enough to decode the substitute t 匕. In an implementation, the towel of the money is preferably replaced by a frame having no sound_capacity, i.e., a so-called scaly system is performed. The noise (4) tastes that the whole spectrum is set to zero. Specifically, by using the AAC standard (MPEG.4 Audio), the change is preferably used to generate the frequency 4, and the highest scale factor band used to transmit the spectral value is set to zero (face) in the zero spectrum towel (4) ). Alternatively, it is also possible to transmit the spectral values of the scale factor bands of the second, second order. This can in principle be implemented using any of the available Huffman codebooks in the different codebooks provided, and when using the codebook "ZER0_HCB" (~. crypto code table), the wire does not have to transmit these lines. Note that when indicating that the error frame or the lost frame is detected, the base station that has detected the error frame or the lost frame may have performed the difference. However, according to the present invention, since the error concealment measure is complicated in the face 2, and additionally it is necessary to estimate the previous (and possibly two) 9-band k number ', the base station preferably does not perform such error concealment ^ specifically, When the DECT base station "serves" for multiple mobile units, in order to be able to perform complex error concealment, the result will be that the base station person continuously decodes and encodes all audio content. In addition to the higher processor and memory resources for 16 200845644 In addition to the related requirements, especially when using lossy encoders, the result will be an additional quality degradation due to the tandem coding effect. In addition, the delay will be greatly increased.
ίοΊο
由於在AAC標準中未提供用信號通知幀丟失的顯式方 式，、在一個實現中，使用常規的解碼器所忽略的信號^知 方式。這裏保持資料語法和/或分組語法。另一方面，由於 這種類型_也可以在常規操作中出現（例如在解碼器= 入端處沒有信號時），因此簡單地將幀的頻譜設置為愛不足 以提供關於這是替代幀和/或替代分組的事實的安全的抨 對於幀是替代幀和/或替代分組的事實的指示向擴展铲 碼器提供了以下資訊：當前巾貞並非頻譜實際為零的幢，: 是由於傳輸差錯而被引人職財以避免解碼器故障和 語音鍵路故障的巾貞。 一 音頻編碼標準典型地提供了允許附加有效載荷 用戶專用資料區域’然而，其中該有效難被傳統的解 器（即不具有擴展功能的基本解碼器）所忽略。在aa = 準中，如第六圖A中狀義的，這種用戶^義的有效载矿; 是所謂的“extension—payload” 。如第六圖B所示，根】 二extenSi〇n_type”變數的值，該標準提供了不同的目^。 第六圖A和第六圖B摘錄自標準IS〇/IEC14496_3 : 2^°5 (E)。在該標準中出於以下目的而提供了其中所述的 單元（FILL)的使用。當針_有音頻:#_及所有附 資料的位元總數小於該巾貞中用於實現目標位元率的所允； 20 200845644 的最小位元數時，必須將填充單元添加至位元流中。當編 碼器想要引入這種DRC資訊時，添加動態範圍控制位元 _(drc位元）。如同標準所述，在正常情況下，避免填充位 兀並使用自由位元（free bit)來填滿位元儲存和/或位元保 5存庫。只有在位元儲存已滿時，才寫入填充位元。允許任 何數目的填充位元。 在本發明的一個實現中，如第七圖在7〇處所示，將 “extension—type”設置為“〇〇〇〇” ，以便與填充位元的標 準使用相反地將對巾貞丢失的指示寫人“Qthei<—bits”搁位。 1〇 在用於填充位元的標準中所提供的用戶專用資料區域 用於提供對替代幀的信號通知，即用於容納對於替代幀的 指示。然而，根據實現，可以根據“extenskm—type”值的 不同设置來使用其他“extensi〇n—payl〇ad ()，，。由於優選 地（通過將max」fb設置為零）有效地傳輸零頻譜或^使 I5用Zero Huffman碼本，因此存在可用於各個 extension_payload ()的足夠的位元。 應當指出’在71處示出了針對替代分組的典型的遵從 MPEG-4的資料流程和/或分組語法和/或龄法，其中，如 72處所示’使用雙數max—sfb=〇，，。優選地，還將獲得有 2〇效分組語法所需要的所有其他資料設置為零。然而，應當 指出，該資料自身並非對於替代巾貞的可靠的指示。由於正 常的編碼器不會寫入具有零頻譜的巾貞而是寫入特定的 extension—payl〇ad，因此只有 extensi〇n』ay_ 7〇 將產生 可靠的指示。 18 200845644 第八圖示出了針對採樣速率為48kHZ、單聲道信號、 位元速率為64kBit/s的新一代DECT的示例性替代幀。應 當指出’由於位元流中的eXtenSi〇n-payl〇a(J並非“按位元 組對背” ’因此在第八圖中不容易看出如第七圖在7〇處所 5 示的 extension—payload ° 此外，應當指出，“extension_payload”中的位元組合 包括也被稱為“差錯模式”的位元模式，該位元模式與 “FRAME一LOSS”的ASCII碼相對應。通過該位元模式來 保證不產生與extension—payload的其他用戶的衝突，這是 10因為另外的用戶幾乎不可能使用“FRAME—LOSS”的 ASCII碼來用信號通知與“FRAME-L〇ss，，無關的事。 