CN102982827B - Mode matching method of audio data for digital audio tape (DAT) audio repair - Google Patents

Mode matching method of audio data for digital audio tape (DAT) audio repair Download PDF

Info

Publication number
CN102982827B
CN102982827B CN 201210510014 CN201210510014A CN102982827B CN 102982827 B CN102982827 B CN 102982827B CN 201210510014 CN201210510014 CN 201210510014 CN 201210510014 A CN201210510014 A CN 201210510014A CN 102982827 B CN102982827 B CN 102982827B
Authority
CN
Grant status
Grant
Patent type
Prior art keywords
file
sample
secondary
audio
step
Prior art date
Application number
CN 201210510014
Other languages
Chinese (zh)
Other versions
CN102982827A (en )
Inventor
许舒亚
李柏岩
宋晖
陈强斌
王劲松
袁征
黄钢
韩斌
Original Assignee
上海音乐学院
上海协言科学技术服务有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Grant date

Links

Abstract

本发明公开了一种音频数据的模式匹配方法,包括:使用转录机对磁带进行第一及第二次转录,获得第一及第二转录音频;将第一及第二转录音频的数据进行归一化处理,并将两者中长度较长的作为主文件,长度较短的作为辅文件;将辅文件相对于主文件对齐;分割主文件,得到多个主文件样本段;在辅文件中搜索与各个主文件样本段波形匹配的辅文件样本段,得到各个主文件样本段与各个辅文件样本段间的匹配关系;分析匹配关系,得到可用的主文件样本段和辅文件样本段,并将其输出作为用于对磁带的转录音频进行修复的数据。 The present invention discloses a pattern matching method of audio data, comprising: using a first transcription unit for the tape and the secondary transcription to obtain first and second audio transcription; transcription of the first and second audio data are normalized a process, and both longer length as the main document, a short length of the secondary document; secondary document with respect to the alignment master file; split master files, master file of samples to obtain a plurality of segments; in the secondary file secondary file sample segment waveform matching searches the respective master file sample segment, is matched relationship between the respective master file sample segments with respective auxiliary document sample period; analysis matching relation to afford available primary file sample segments and the secondary file sample section, and the outputs it as an audio tape for transcription of data repair. 本发明通过对磁带进行两次转录得到两个转录音频并进行匹配,得到这两个转录音频中可用的样本段,为磁带的转录音频的修复创造了条件。 The present invention is twice the tape by two transcripts transcribed audio and matched to obtain two segments transcribed audio samples available, create the conditions for an audio tape transcription repair.

Description

用于DAT音频修复的音频数据的模式匹配方法 Pattern matching method for repairing an audio DAT audio data

技术领域 FIELD

[0001] 本发明涉及一种匹配方法,尤其涉及一种音频数据的模式匹配方法。 [0001] The present invention relates to a matching method, particularly to a pattern matching method of audio data.

背景技术 Background technique

[0002] DAT(digtal aud1 tape)指的是数字音频磁带,这是一种高密度数字磁带录音介质,频率采样为三种模拟LP32KHZ、数字44.1KHZ及模拟48KHZ,主要流行于上世纪九十年代。 [0002] DAT (digtal aud1 tape) refers to a digital audio tape, which is a high density digital tape recording medium, sampling frequency into three analog LP32KHZ, digital and analog 48KHZ 44.1KHZ, is prevalent in the nineties . 由于开发代价及制造代价高昂及出于对版权的考虑,其在当时面向的主要用途为专业领域,多为灌制CD前的母带制作。 Due to the cost of development and manufacturing costly and for reasons of copyright, its main purpose was oriented to professional fields, mostly pre-mastering mastering CD. 因此目前世界各国的电台、电视台等机构留存着很多存有珍贵录音资料的DAT磁带,由于磁带不利于保存,并且面临着DAT技术逐步退出市场,适用的DAT转录机越来越少,因此对这些DAT磁带的转录工作是刻不容缓的。 So now the world radio, television and other agencies retained a lot of information there rare recordings of DAT tape, because the tape is not conducive to preservation, and faced with the DAT technology gradually withdraw from the market for the DAT machine transcription less and less, and therefore these DAT tape transcription work is urgent.

[0003] 然而,在使用DAT转录机将DAT磁带上的数据转录为数字音频数据时,由于该类型的DAT转录机器本身固有的缺陷,使转录数据会随机产生小段的扭曲、丢失等现象。 [0003] However, when using the data DAT DAT machine transcribed on the tape transcribed into digital audio data, due to the inherent disadvantage of this type of machine itself DAT transcription, transcription data randomly twisted subparagraph loss and so on. 我们在实践中发现,转录数据中丢失的数据的持续时间约为0.1〜1.5秒,这严重了影响转录音频的最终试听效果。 We found that in practice, the transcription of data loss in the duration of data about 0.1~1.5 seconds, which seriously affects the transcription of the audio of the final audition results. 并且我们发现,这种丢失数据段出现的频率和位置也不是固定的,而是和具体的机器、录制开始时间、转录数据长度等因素有关,较长的数据丢失现象一般每10分钟左右出现一次,在一段20分钟左右的音频数据中数据丢失或损坏现象出现频率约为2〜8次。 And we found that the frequency and location of such loss of data segment appears not fixed, but rather concrete and machines, recording start time, length of data transcription and other factors, generally occur once every 10 minutes or so longer data loss , data loss or corruption in the period of about 20 minutes of audio data in the frequency of the phenomenon is about 2~8 times. 因此,经过DAT转录机转录得到的数字音频需要进行修复。 Therefore, after transcription of digital audio DAT machine transcribed needs to be fixed.

