Aligning text to audio and video using ELAN - Linguistics
Contents
1. Date Modified File Format Subtitles text file srt Encoding UTF 16 Hy Ce E New Folder Cancel Illustration 14 Exporting subtitles in UTF 16 stl format for DVD Studio Pro The resulting stl subtitles can then be imported into DVD Studio Pro by right clicking on one of the subtitles tracks S1 to 832 and choosing Import Subtitle File from the context menu which appears You can add up to 32 separate subtitles in this way By default each of the subtitles tracks added can be accessed on the final DVD by pressing the Subtitles button on the DVD player s remote for a stand alone DVD player or by selecting a particular subtitle track for DVD playback software on a computer 23 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox Section Three Producing YouTube videos with subtitles One of the advantages of investing effort in time aligned transcription using software like ELAN is the variety of output formats which can be produced automatically from the resulting transcripts Among other export options ELAN 1s capable of producing so called subtitles text files which present the contents of a selected set of ELAN tiers as captions These files can be used to provide user selectable subtitles for both video DVDs and popular on line video sharing sites such as YouTube This section describes how to produce capti2. Export location Users chris Desktop word_list_output E Choose Create Split files based on Name files Labels Using Label Track Name _ Include audio before first label C Numbering before Label Track Name First file name word _list C Numbering after Label Track Name Tracer File name prefix word_list MI Overwrite existing files Illustration 5 Exporting multiple audio files in Audacity After a short delay Audacity should inform you once it has completed extraction The result of this export process should be a series of audio files each named according to the contents of the label to which it corresponds 11 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox Section Two Producing DVDs with subtitles Another way of distributing video materials that have been annotated in ELAN 1s by producing video DVDs with multilingual subtitles Much as with the YouTube captions described in the preceding section viewers of these videos can choose which language if any will be displayed as subtitles during playback Unlike YouTube cf Section 3 or CuPED cf Section 4 however the resulting discs can be played by anyone who has access to a DVD player without requiring a computer an internet connection or access to any particular on line video hosting service This section provides information about how to produce video DVDs with multiple subtitles tracks taken from a
3. 18 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox 5 Reviewing the final DVD If you decided to burn a copy of the DVD to disc then reviewing the final copy is as simple as inserting the disc in a stand alone DVD player or in computer with a DVD drive and appropriate playback software To view the subtitle tracks added earlier select them from within the playback program on a computer or using the Subtitles button on the DVD player remote on a stand alone DVD player Since DVD Flick does not currently provide a preview option which includes the added subtitles you may want to consider creating an ISO image of your DVD prior to burning then loading or mounting this file as the virtual contents of a DVD drive Under Windows Daemon Tools www filehippo com download_daemon_tools is able to load an ISO image as if it were an actual DVD allowing you to test out a DVD image before burning it to disc Likewise Mac users should be able to double click on the ISO image itself and it should load just as a normal video DVD would Some video players like VLC www videolan org vlc are also able to play videos contained in a ISO image directly without the need for services like Daemon Tools 6 Alternatives Although the preceding instructions have concentrated on using DVD Flick under Windows to produce subtitled DVDs there are certainly other DVD authoring packages both Free and
4. you may want one version which just shows translations in English or another which only shows your video in the original language You can export as many different sets of subtitles as you like we can add each set of subtitles to our new YouTube video and then allow users to choose between them Once you have created subtitles text in ELAN for your video clip we can add the subtitles to the new YouTube video In YouTube you can view all of the video clips which you have added to the site by selecting the My Videos link found under your user name in the set of links in the top right hand corner of the screen 27 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox You Tube Canada English 0 3 amp cilldi2009 v Account QuickList 0 Help Sign Out My Videos Home Subscriptions Videos Channels e Upload v Add Remove Modules Playlists a Subscriptions ij Spotlight Sundance Shorts view all edit x a onal Messages nbox ytscreeningroom smeg red with you These award winning short films have two things in common they all played at the Sundance Film E 0 Comments AAA Festival and they are all early works from some of today s hottest indie directors Li Illustration 19 Accessing your videos in YouTube Following this link will bring you to a page which lists all of the videos which you have added to YouTube What we want to do here is edit the new video and add th
5. Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox a8 Export as Subtitles text Export as Subtitles text Select tiers Orthographic Plautdietsch Transcription Orthographic Transcription English Translation Orthographic Plautdietsch ITU Output options Restrict to selected time interval Eil Add master media time offset to annotation times _ Minimal duration per subtitle in ms Eil Override default timecode C PAL C NTSC DF NTSC NDF B E OK Close a Illustration 13 Selecting tiers to export as stl subtitles in ELAN 3 Once you have selected your tiers and a timecode for your subtitles click OK You will then be prompted to select a file name and encoding for the subtitles you ve chosen to export You can save the file wherever you want to make sure that ELAN produces stl subtitles that DVD Studio Pro can read be sure to save your file with the extension stl rather than the default srt e g MySubtitles stl and choose UTF 16 encoding in the drop down selection at the bottom of the window rather than the default UTF 8 Illustration 14 shows one such set of subtitles being saved under the name Aupelkoose stl and with the encoding UTF 16 selected 22 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox aM Export as Subtitles text Save As Aupelkoose stl Plautdietsch Martha_Klas
6. Cox My Account Captions and Subtitles Info amp Settings AudioSwap Annotations Captions and Subtitles W Closed Captions New captions file Choose File PlautCast46 eng srt Track Name optional English Track Language English Upload or cancel Illustration 21 Adding subtitles to a YouTube video Within the Captions and Subtitles section we can now add subtitles to our new video Select the Choose File button and find one of the subtitles text srt files that you made in ELAN YouTube will then give you the option to give a descriptive name to this set of subtitles e g English translation Cree syllabics etc and let you choose what language these subtitles are in Don t worry if your language isn t listed YouTube currently presents a very limited list of languages but we can get around that for now by putting the name of the language of the subtitles in the description instead Once you have these fields entered you can select the Upload button and your subtitles will be sent to YouTube and added to your video If you want to add more than one set of subtitles one for what is being spoken in your language for instance and one for a translation you can repeat the steps above selecting another subtitles text file on your computer and adding it to your new YouTube video as well 4 Viewing your YouTube video with subtitles When you are done adding su
7. click on the Browse button to select a directory into which CuPED should save the final clickable story then click on Process Template to have CuPED begin the conversion AOO CuPED Congratulations CuPED has finished processing your template You can safely close this window or select from one of the following options to continue working with the processed template DA Save a copy of the finished template for later use View the template output using the default application Fr Open the output directory l Customize another template About CuPED r1 Go Back Illustration 26 Reviewing the finished clickable story in CuPED If you are working with a transcript of video material you will note that the preview screen does not show the video itself Although the video will be present in the final clickable story most media require some conversion before they can be displayed in the preview and that processing is usually much too computationally intensive to perform for a quick and dirty preview In other words don t panic Reviewing and sharing a clickable story Once CuPED has finished processing your template you may now access the clickable story it has produced In the final congratulations screen you may click on the View the template output button and CuPED will automatically open a web browser and load your new clickable story or you may simply find the folder where you asked CuPED to save the
8. y dialect between participants in the Elders and Crafters Forum at the Canadian Indigenous Languages and Literacy Development Institute CILLDI 2009 University of Alberta and optionally a set of tags which are keywords which you think might help people find your video e g Cree Nehiyawewin CILLDI If you intend to make your video public for anyone on the internet to see then this information is very important since it will be what most people use to find your video You Tube Canada English 0 amp cilldi2009 Account QuickList 0 Help Sign Out Home Subscriptions Videos Channels Search Video File Upload Record from Webcam Im t Do not upload TV sh gt PlautCast46 mp4 Uploading Cancel a oerras ise aie or of 37 8MB 9 33 ini commercials without permission unless CZ sanam they consist entirely of content you created Add Video Name and Description and Edit Privacy Settings yourself The Copyright Tips page and the Community Title Guidelines can help you determine whether Test vid your video infringes someone else s copyright oe By clicking Upload Video you are Description representing that this video does not violate YouTube s Terms of Use and that you own all This is a test video showing how to create a subtitled YouTube video from ELAN sources copyrights in this video or have authorization to upload it Illustration 18 Uploading a new video to YouTube 26