下面將芩照第五圖來討論可能在從分組的發射機至解 碼器和/或至解碼器中的音頻呈現裝置的傳輸場景中的不同 點處出現的分組或幀的不同順序。 15 第五圖A示出了編號為（i-D、i、（i+1)、（i+2)的分 組的序列。這種正確的分組或幀序列在面向分組的傳輸網 路（例如網際網路）的輸出端處出現。 第五圖B示出了在基站的輸入端處的分組序列，其中 可以看出，在第五圖B所考慮的時域部分中分組丨尚未到 20達，但是分組順序已經變得混淆。其原因在於，分組丨已經 凡全丟失或者獲得從發射機至接收機的非常長的路徑。另 一方面，分組（i+2)已經獲得了非常有利的路徑，因此該 分組在從發射機（即編碼器）至基站輸入端處的途中“追 上 了分組（i+i )。 200845644 在第五圖B中，第-圖中的接收裝置u將在分 時以正確的順序再次對分組進行分類。此外，解碼裝置U 將發現具有編號i的分組尚未出現或者發生錯誤。因衣此，如 第五圖C所示，將產生針對編號丨的替代分組。因此第五 圖c示出了第一圖中的輸出装置15所輸出的分組 從根據第-圖的基站至根據第二_移動單元的路徑上， 分組順序沒有改變。然而’如第二圖所示，擴展解石: ίο 15 識別編號為i的分組是較分組。與第五圖d所^生 音頻内容紅常呈關其他分_反地，針料 生返回至差錯隱藏措施的合成音頻内容。 本發明的方法可根據情況以硬體或軟體來實現。以 ，可以基於數位儲存介f，具體是基於具有可被電子地接 ==信號_片或CD ’該磁片或CD可與可編程電: 電月行相應的方法。總體上，本發明還可在 式中貫現’該電腦程式產品包括儲存在機器可 二夂用於在該電腦程式產品在電腦上運行 式碼^ 言之，本發㈣此可以被實現為具有程 =電腦程式’用於在該電腦程式在電腦上運行時:;亍 20 20 200845644 【圖式簡單說明】 第一圖示出了用於傳輸資料分組序列的設備的實現的 電路框圖； 第二圖是用於對資料分組序列進行解碼的解碼器的電 5 路框圖； 第三圖是用於產生資料分組序列的音頻解碼器的電路 框圖， 第四圖特別示出了第二圖的音頻呈現裝置的更特定的 實現； 10 第五圖A示出了編碼器輸出的資料分組序列； 弟五圖B不出了基站接收到的貢料分組序列， 第五圖C示出了由基站輸出並由移動單元接收的、已 插入替代分組的資料分組序列； 第五圖D示出了音頻呈現裝置中所產生的音頻内容的 15 序列； 第六圖 A 示出 了根據 ISO/IEC 14496-3 : 2005 (E) MPEG4 的 extension—payload 的語法； 第六圖B示出了用於例證extensionjype欄位的值的 表； 20 第七圖示出了用於基於變換的音頻編碼器/解碼器的示 例性分組語法； 第八圖示出了包括有效分組語法和預定内容特性的替 代分組的示例。 【主要元件符號說明】 21 200845644 基站ίο 網際網路12 用於提供替代分組的裝置14 天線16 5 替代分組線路18 移動單元20 檢測器22 音頻呈現裝置24 分組線路26 10 時頻轉換級30 心理聲學模型32 分組打包器34 分組拆包器40 反量化器42 接收裝置11 檢測裝置13 輸出裝置15 分組線路17 控制線路19 接收機21 差錯隱藏裝置23 輸出單元25 替代分組線路27 量化器31 熵編碼器33 熵編碼器41 頻時轉換43 22Since an explicit way of signaling frame loss is not provided in the AAC standard, in one implementation, the signal neglected by conventional decoders is used. Keep the data syntax and/or group syntax here. On the other hand, since this type _ can also occur in normal operation (for example, when there is no signal at the decoder = input), simply setting the spectrum of the frame to be insufficient for love provides information about this being a substitute frame and / Or the security of the fact of substituting the packet. The indication of the fact that the frame is a substitute frame and/or a substitute packet provides the following information to the extended code scanner: the current frame is not a block whose spectrum is actually zero, because of a transmission error. Being instructed to avoid decoder failures and voice key failures. An audio coding standard typically provides for allowing additional payloads of user-specific data areas' however, where the validity is difficult to ignore by conventional decoders (i.e., base decoders that do not have extended functionality). In aa = quasi-middle, as shown in Figure 6A, this user's valid payload; is called "extension-payload". As shown in Figure 6B, the roots of the two extenSi〇n_type variables, the standard provides different objectives. Figure 6 and Figure 6B are extracted from the standard IS〇/IEC14496_3: 2^°5 ( E). The use of the unit (FILL) described therein is provided in the standard for the following purposes: When the pin_with audio: #_ and all the attached bits of the data are smaller than the number of bits used in the frame to achieve the goal The bit rate is allowed; 20 The minimum number of bits in 200845644 must be added to the bit stream. When the encoder wants to introduce this DRC information, add the dynamic range control bit_(drc bit) As described in the standard, under normal circumstances, avoid