[0004] 上述的这种DAT转录过程中发生的现象不是在个别的DAT转录机上发生的偶然现象,而是几乎每台DAT转录机在转录时都会发生上述的现象,只是出现的频率或时间不同。 [0004] The above-described phenomenon occurs by chance during transcription of such DAT not occur on individual transcription DAT machine, but will almost every DAT transcriptional machinery above phenomenon occurs when transcribed, different frequency or time just occurring . 由于目前DAT技术逐步退出了市场,这种DAT转录机已经停产,而厂家也不再提供技术支持,因此在硬件上修复转录得到的数字音频是不可能的。 As the current DAT technology gradually withdraw from the market, this DAT transcriptional machinery has been discontinued, while manufacturers will no longer provide technical support, repair of digital audio transcription thus resulting in the hardware is not possible. 而到国外专门的修复机构去修复的话,成本会很高。 By specialized agencies abroad to repair the repair, the cost will be high. 因此必须考虑通过数字处理等技术手段来对转录得到的音频结果进行分析、研宄,以找到修复方法、手段,实现对DAT的转录音频的修复。 Must be considered to analyze the results of the audio transcript obtained by digital processing and other technical means, traitor research to find a fix, means to achieve the repair of DAT audio transcription.

[0005] 因此,本领域的技术人员致力于开发一种音频数据的模式匹配方法,在DAT转录音频中找到可用的数据用于DAT音频修复。 [0005] Accordingly, those skilled in the art to develop a method of pattern matching audio data, finds an available data DAT DAT audio dubbing audio for repair.

发明内容 SUMMARY

[0006] 有鉴于现有技术的上述缺陷,本发明所要解决的技术间题是提供一种音频数据的模式匹配方法,通过对磁带的两次转录得到的两个转录音频进行模式匹配,获得可用于转录音频的修复的数据,尤其是对DAT磁带的两次转录得到的两个转录音频进行模式匹配,获得可用于DAT磁带的转录音频的修复的数据。 [0006] In view of the above-described drawbacks of the prior art, the present invention is to provide an inter-technology problem is audio data pattern matching method, two by two transcriptional transcribed audio tape was subjected to pattern matching, are available to obtain transcribed audio data to repair, especially two dubbing audio tape DAT twice transcription was subjected to pattern matching, data may be obtained for dubbing audio tape DAT repair.

[0007] 为实现上述目的,本发明提供了一种音频数据的模式匹配方法,其特征在于,包括: [0007] To achieve the above object, the present invention provides a method of pattern matching audio data, characterized by comprising:

[0008] 步骤100使用转录机对磁带进行第一次转录,获得的数字波形音频作为第一转录音频;使用所述转录机对所述磁带进行第二次转录,获得的数字波形音频作为第二转录音频;所述第一转录音频和所述第二转录音频的采样频率相同; [0008] Step 100 using the transcriptional machinery for transcription cassette for the first time, the digital audio waveform obtained as the first audio transcription; transcription using the machine for the second transcription cassette, obtained as a second digital audio waveform the audio transcription; transcription of the same first and second audio dubbing audio sampling frequency;

[0009] 步骤200将所述第一转录音频和所述第二转录音频的数据进行归一化处理;将完成所述归一化处理的所述第一转录音频和所述第二转录音频中长度较长的一个作为主文件,长度较短的一个作为辅文件; The transcription of the first 200 and the second audio dubbing audio data is normalized [0009] Step; to complete the normalization processing said first and said second audio transcription audio transcription a longer length as a main document, a short length of a secondary document;

[0010] 步骤300将所述辅文件相对于所述主文件对齐; [0010] step 300 the secondary documents are aligned relative to the main file;

[0011] 步骤400分割所述主文件,得到多个样本段作为主文件样本段,并对各个所述主文件样本段编号; [0011] The step of dividing said master file 400, a plurality of samples to obtain a sample file segments as a main section, and each of the main section sample file number;

[0012] 步骤500在所述辅文件中搜索与各个所述主文件样本段的波形匹配的样本段作为辅文件样本段,得到各个所述主文件样本段与各个所述辅文件样本段之间的匹配关系; [0012] Step 500 searches the file in the secondary sample waveform segments match each of the main section of the sample as the secondary file file sample segment, between each of the main file to obtain samples each of the secondary section and section file samples the matching relation;

[0013] 步骤600分析所述匹配关系,得到可用的所述主文件样本段和可用的所述辅文件样本段; [0013] Step 600 analyzes the matching relation, said master file to obtain the available sample file section and the secondary section available sample;

[0014] 步骤700输出所述可用的所述主文件样本段和所述可用的所述辅文件样本段作为用于对所述磁带进行音频修复的数据。 Said master file sample period [0014] Step 700 outputs the available sample and the secondary segment of the file as the data available for the repair of audio tape.

[0015] 进一步地,所述步骤100中的所述转录机为DAT转录机,所述磁带为DAT磁带。 [0015] Furthermore, the transcription in the step 100 is a transcriptional machinery DAT, the tape is a DAT tape.

[0016] 进一步地,所述步骤100中的采样频率为44.1kHz ο [0016] Further, in step 100 the sampling frequency is 44.1kHz ο

[0017] 进一步地,所述步骤300中的将所述辅文件相对于所述主文件对齐包括: [0017] Furthermore, the secondary file in the step 300 relative to the main document alignment comprises:

[0018] 步骤310将所述辅文件中的前20%位置处取一个样本段,在所述主文件的前10% -30%的范围内搜索与所述样本段的起点位置Pbdi的数据匹配的数据的位置P min; [0018] Step 310 will take a sample segment of the position of the top 20% in the secondary file in the range of 10% -30% of the front of the main data file matches the search start position of the sample segment of Pbdi position data P min;

[0019] 步骤320若Pmin< 0,则对齐失败;若P min> P Bm,则设定所述主文件上与所述辅文件匹配的起点位置Pa1iS =Pa1i= Pmin-Pail,所述辅文件上与所述主文件匹配的起点位置PBh为所述辅文件的起点位置Pbci;若P min< P Bm,则设定所述主文件上与所述辅文件匹配的起点位置PAh为所述主文件的起点位置P ω,所述辅文件上与所述主文件匹配的起点位置PBh为:P Bh=PBnTPmin;g P min= PBm,则所述主文件和所述辅文件已对齐,设定所述主文件上与所述辅文件匹配的起点位置Pa1i为所述主文件的起点位置P ω,所述辅文件上与所述主文件匹配的起点位置PBh为所述辅文件的起点位置PK)。 [0019] Step 320 if Pmin <0, the alignment failure; if P min> P Bm, is set on the master file and the file matching the start position of the secondary Pa1iS = Pa1i = Pmin-Pail, the secondary file the matching of the main file start position as the secondary file PBh start position Pbci; if P min <P Bm, the master file is set with the secondary file matches the start position of the main PAh file start position P ω, the secondary file of the main file is matched to the starting position PBh: P Bh = PBnTPmin; g P min = PBm, then the master file and the auxiliary file aligned, set the main file and auxiliary files matching the start position to the start position of the main Pa1i file P ω, the secondary file on the primary file matches the start position of the secondary PBh file start position PK ).