9. File b Selection 00 00 00 000 00 00 00 000 0 Close sw C Sekcion Mede C Loop M Save a6 5 Save As GHS rr w l S a O O a A a E DERIT REJALA l ENET REFRE ty A l i D Si Ooa a PaE i TE tpt Save as Template F385 00 02 000 00 00 03 000 00 00 04 000 00 00 05 000 00 00 06 000 00 0C Save Selection as eaf Merge Transcriptions Automatic Backup t I l Wm Wal EA s srt A AAS VRAA l Timers lense rst A Sines Vine a S l Terre eE AE Ul E Wine a Sam l SOAS UOA A AE seman D S Sanaa l i en E Sela 00 02 000 00 00 03 000 00 00 04 000 00 00 05 000 00 00 06 000 00 0C Page Setup AEP ative peor J pt Print Preview XP Print qf p two three four five Export As b Shoebox File Export Multiple Files As f Toolbox File UTF 8 COR Import gt CHAT File Se U gt UA Tab delimited Text Tiger xml Interlinear Text HTML Traditional Transcript Text Praat TextGrid List of Words SMIL QuickTime text Subtitles Text Media Clip Image from ELAN Window Filmstrip Image Exit 20 Export as tab delimited text ae Illustration 1 Exporting tab delimited text from within ELAN The tab delimited text format that ELAN produces is almost identical to the format that Audacity uses for its labels containing the start time end time and text of each annotation on some number of tiers To make sure that the two formats are c
10. commercial available for other platforms that are capable of accomplishing much the same task As mentioned in the introduction DVD Styler http www dvdstyler org can be used under Windows and Linux to produce subtitled DVDs although issues have been noted in recent versions with its support for Unicode characters outside of the extended Latin alphabet If your language s orthography only uses letters found in Western European languages you may still be able to use DVD Styler to process your subtitles with the following caveats 1 When exporting your subtitles from ELAN File gt Export As gt Subtitles Text 3 More specifically it does not appear that DVD Styler 1 8 1 is able to use srt subtitles exported from ELAN in anything other than ISO 8859 1 encoding informal tests suggest that neither UTF 8 or UTF 16 subtitles are fully supported in this version of DVD Styler 19 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox be sure to select the encoding ISO 8859 1 rather than the default UTF 8 for your srt subtitles 2 Once ELAN has produced your subtitles they can be loaded into DVD Styler by dragging them onto the thumbnail image of the corresponding video track in the bottom toolbar in DVD Styler Alternatively you can right click on the video track in DVD Styler select Properties from the menu and proceed to add each subtitle track using the plus and minus buttons
11. nakaasdat la ut i yi mikala ii yadzanita And then her husband was suspicious of her Ut i mikala ii yadzani a F all di din dia dinama Pe PD bie td Pd AI Output directory Users chris Desktop cuped f Browse gt Go Back Process Template Illustration 25 Previewing a template in CuPED Once you have added your transcript you should be able to select a tier within it to use as the primary tier in the clickable story The primary tier will be the one displayed most prominently in the final story along with all of its child tiers Click on the Primary tier option in the left hand list of template settings then check one of the checkboxes that appears next to the names of tiers in the Current setting box You can also customize how that tier will be displayed by clicking on the Options button but that step can also be omitted Once all of the required options have been provided you may click on Continue This will take you to a preview of the final template If you want to make changes to any of the display settings selecting better fonts or hiding particular child tiers all of which can be done by clicking on the Options button when a particular tier is selected you can click on the Back button to return to the previous screen where the template can 36 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox be customized further If everything looks alright