padding bits and use free bits to fill the bit store and/or bit store 5. Only when the bit store is full, Write padding bits. Allow any number of padding bits. In one implementation of the invention, as shown in Figure 7 at 7〇, set "extension_type" to "〇〇〇〇" so that In contrast to the standard use of padding bits, the indication of the loss of the frame is written as "Qthei<-bits "Stop. 1" The user-specific material area provided in the standard for padding bits is used to provide a signal to the substitute frame, ie to accommodate an indication of the substitute frame. However, depending on the implementation, Different settings for the extenskm-type value are used to use the other "extensi〇n-payl〇ad (),,. Since the zero spectrum is efficiently transmitted (by setting max "fb to zero) or the Zero Huffman codebook is used for I5, there are enough bits available for each extension_payload(). It should be noted that a typical MPEG-4 compliant data flow and/or packet grammar and/or ageing method for alternative groupings is shown at 71, where the use of the double number max-sfb=〇, as shown at 72, . Preferably, all other data required to obtain a valid packet syntax is also set to zero. However, it should be noted that the information itself is not a reliable indication of the replacement of the frame. Since a normal encoder does not write a frame with a zero spectrum but writes a specific extension—payl〇ad, only extensi〇n』ay_ 7〇 will produce a reliable indication. 18 200845644 The eighth figure shows an exemplary alternative frame for a new generation DECT with a sampling rate of 48 kHz, a mono signal, and a bit rate of 64 kBit/s. It should be noted that 'because of the eXtenSi〇n-payl〇a in the bit stream (J is not "by bit-to-bit"", it is not easy to see in the eighth figure that the extension shown in Figure 7 is in the 7th place. -payload ° In addition, it should be noted that the bit combination in "extension_payload" includes a bit pattern, also referred to as "error mode", which corresponds to the ASCII code of "FRAME-LOSS". Mode to ensure that no conflicts with other users of extension-payload occur, which is 10 because it is almost impossible for another user to use the "FRAME-LOSS" ASCII code to signal "FRAME-L〇ss, nothing to do." The different sequences of packets or frames that may occur at different points in the transmission scene from the transmitter of the packet to the decoder and/or to the audio rendering device in the decoder will be discussed below with reference to the fifth figure. Figure 5 shows a sequence of packets numbered (iD, i, (i+1), (i+2). This correct packet or sequence of frames is in a packet-oriented transport network (eg internet) Appears at the output. Figure 5 shows A sequence of packets at the input of the base station, where it can be seen that the packet 丨 has not reached 20 in the time domain portion considered in Figure 5B, but the order of the packets has become confusing. All lost or obtained a very long path from the transmitter to the receiver. On the other hand, the packet (i+2) has obtained a very favorable path, so the packet is input from the transmitter (ie encoder) to the base station On the way to the end, "the packet (i+i) is caught up. 200845644 In the fifth diagram B, the receiving device u in the first figure will classify the packets again