[0020] 进一步地,所述步骤310中的所述样本段的长度为217。 [0020] Further, the length of the sample in the section 310 in step 217.

[0021] 进一步地,所述步骤400中的对所述主文件的所述分割是等长度的分割,经过所述分割得到的所述各个主文件样本段的长度皆为214。 [0021] Further, dividing the said master file in the step 400 is equally divided length, the length obtained through the division of each section of the main file are all 214 samples.

[0022] 进一步地,所述步骤400中的对各个所述主文件样本段的编号是从所述主文件的所述起点位置pAh开始依次递增的。 [0022] Further, the number of samples of each of the main section of the document in the step 400 from the start position of the master file pAh sequentially incremented.

[0023] 进一步地,所述步骤500中的在所述辅文件中进行的所述搜索是从所述辅文件的所述起点位置PBh开始的。 [0023] Further, the search is performed in the secondary file in the step 500 is the start position being the secondary file PBh started.

[0024] 进一步地,所述步骤500中的与一个所述主文件样本段波形匹配的一个所述辅文件样本段中的各个数据与所述主文件样本段中的各个数据依次对应且相等。 [0024] Further, each of the secondary data file in a sample section of said master file with a sample waveform segment matching step 500 with the respective data files in the primary segment sequentially corresponding to the sample and equal.

[0025] 进一步地,所述步骤600中的分析所述匹配关系包括: [0025] Further, in the step 600 of analyzing the matching relationship comprises:

[0026] 步骤610对于每一个所述主文件样本段和与其波形匹配的一个所述辅文件样本段,获取所述主文件样本段在所述主文件中相对于所述起点位置pAh的位置PA和所述辅文件样本段在所述辅文件中相对于所述起点位置PBh的位置PB; [0026] Step 610 of the secondary file for each sample period of said sample master file and its waveform matching section acquires the sample master file in said master file segment relative to the start position position PA of pAh sample and the secondary file in the secondary file segment relative to the start point position PBh position PB;

[0027] 步骤620若Pa= PB,则所述主文件样本段和所述辅文件样本段为所述可用的所述主文件样本段和所述可用的所述辅文件样本段;若PA> Pb,则所述主文件样本段为所述可用的所述主文件样本段;若PB> PA,则所述辅文件样本段为所述可用的所述辅文件样本段。 [0027] If Step 620 Pa = PB, then the primary section and the secondary file samples file segment of the sample available sample said master file section and the secondary section of the sample files available; if PA> Pb, then said master file to said master file sample period of the sample period is available; if PB> PA, then the secondary section to the secondary sample document file segment of the sample available.

[0028] 在本发明的较佳实施方式中,使用本发明的音频数据的模式匹配方法对DAT磁带进行了音频修复,包括步骤:使用DAT转录机对DAT磁带进行第一次转录,获得的数字波形音频作为第一转录音频,使用上述DAT转录机对该DAT磁带进行第二次转录,获得的数字波形音频作为第二转录音频,其中,第一转录音频和第二转录音频的采样频率皆为44.1kHz ;将第一转录音频和第二转录音频的数据进行归一化处理,并将其中长度较长的一个作为主文件,长度较短的一个作为辅文件;确定主文件上与辅文件匹配的起点位置PAh,确定辅文件上与主文件匹配的起点位置PBh,使辅文件相对于主文件对齐;等长度分割主文件,得到多个长度皆为214的样本段作为主文件样本段,并对各个主文件样本段编号;在从辅文件上与主文件匹配的起点位置PBh开始搜索辅文件中与各个主文件样 [0028] In a preferred embodiment of the present invention, a pattern matching method according to the present invention, audio data of a DAT audio tape the repair, comprising the step of: using DAT DAT tape transcription machine transcription for the first time, the obtained digital as a first waveform of an audio dubbing audio, using the DAT machine transcription of the second transcriptional DAT tape, obtained as a second digital waveform audio dubbing audio, wherein transcription of the first and second audio sampling frequency are all transcribed audio 44.1kHz; the first and second transcription audio dubbing audio data is normalized, and wherein a longer length as a master file, a shorter length as auxiliary file; determining the main file and auxiliary files match PAH to the starting position, on the secondary is determined with the master document file matching PBH to the starting position, so that the secondary document alignment relative to the main file; equal length divided main file, to obtain a plurality of sample segments 214 are all of the length of the sample period as a main file, and master file for each sample a segment number; in the main file, the auxiliary file matches from the start position to start the search PBh respective main and secondary file key file 本段的波形匹配的样本段作为辅文件样本段,得到各个主文件样本段与各个辅文件样本段之间的匹配关系;分析该匹配关系,得到可用的主文件样本段和可用的辅文件样本段;将可用的主文件样本段和可用的辅文件样本段输出,作为用于对该DAT磁带的转录音频进行修复的数据。 Sample waveform segments of this paragraph as a secondary matching file sample period be matched between the various master files with the respective secondary sample period sample file segments; analyzing the matching relation, it can be used to obtain a sample of the master file and available secondary segment file samples section; master file will be available and the available sample segments secondary segment output file sample, as for repair of the transcribed audio data DAT tape.