12. s possible to customize the font colour and positioning of any text that DVD Flick will show as subtitles To customize a subtitle track click on that track s name as listed in the Subtitle tracks portion of the Properties window for your video then click on the Edit button on the right hand side 15 Aligning text to audio and video using ELAN June 22 July 2 2010 2 DVD Flick Unnamed project New project Su Properties of General Chapters video source Audio tracks Subtitle track IE Open project Save project btitle Language English EN Top margin Bottom margin Left margin cP we ee we Right margin x x 3a w Font Palatino Linotype Character set Latin 1 Size Bold Alignment gt o o O O o a Anti alias Transparent background C Display by default iz Project settings Menu settings A Berez C Cox ro oa Create DYD Guide Update Add title Edit title Remove title Move up Move down Compact list I thought so arddisk space required 36 Mb C 5 1 pixels Text color Outline color 26519 Kb Background color Browse Outline Illustration 9 Customizing subtitle display options in DVD Flick From within the Subtitle window which should appear it 1s possible to tweak various text formatting and positioning options Although the default values are usually more than acceptable this scree
13. story and open the index html file there yourself 37 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox The clickable stories that CuPED produces use much the same technology as YouTube to provide audio and video through a web based interface Unlike YouTube however the CuPED output can be placed on individual computers or on a private web server or burned onto a data CD or data DVD for off line distribution To make the story available to others simply copy the entire output folder s contents to the desired location if placed on a web server the story will immediately be available over the web or if on a memory stick it will be accessible to anyone able to open the index html in the story folder ANO CuPED Flowplayer template a a ro ates f te eens TF rE Lak ats mite er fits ETT A i m k Y DEE tbr E La Fi J F he ETI ey f faa Erai fa s s 7 a j k Ta m ai Eim Til a ik i pa i thaks ihat Pa aa gt i E il am e ae i E a Mg S rat shots hg EER dy ji dzanagu diyi tsika nakaasdat la 4 A long time ago this woman she was flirting around Dzanagu diyi tsika nakaasdat la Illustration 27 A clickable story produced by CuPED In some cases the technology that CuPED uses to play audio and video in a web browser places restrictions on some of the interactive features of the clickable story the individual annotations may not be h
14. with a video clip that we have already saved using MPEG Streamclip and transcribed using ELAN we will choose the first option Upload Video File 0 ea amp cilldi2009 Account QuickList 0 Help Sign Out Search Inload Vic j Inbox Record from Webcam Ia taal Illustration 16 Uploading a video file to YouTube Selecting this option will take you to a screen where you can choose the video file that you want to upload to YouTube YouTube places several restrictions upon the videos that it accepts perhaps the most important of which is that clips are limited to a maximum length of ten minutes If your video clip is longer than ten minutes then you may want to consider software such as CuPED for helping you share your transcribed language resources over the web If your video clip is less than ten minutes long you can add it to YouTube on this page by selecting the Upload Video button You Tube Canada English 0 cilldi2009 v Account QuickList 0 Help Sign Out Home Subscriptions Videos Channels Search Video File Upload Record from Webcam Important Do not upload any TV E Press Upload Video to select and upload a video file Upload video music videos music commercials without maaria hal unless they consist entirely of content you created AutoShare Options 5 yourself Connect accounts and automatically post your uploads to these popular services and websites The Copyri
15. Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox ASA CuPED Customizable Presentation of ELAN Documents Introducing CUPED Customizable Presentation of ELAN Documents About CuPED CuPED Customizable Presentation of ELAN Documents is a tool for transforming time aligned transcripts such as those produced by ELAN into a variety of presentation formats Where ELAN and similar tools commonly concentrate upon the production of transcripts suitable for long term archival preservation CUPED aims to provide a simple and user friendly means of converting these archival sources into formats more readily accessible to a general audience Please note that CuPED is not affiliated in any way with ELAN or the Max Planck Institute l although both the institution and the software it creates through its Language Archiving T Technology initiative deserve acknowledgement for the positive example they set in the Illustration 22 CuPED homepage sweet artsrn ualberta ca cdcox cuped While the CuPED source code is available for use on any operating system where Python is available precompiled installers have been prepared for Windows and Mac OS X Intel PPC users Assuming that you are using either one of these operating systems you may visit the CuPED homepage sweet artsrn ualberta ca cdcox cuped and download and run the appropriate installer which will guide you through the installati
16. Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox Besides allowing you to describe your video clip this page also allows you some basic control over who is allowed access to it In the Privacy section towards the bottom of the page YouTube allows you to specify whether your video should be public and available for anyone in the world to watch or private and visible only to you and a short list of people you choose Depending on the conditions which exist around the video clip you transcribed you may want to let everyone see your video or you may need to restrict access to only a short list of people the choice is up to you Once you have described your video and decided who should be allowed to access it you can select the Save Changes button towards the bottom of the screen YouTube should then tell you that your video settings have been successfully saved Your video clip is now on YouTube it may take a couple minutes to show up in the YouTube search listings if you decided to make your video visible to the general public but it s now on the site 3 Adding subtitles What we want to do now is add subtitles to this video The subtitles text files that YouTube uses for its subtitles are the same ones that ELAN produces when we wanted to add captions to our DVDs You can select which tiers you want to include in your subtitles from within ELAN File gt Export As gt Subtitles Text
17. Aligning text to audio and video using ELAN Institute on Field Linguistics and Language Documentation Workshop June 22 July 2 2010 University of Oregon Andrea Berez aberez umail ucsb edu Christopher Cox christopher cox Qvualberta ca Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox Workshop information 1 Workshop description This workshop is intended to provide a basic hands on introduction to developing multimedia language resources using ELAN concentrating upon the application of this software to language documentation and description Participants will gain experience in time aligned transcription in ELAN as well as in transforming ELAN annotated materials into a variety of output formats using a suite of readily available open source software tools 2 Workshop goals The primary goal of this workshop is to provide participants with an opportunity to 1 Gain a better understanding of the fundamental concepts behind ELAN as well as practical experience applying this knowledge to the annotation of audiovisual language materials 2 Consider how ELAN might be of use in their own language related projects where ELAN might be used to produce standards compliant future friendly language resources or assist in facilitating reuse and dissemination of language materials and weigh these potential benefits against the advantages of other comparable tools and tec
18. ard limit on the amount of information which may be included in your transcripts There are no restrictions on how many speakers can appear in a transcript for instance or on what languages can be represented ELAN is capable of annotating a variety of common audio and video formats works happily with multiple languages and multiple speakers in a single transcript and can import and export its transcripts 1n a variety of common formats Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox 2 What is ELAN not For all of its strengths it should be noted that ELAN does not attempt to serve every function to which time aligned transcripts could be put There are two particular limitations that are worth bearing in mind l 2 ELAN does annotation not editing ELAN concentrates upon adding textual annotations to multimedia resources rather than making any changes to those resources themselves ELAN will never change your media in any way during transcription it can play recordings but it can t edit out silences false starts or any other features you may want to remove before transcription This is arguably a good thing since ELAN has no media editing functions there is much less of a chance of accidentally deleting part of an irreplaceable recording while doing transcription in ELAN ELAN s annotations are plain Unicode text ELAN doesn t provide a built in way of using underlin
19. btitles to your YouTube video you can review your work by watching your video in the normal YouTube video display To access your new video select the My Videos link within the menu under your user name found in the set of links in the top right hand corner of each YouTube page and then select your new video 29 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox Tum on Captions 4 b oo 0022 1001 fs a CS Illustration Turning on captions for your new YouTube video By default no captions are turned on for your video To enable captions click on the arrow in the bottom right hand corner of your video and select the CC option from the menu which appears there This will bring up the list of available subtitle tracks which you have added to your video You can turn individual sets of subtitles on and off from this menu Sometimes it takes a few minutes for newly added subtitles to appear in this menu If your new subtitles aren t showing up right away just wait a couple of minutes and then reload your video and the subtitles should appear You should now see the subtitles appearing underneath your video during playback 5 Limitations l Although YouTube can be a handy tool for distributing language materials there are a number of limitations to the site that should be kept in mind YouTube places a limit on the length and file siz