in the correct order at time division. Further, the decoding device U It will be found that the packet with the number i has not yet appeared or an error has occurred. As a result, as shown in the fifth figure C, an alternative packet for the number 丨 will be generated. Thus the fifth figure c shows the output device 15 in the first figure. The output packet is changed from the base station according to the first picture to the path according to the second mobile unit. However, as shown in the second figure, the extended solution: ίο 15 identifies that the packet with the number i is Grouping The audio content of the fifth picture d is often separated from other points. In other words, the needle material returns to the synthesized audio content of the error concealment measure. The method of the present invention can be implemented by hardware or software depending on the situation. The method can be based on the digital storage f, in particular based on a method in which the magnetic disk or CD can be electronically connected to ==signal_CD or CD can be associated with programmable electric power: electric moon row. In general, the invention can also be In the middle of the computer program product, which is stored in the machine for use in the computer program product running on the computer, the hair (4) can be implemented as having a program=computer program' When the computer program runs on the computer: ;亍20 20 200845644 [Simple description of the diagram] The first figure shows a circuit block diagram of the implementation of the device for transmitting the data packet sequence; the second figure is for the data packet sequence An electrical 5 block diagram of the decoder performing the decoding; a third diagram is a circuit block diagram of an audio decoder for generating a sequence of data packets, and a fourth diagram specifically showing a more specific implementation of the audio rendering device of the second diagram ; 10 Figure A shows the sequence of data packets output by the encoder; Figure 5B shows the sequence of tribute packets received by the base station, and Figure 5C shows the inserted substitute packets that are output by the base station and received by the mobile unit. a data packet sequence; a fifth diagram D shows a sequence of 15 audio content generated in an audio presentation device; a sixth diagram A shows an extension-payload according to ISO/IEC 14496-3:2005 (E) MPEG4 Syntax; Figure 6B shows a table for illustrating the value of the extensionjype field; 20 Figure 7 shows an exemplary grouping syntax for a transform-based audio encoder/decoder; Figure 8 shows An example of an alternate packet that includes a valid packet syntax and predetermined content characteristics. [Main component symbol description] 21 200845644 Base station ίο Internet 12 device for providing alternative packets 14 Antenna 16 5 Substitute packet line 18 Mobile unit 20 Detector 22 Audio presentation device 24 Packet line 26 10 Time-frequency conversion stage 30 Psychoacoustics Model 32 Packet packetizer 34 Packet unpacker 40 Dequantizer 42 Receiving device 11 Detection device 13 Output device 15 Packet line 17 Control line 19 Receiver 21 Error concealing device 23 Output unit 25 Substitute packet line 27 Quantizer 31 Entropy encoder 33 Entropy Encoder 41 Frequency Conversion 43 22