[0029] 由此可见,本发明的音频数据的模式匹配方法通过对DAT磁带进行两次转录得到两个转录音频,通过对这两个转录音频进行匹配,得到这两个转录音频中可用的样本段,这些可用的样本段可用于对DAT磁带的转录音频进行修复。 [0029] Thus, the audio data of the pattern matching method of the present invention, DAT tape twice by two transcription transcribed audio dubbing audio by the two match, to obtain two samples of audio transcription available segments, these segments may be used for samples dubbing audio DAT tape to repair. 因此本发明为DAT的转录音频的修复创造了条件,从而有利于转录DAT磁带以抢救这些磁带中存储的珍贵录音资料。 The invention thus creating the conditions for the transcription of audio DAT repair, to facilitate transcription DAT tape recordings to rescue valuable information stored in these tapes. 并且,本发明的音频数据的模式匹配方法并不限于用于DAT磁带,还可以用于其他磁带或存储媒介,因而具有广泛的应用领域和应用前景。 Further, the pattern matching method of audio data according to the present invention is not limited to use DAT tape, or magnetic tape may also be used for other storage media, which has a wide range of applications and application prospects.

[0030] 以下将结合附图对本发明的构思、具体结构及产生的技术效果作进一步说明,以充分地了解本发明的目的、特征和效果。 [0030] Hereinafter, in conjunction with the accompanying drawings of the inventive concept, the specific structure and technical effect produced is further described to fully understand the objects, features and advantages of the present invention.

附图说明 BRIEF DESCRIPTION

[0031] 图1是本发明的音频数据的模式匹配方法的流程图。 [0031] FIG. 1 is a flowchart of pattern matching method of the present invention audio data.

[0032]图2是作为数字波形音频的转录音频的示意图。 [0032] FIG. 2 is a schematic diagram of a digital audio waveform in the audio transcription.

[0033] 图3是本发明的音频数据的模式匹配方法中确定匹配的起点位置的一个示例的示意图。 [0033] FIG. 3 is a schematic view of one example of a pattern matching method of the present invention, audio data that matches the determined starting position.

[0034] 图4是本发明的音频数据的模式匹配方法中确定匹配的起点位置的另一个示例的示意图。 [0034] FIG. 4 is a schematic diagram of another example of a pattern matching method according to the present invention, audio data that matches the determined starting position.

[0035] 图5是在辅文件B中搜索与主文件A的各个主文件样本段匹配的辅文件样本段的示意图。 [0035] FIG. 5 is a schematic view of the secondary file B Sample respective master file search section of the main file A sample file matches the secondary segments.

具体实施方式 detailed description

[0036] 在本实施例中,应用本发明的音频数据的模式匹配方法对DAT磁带进行音频修复,如图1所示,包括步骤: [0036] In the present embodiment, the pattern matching method applied to the present invention the audio data to an audio tape DAT repair, shown in Figure 1, comprising the steps of:

[0037] 步骤100,获得第一及第二转录音频。 [0037] Step 100 to obtain a first and a second audio transcription.

[0038] 本步骤中,首先使用一台DAT转录机对需要修复的DAT磁带进行第一次转录,获得一个数字波形音频,将该数字波形音频作为第一转录音频;然后再使用该DAT转录机对该DAT磁带进行第二次转录,获得另一个数字波形音频,将该数字波形音频作为第二转录音频。 [0038] In this step, first, using a DAT machine in need of repair transcription DAT tape transcription for the first time, to obtain a digital waveform audio, the digital audio waveform as the first audio transcription; transcription DAT and then use the machine DAT tape for the second transcription obtain another digital waveform audio, the digital audio waveform as the second audio transcription.

[0039] 其中,第一转录音频和第二转录音频的采样频率皆为44.1kHz„ [0039] wherein transcription of the first and second audio sampling frequencies are all transcribed audio 44.1kHz "

[0040] 步骤200,对转录音频进行预处理。 [0040] Step 200, the audio transcription pretreatment.

[0041] 本步骤中,对步骤100中获得的第一及第二转录音频进行预处理,预处理包括对该第一及第二转录音频的数据进行归一化处理,以使两个转录音频能够被比较、匹配。 [0041] In this step, the first and second audio transcription obtained in step 100 is pre-processed, pre-processing comprises normalizing the first and second dubbing audio data, so that transcription of two audio It can be compared to a match. 本实施例中,使经过归一化处理的第一及第二转录音频的数据的值在-1.0〜1.0的范围内。 In this embodiment, the first and the second value after dubbing audio data normalized in the range of -1.0~1.0.

[0042] 然后将两者中长度较长的一个作为主文件A,长度较短的一个作为辅文件B。 [0042] and then both a longer length as the main file A, as a shorter length secondary file B. 其中,(如图2所示),作为数字波形音频的转录音频中的各个数据的序号(即采样点)从I开始递增(增量为I),其值与其序号在坐标系中形成波形曲线,例如图2中的数据a(Ai,i)表示该数据是转录音频中的第i个数据,其值为Ai,其序号为i。 Wherein, (2), a digital audio waveform in the audio transcript of each data sequence number (i.e., sampling point) from the beginning I is incremented (increment I), its value and its ID are formed in the waveform plot coordinates , for example, data in Figure 2 a (Ai, i) indicates that the data is an audio transcript of i-th data, the value Ai, which is the serial number i. 这样,转录音频的长度即其包含的数据的个数,典型地,转录音频的长度为228。 Thus, the number of transcript length of the audio data i.e., it contains, typically, a length of 228 audio transcription.

[0043] 步骤300,将主文件与辅文件对齐。 [0043] Step 300, to align with the main file and auxiliary files.

[0044] 由于后续的步骤中的比较、匹配皆是从文件头开始,因此将到步骤200中确定的主文件A和辅文件B对齐并确定匹配的起点位置是很重要的,具体地包括步骤310和320。 [0044] Since Comparative subsequent step, the matching are all beginning of the file, thus in the step 200 determines that the primary document A and the secondary file B are aligned and a match is determined starting position is very important, in particular comprising the steps of 310 and 320.

[0045] 步骤310,在辅文件中取一个样本段。 [0045] Step 310, taking a sample segment in the secondary file.

[0046] 具体地,在步骤200中确定的辅文件B中的前20%位置(即从辅文件B的起点位置开始距离该起点位置20%辅文件B的长度处的位置)处取一个样本段,该样本段的长度为217(即该样本段中包含217个数据,以下不作说明的长度皆表示其中包含的数据的个数),记录该样本段的起点位置PBm(可以认为起点位置是起始的数据在整个转录音频的数据中的序号)。 (At position i.e. starting from the starting position of the secondary file B from the start position of 20% of the secondary file B length) [0046] Specifically, the secondary file B determined in step 200 in the top 20% at position taking a sample segment, the length of the sample stage 217 (i.e., in the sample period contains 217 data, the following is not the length of the description are represented by the number of data contained therein), the recording start position of the sample segment PBm (may be considered the starting position is starting data number in the entire transcribed audio data). 在步骤200中确定的主文件A的前10% -30%的范围内搜索与该样本段的起点位置PBm的数据匹配的数据的位置P min。 The position of the search data matches the start position of the sample segment PBm range of 10% -30% of pre determined in step 200 of the main file A data P min.

[0047] 步骤320,确定匹配的起点位置。 [0047] Step 320, a match is determined start position.

[0048] 具体地为: [0048] specifically:

[0049] 若Pmin< 0,则对齐失败,进入步骤800 ; [0049] if Pmin <0, the alignment failure, proceeds to step 800;

[0050] 若Pmin> P Bm,如图3所示,则设定主文件A上与辅文件B匹配的起点位置PAh为:Pai1= P min_PBm,辅文件B i:与主文件A匹配的起点位置为辅文件B的起点位置Pbci;删除主文件A上在匹配的起点位置PAh之前的部分,进入步骤400 ; [0050] When the Pmin> P Bm, shown in Figure 3, file B is set to match with the secondary master file starting position A of PAh: Pai1 = P min_PBm, auxiliary file B i: A file matching the main starting point starting position of the file B supplemented Pbci; PAh section prior to the starting position a in the primary file delete matched, proceeds to step 400;

[0051] 若Pmin< PBm,如图4所示,则设定主文件A上与辅文件B匹配的起点位置为主文件A的起点位置PA(I,辅文件B上与主文件A匹配的起点位置PBh为:P Bh= P Bm_Pmin;删除辅文件B上在匹配的起点位置PBh之前的部分,进入步骤400 ; [0051] When Pmin <PBm, as shown, is set on the main file A and the file B match auxiliary start position 4 based start position PA of the file A (I, B as the main file and auxiliary files that match A PBh starting position is: P Bh = P Bm_Pmin; PBh section prior to the starting position in the file B to delete the secondary matching proceeds to step 400;

[0052] 若Pniin= P Bm(未图示),则主文件A和辅文件B已对齐,主文件A上与辅文件B匹配的起点位置PAh为主文件A的起点位置PA(l,辅文件B上与主文件A匹配的起点位置PBh为辅文件B的起点位置PB(I,进入步骤400。 [0052] When Pniin = P Bm (not shown), the main file A and the file B are aligned secondary, and auxiliary file starting position B PAh matching document A based on the master file starting position A of PA (l, secondary file B with a primary file starting position a of the start position matches the supplemented PBh file B PB (I, proceeds to step 400.

[0053] 步骤400,分割主文件。 [0053] Step 400, the divided master file.

[0054] 在本步骤中,对经过步骤320的主文件A进行等长度分割,得到多个主文件样本段,这些主文件样本段的长度皆为214(即每个主文件样本段中皆包含214个数据)。 [0054] In this step, after the step of the master file A 320 is the length of dividing the like, to obtain a plurality of samples of the master file segments, the length of these segments are all master file 214 samples (i.e., sample of each master file contains all segments 214 data). 从主文件A的起点位置PAh开始(即从主文件A的文件头开始,由于在步骤320中已删除了主文件A上在匹配的起点位置PAh之前的部分),对每个主文件样本段进行递增编号,例如A01、A02、A03、...ο Starting at the beginning of the main file A PAh location (i.e., beginning of the file A master file, since the step portion 320 has been removed prior to the starting position of the match PAh main file A), for each main segment file samples incrementing number, such as A01, A02, A03, ... ο

[0055] 步骤500,获得主文件样本段与辅文件样本段之间的匹配关系。 [0055] Step 500, the matching relation is obtained between the sample and the sample file segment auxiliary section master file.

[0056] 具体地,从经过320的辅文件B的起点位置PBh(即从辅文件B的文件头开始,由于在步骤320中已删除了辅文件B上在匹配的起点位置PBh之前的部分)开始搜索与步骤400中得到的各个主文件样本段匹配的辅文件样本段,并对该辅文件样本段编号。 [0056] In particular, (i.e., start of the file a secondary file B, as in step 320 have removed some previous start position PBh on the secondary file B in the matching) from the through secondary file B 320 in the start position PBh sample segments each secondary document segment matching master file sample obtained in step 400 to start the search, and the file number of the auxiliary sample period. 例如,如图5所示,辅文件样本段BOl与一个主文件样本段AOl匹配,辅文件样本段B02与一个主文件样本段A02匹配。 For example, as shown, the secondary file sample segment with a master file BOl sample AOl matching section, the secondary sample file with a master file segment B02 A02 sample segments 5 match. 当一个辅文件样本段B02与一个主文件样本段A02匹配时,辅文件样本段B02中的各个数据与主文件样本段A02中的各个数据依次对应且相等,即辅文件样本段B02中的的第j个数据与主文件样本段A02中第j个数据的幅度相等(或者两者的幅度差小于1% ),其中j遍历主文件样本段A02和辅文件样本段B02中的所有数据的个数。 When a secondary file sample segment B02 with a master file sample segment A02 match, the respective data of respective secondary file sample section B02 in the sample section in the main file in A02 sequentially correspond and are equal, i.e., the secondary file sample segment B02 of the the amplitude of the j-th sample data file with the master in paragraph A02 j-th data is equal to (or both the amplitude difference is less than 1%), where all the data of a master file sample j traverse sections A02 and B02 in the secondary section of the sample files number.

[0057] 具体的匹配过程为: [0057] The specific matching process is:

[0058] 对于每一个主文件样本段,在辅文件B中从辅文件B的起点位置PBh开始上述的搜索,第一次搜索的范围为[PBh,PBh+27X214]; [0058] For each sample period a master file, the auxiliary file B from the start position of the secondary file B above PBH start search for the first search range [PBh, PBh + 27X214];

[0059] 如果未找到,则将搜索的范围向辅文件B的尾部移动长度211,即对于第k次搜索,搜索的范围为[(k-1) X2n+PBh, (k-1) X2n+PBh+27X214],其中k = 1、2、…; [0059] If it is not found, then a search range of length of the tail moving the secondary file B 211, i.e., for the k th search, the search range is [(k-1) X2n + PBh, (k-1) X2n + PBh + 27X214], where k = 1,2, ...;

[0060] 如果搜索到达辅文件B的尾部仍未找到匹配该主文件样本段的辅文件样本段,则跳到下一个主文件样本段,开始搜索与该主文件样本段匹配的辅文件样本段。 [0060] If the search reaches the trailing secondary file B not find a matching sample of the main document file sample stage secondary segment, skip to the next sample period a master file, the file start searching the secondary sample section matching the section of the main file samples .

[0061 ] 步骤600,分析匹配关系。 [0061] Step 600, analyzing the matching relationship.

[0062] 本步骤通过分析步骤500中得到的匹配关系,得到可用的主文件样本段和可用的辅文件样本段,具体地包括步骤610和620。 [0062] The step of matching relations analysis step 500, to obtain a sample segments available primary and secondary file segments files available samples, particularly 610 and 620 comprising the step of.

[0063] 步骤610,对于每一个主文件样本段和与其波形匹配的一个辅文件样本段,获取该主文件样本段在主文件A中的位置? [0063] Step 610, a secondary document for each sample period a master file segment and sample waveform matching therewith, acquires the position of the main section in the main document file A sample of? 4(即其开始位置相对于主文件A的文件头的位置),以及获取该辅文件样本段在辅文件B中的位置Pb (即其开始位置相对于辅文件B文件头的位置)。 4 (i.e., the position of its starting position relative to the main file A file header), and obtain the secondary file in the secondary position of the sample segment in file B Pb (i.e. its starting position relative to the position of the secondary file header file B).

[0064] 步骤620,包括: [0064] Step 620, comprising:

[0065] SPa= Pb,则该主文件样本段和该辅文件样本段为可用的主文件样本段和可用的辅文件样本段; [0065] SPa = Pb, the master file and the auxiliary file segment sample sample sample period is a front section and sample files available secondary segment files available;

[0066] 若PA> PB,则该主文件样本段为可用的主文件样本段,该辅文件样本段为不可用的辅文件样本段; [0066] If PA> PB, the sample period is available master file master file sample period, the sample period is auxiliary secondary file section file samples unavailable;

[0067] 若PB> PA,则该辅文件样本段为可用的辅文件样本段,该主文件样本段为不可用的主文件样本段。 [0067] When PB> PA, then the secondary section is a sample document file sample available auxiliary section, the master file as the primary file sample period sample segments unavailable.

[0068] 如图5所示的主文件样本段A02在主文件A中的位置PjP辅文件样本段B02在辅文件B中的位置的关系为PA= PB,则可知主文件样本段A02和辅文件样本段B02为可用的主文件样本段和可用的辅文件样本段。 Relationship between the position of [0068] the master file shown in FIG. 5 PjP sample segment position A02 in the main file and auxiliary files A sample of B02 in the secondary segment file B is PA = PB, it is known that the master file and the sub sample segment A02 sample B02 is a file section file samples available primary and secondary segments file sample period is available.

[0069] 步骤700,输出步骤600中得到的可用的主文件样本段和可用的辅文件样本段,作为用于对DAT磁带的转录音频进行修复的数据。 [0069] Step 700, the step 600 resulting output file samples available in the main section and auxiliary files available sample period, as a data repair transcription DAT audio tapes.

[0070] 步骤800,结束。 [0070] Step 800, the end.

[0071] 以上详细描述了本发明的较佳具体实施例。 [0071] The above detailed description of particular preferred embodiments of the present invention. 应当理解,本领域的普通技术人员无需创造性劳动就可以根据本发明的构思做出诸多修改和变化。 It should be understood by those of ordinary skill in the art without creative work can make many modifications and variations of the inventive concept. 因此,凡本技术领域的技术人员依本发明的构思在现有技术的基础上通过逻辑分析、推理或者有限的实验可以得到的技术方案,皆应在由权利要求书所确定的保护范围内。 Therefore, all those skilled in the art under this inventive concept on the basis of prior art technical solutions through logical analysis, reasoning or limited experiments can be obtained, are to be within the scope defined by the claims of.

Claims (10)

  1. 1.一种音频数据的模式匹配方法,其特征在于,包括: 步骤(100)使用转录机对磁带进行第一次转录,获得的数字波形音频作为第一转录音频;使用所述转录机对所述磁带进行第二次转录,获得的数字波形音频作为第二转录音频;所述第一转录音频和所述第二转录音频的采样频率相同; 步骤(200)将所述第一转录音频和所述第二转录音频的数据进行归一化处理;将完成所述归一化处理的所述第一转录音频和所述第二转录音频中长度较长的一个作为主文件,长度较短的一个作为辅文件; 步骤(300)将所述辅文件相对于所述主文件对齐; 步骤(400)分割所述主文件,得到多个样本段作为主文件样本段,并对各个所述主文件样本段编号; 步骤(500)在所述辅文件中搜索与各个所述主文件样本段的波形匹配的样本段作为辅文件样本段,得到各个所述主文件样本段与各个 1. A method of pattern matching audio data, characterized by comprising: the step (100) using the transcriptional machinery of the first tape transcription, audio digital waveform obtained as the first audio transcription; transcription using the machine the said second magnetic tape transcription, obtained as a second digital audio waveform audio transcription; transcription same as the first audio and the second audio sampling frequency of transcription; step (200) of the first audio and the transcription said second transcribed audio data normalized; the completion of the second audio transcription in a longer length, a shorter length of the normalization processing of the first audio and the transcription of a file as the primary as a secondary document; step (300) the file to the secondary alignment relative to the main file; dividing step (400) of the master file, a plurality of samples to obtain a sample file segments as a main section, and each of the main file sample segment number; step (500) in the secondary file in the search for matching with the sample waveform segments each of said master file as auxiliary sample segment file sample segments, each of said master file to obtain samples each segment 述辅文件样本段之间的匹配关系;所述匹配过程如下: 对于每一个主文件样本段,在辅文件B中从辅文件B的起点位置PBh开始上述的搜索,第一次搜索的范围为[PBh,PBh+27X214]; 如果未找到,则将搜索的范围向辅文件B的尾部移动长度211,即对于第k次搜索,搜索的范围为[(k-1) X2n+Pa,(k-1) X2n+PBh+27X214],其中k = 1、2、…; 如果搜索到达辅文件B的尾部仍未找到匹配该主文件样本段的辅文件样本段,则跳到下一个主文件样本段,开始搜索与该主文件样本段匹配的辅文件样本段; 步骤(600)分析所述匹配关系,得到可用的所述主文件样本段和可用的所述辅文件样本段; 步骤(700)输出所述可用的所述主文件样本段和所述可用的所述辅文件样本段作为用于对所述磁带的转录音频进行修复的数据。 Matching relationship between said secondary segment file samples; the matching process is as follows: For each sample of a master file section, in the secondary file B from file starting position B of the auxiliary start PBh above search for the first search range [PBh, PBh + 27X214]; if not found, then a search range of length of the tail moving the secondary file B 211, i.e., for the k th search, the search range is [(k-1) X2n + Pa, (k -1) X2n + PBh + 27X214], where k = 1,2, ...; if the search reaches the trailing secondary file B not find a matching sample of the main document file sample stage secondary segment, skip to the next master file sample segment, segment start searching auxiliary files match the sample file master sample periods; the step (600) analyzing said matching relationship, said master file to obtain the available sample file section and the secondary section available sample; step (700) sample output file of the main section of the available sample and the secondary section of the file is available as a data repair of the transcribed audio tape.
  2. 2.如权利要求1所述的音频数据的模式匹配方法,其中所述步骤(100)中的所述转录机为DAT转录机,所述磁带为DAT磁带。 2. The pattern matching method according to claim 1 audio data, wherein said step of transcription (100) of the transcriptional machinery of DAT, the tape is a DAT tape.
  3. 3.如权利要求1或2所述的音频数据的模式匹配方法,其中所述步骤(100)中的采样频率为44.1kHz ο 3. The pattern of claim 12 or claim audio data matching method, wherein said step of sampling frequency (100) is 44.1kHz ο
  4. 4.如权利要求3所述的音频数据的模式匹配方法,其中所述步骤(300)中的将所述辅文件相对于所述主文件对齐包括: 步骤(310)将所述辅文件中的前20%位置处取一个样本段,在所述主文件的前10%-30%的范围内搜索与所述样本段的起点位置(PBm)的数据匹配的数据的位置(Pmin); 步骤(320)若Pmin〈0,则对齐失败^Pmin>PBm,则设定所述主文件上与所述辅文件匹配的起点位置(PAh)为=Pa1i= P min_PBm,所述辅文件上与所述主文件匹配的起点位置(PBh)为所述辅文件的起点位置(Pbci) ;SPmin〈PBm,则设定所述主文件上与所述辅文件匹配的起点位置(PAh)为所述主文件的起点位置(Paci),所述辅文件上与所述主文件匹配的起点位置(PBh)为:PBh= PBm_Pmin;gPmin= PBm,则所述主文件和所述辅文件已对齐,设定所述主文件上与所述辅文件匹配的起点位置(PiJ为所述主文件的起点位置(Paci),所述辅文件上与所述 Step (310) the secondary file: 4. The mode according to claim 3 audio data matching method, wherein said step of said secondary file (300) is aligned relative to the main file comprising 20% taken at a front position of a sample segment position (Pmin is) in the range of 10% -30% of the primary file before searching the start position (PBm) section of the sample data matches the data; a step ( 320) if Pmin <0, the alignment failure ^ Pmin> PBm, is set on the master file and the file matching the secondary to the starting position (PAH) is a = Pa1i = P min_PBm, the file with the secondary matching primary file starting position (PBH) starting from a position of said secondary file (Pbci); SPmin <PBm, is set on the master file and the file matching the secondary to the starting position (PAH) to said master file the starting position (Paci), the secondary of the main file that matches the file starting position (PBH) is: PBh = PBm_Pmin; gPmin = PBm, then the master file and the auxiliary file aligned, the set said master file on the secondary file matched with the start position (start position Pij is said master file (Paci), the file with the secondary 文件匹配的起点位置(PBh)为所述辅文件的起点位置(Pbq)。 Matching file starting position (PBH) starting from a position of said secondary file (Pbq).
  5. 5.如权利要求4所述的音频数据的模式匹配方法,其中所述步骤(310)中的所述样本段的长度为217。 5. The pattern matching method as recited in claim audio data, wherein the length of the sample period of said step (310) 217.
  6. 6.如权利要求5所述的音频数据的模式匹配方法,其中所述步骤(400)中的对所述主文件的所述分割是等长度的分割,经过所述分割得到的所述各个主文件样本段的长度皆为214。 6. The method of claim 5, wherein the pattern matching of the audio data obtained through the division respective main claims, wherein said master file dividing the said step (400) is equally divided in length, length of the file segment are all 214 samples.
  7. 7.如权利要求6所述的音频数据的模式匹配方法,其中所述步骤(400)中的对各个所述主文件样本段的编号是从所述主文件的所述起点位置(Pa1i)开始依次递增的。 7. The pattern matching method according to claim 6 audio data, wherein the number of samples of each of the main section in the file (400) step is to start from the start point position of the master file (Pa1i) incrementally.
  8. 8.如权利要求7所述的音频数据的模式匹配方法,其中所述步骤(500)中的在所述辅文件中进行的所述搜索是从所述辅文件的所述起点位置(PBh)开始的。 8. The pattern of the audio data matching method of claim 7, wherein in said step (500) performed in the secondary file in the search from the start position of the secondary file (PBH) It began.
  9. 9.如权利要求8所述的音频数据的模式匹配方法,其中所述步骤(500)中的与一个所述主文件样本段波形匹配的一个所述辅文件样本段中的各个数据与所述主文件样本段中的各个数据依次对应且相等。 9. The pattern matching method of audio data according to claim 8, wherein each one of said secondary data file with a section of said sample master file matches the sample waveform segments in the step (500) and the the data in each master file in turn corresponds to the sample period and equal.
  10. 10.如权利要求9所述的音频数据的模式匹配方法,其中所述步骤(600)中的分析所述匹配关系包括: 步骤(610)对于每一个所述主文件样本段和与其波形匹配的一个所述辅文件样本段,获取所述主文件样本段在所述主文件中相对于所述起点位置(PAh)的位置(Pa)和所述辅文件样本段在所述辅文件中相对于所述起点位置(PBh)的位置(Pb); 步骤(620) SPa= Pb,则所述主文件样本段和所述辅文件样本段为所述可用的所述主文件样本段和所述可用的所述辅文件样本段;若PA>PB,则所述主文件样本段为所述可用的所述主文件样本段;若PB>PA,则所述辅文件样本段为所述可用的所述辅文件样本段。 10. The pattern according to claim 9 audio data matching method, wherein said step of analyzing (600) said matching relationship comprising: a step (610) for each of said master file sample waveform segments matched thereto a sample period of said secondary file, said master file obtaining sample segments in the main document with respect to said starting position (PAH) (Pa) of the sample and the secondary file segment relative to the secondary file the starting position (PBH) position (Pb); step (620) SPa = Pb, said main section and said secondary file sample file of the sample period is available to the master file and sample segments available sample segments of the secondary file; if PA> PB, the sample section of said master file to said master file of the available sample period; if PB> PA, then the secondary section of the sample file is available sample segments of said secondary file.
CN 201210510014 2012-12-03 2012-12-03 Mode matching method of audio data for digital audio tape (DAT) audio repair CN102982827B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201210510014 CN102982827B (en) 2012-12-03 2012-12-03 Mode matching method of audio data for digital audio tape (DAT) audio repair

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201210510014 CN102982827B (en) 2012-12-03 2012-12-03 Mode matching method of audio data for digital audio tape (DAT) audio repair

Publications (2)

Publication Number Publication Date
CN102982827A true CN102982827A (en) 2013-03-20
CN102982827B true CN102982827B (en) 2015-07-22

Family

ID=47856729

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201210510014 CN102982827B (en) 2012-12-03 2012-12-03 Mode matching method of audio data for digital audio tape (DAT) audio repair

Country Status (1)

Country Link
CN (1) CN102982827B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101203917A (en) * 2005-06-22 2008-06-18 弗劳恩霍夫应用研究促进协会 Device and method for determining a point in a film comprising film data applied in chronological order
CN202093836U (en) * 2011-06-15 2011-12-28 深圳市佳音王电子有限公司 Audio converting device

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7266501B2 (en) * 2000-03-02 2007-09-04 Akiba Electronics Institute Llc Method and apparatus for accommodating primary content audio and secondary content remaining audio capability in the digital audio production process

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101203917A (en) * 2005-06-22 2008-06-18 弗劳恩霍夫应用研究促进协会 Device and method for determining a point in a film comprising film data applied in chronological order
CN202093836U (en) * 2011-06-15 2011-12-28 深圳市佳音王电子有限公司 Audio converting device

Also Published As

Publication number Publication date Type
CN102982827A (en) 2013-03-20 application

Similar Documents

Publication Publication Date Title
US20060221788A1 (en) Efficient techniques for modifying audio playback rates
US20070094031A1 (en) Audio time scale modification using decimation-based synchronized overlap-add algorithm
US20080249644A1 (en) Method and apparatus for automatically segueing between audio tracks
US6055495A (en) Speech segmentation
US7477739B2 (en) Efficient storage of fingerprints
US20040059570A1 (en) Feature quantity extracting apparatus
US20060041753A1 (en) Fingerprint extraction
US20040264715A1 (en) Method and apparatus for playback of audio files
US20080319743A1 (en) ASR-Aided Transcription with Segmented Feedback Training
US20130246077A1 (en) Adaptive processing with multiple media processing nodes
CN1338104A (en) Data framing for adaptive-block-length coding system
CN101155357A (en) Device and method for recording and saving voice call on mobile phone
US20020133357A1 (en) Method and system for smart cross-fader for digital audio
US20060224956A1 (en) Intelligent document saving
WO2001045391A1 (en) System and method for video production
Haitsma et al. Speed-change resistant audio fingerprinting using auto-correlation
US6530009B1 (en) Methods for increasing file storage capacity of optical discs
US7085613B2 (en) System for monitoring audio content in a video broadcast
CN101625857A (en) Self-adaptive voice endpoint detection method
JP2005532578A (en) System and method for providing a user control with the Embedded repeated objects in the stream
US20100003006A1 (en) Video searching apparatus, editing apparatus, video searching method, and program
JP2006109160A (en) Information processing device and method therefor, and program
CN103021440A (en) Method and system for tracking audio streaming media
US20090259326A1 (en) Server side audio file beat mixing
US8751022B2 (en) Multi-take compositing of digital media assets

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
C14 Grant of patent or utility model