20. changes Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox eno word_list txt 08 1 64 one 2 2 3 19 two 3 61 4 38 three 4 86 5 65 four 6 05 6 76 five Illustration 3 The contents of an Audacity label file produced by ELAN 2 Automatically segmenting audio in Audacity Now that we have produced a label file containing the times and text contents of the annotations we want to have made into clips we can use Audacity to segment the original recording for us Open the original audio recording in Audacity File gt Open and once it has loaded import the text file containing the labels produced by ELAN File gt Import gt Labels If all goes well you should see something similar to Illustration 4 which shows not only the waveform for your audio recording but also a series of text boxes on a separate label track which represents the contents of the tiers you chose to export from ELAN To make sure that these labels imported properly you can click on the text of any label then click on the green play button in the top left hand toolbar to play the audio associated with that label Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox Ane word_list L L um alale bn M ii mr na 4 ow 24 0 Or 24 0 Pax 4 7 G i z B Internal microphone 1 0 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 I l I I i I I I I I word is
21. e subtitles to it To make changes to a video click on the Edit button found just beneath the little preview picture of the video itself ee PA Uploaded Videos 1 10f1 U ded Videos ae pon Search m a a Favorites Playlists Addto w Delete Subscriptions i Quicklist Sort by Title Time Date Added Views Rating History xe O Test Video Purchases This is a test video showing how to create a subtitled YouTube video from ELAN sources y Added July 23 2009 08 40 AM Time 10 02 What s new TA YouTube Insight TONED namng Find out who is viewing your Comments 0 Responses 0 video where they come from and Broadcast Private Raw File PlautCast46 mp4 more using YouTube s new data feature Insight Live Play Edit Annotations AudioSwap Insight Download MP4 Delete 1 10f1 Illustration 20 Reviewing all of the videos which you have sent to YouTube Editing your video presents a number of options for customizing how your video is presented and distributed to YouTube viewers you can select a location for your video on a world map or choose the thumbnail which will be used as a preview for your video in YouTube search listings or control how your video can be used in other websites For our purposes though we re most interested in the Captions and Subtitles section where we can add our subtitles to the video 28 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C
22. e in ELAN Since ELAN itself doesn t ever make changes to the media files it s given we will have to rely on another software package to extract clips from our audio recording for us In this case we will use Audacity a freely available digital audio editor Audacity has a number of useful features including the ability to automatically extract a series of sound clips from a longer recording on the basis of user defined labels arbitrary bits of text associated with particular start and stop times comparable to ELAN s annotations In order to have Audacity read in the annotations that we want to have extracted we need to convert our ELAN transcript into a format that Audacity can understand As of version 3 9 0 ELAN 1s able to produce the kinds of labels that Audacity expects To get started we need to open an existing ELAN transcript containing annotations that we would like to have extracted This could be anything a transcript of an unscripted conversation or a careful recording of individual words intended for inclusion in a dictionary Once the transcript is loaded we will start by exporting the annotations as tab delimited text File gt Export As gt Tab delimited Text Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox AO Elan word_list eaf aise Edit Annotation Tier Type Search View Options Window Help New Subtitles Audio Recognizer Metadata Open Open Recent
23. e of the materials being uploaded At the time of writing most YouTube videos are not permitted to exceed ten minutes in length or 2GB in size While many language materials in compressed video formats e g MPEG 4 may not have to worry about the file size limit the length limit can prove more troublesome if your recording is longer than ten minutes YouTube will 30 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox 2 most likely reject it out of hand There are certainly ways of working around this limit Perhaps the simplest method is to segment your language material into multiple chunks of ten minutes or less using video editing software e g MPEG Streamclip and then producing multiple sets of subtitles from inside ELAN i e by highlighting the sections of video in your original transcript corresponding to each ten minute chunk and for each one selecting Restrict to selected time interval and unselecting Add master media time offset to annotation times in the Export as Subtitles text dialogue This way a separate srt file is created for each chunk and can be uploaded to YouTube as described above This is a rather time consuming procedure if you have many longer video files to process it may be simpler to use a tool like CuPED instead cf Section 4 YouTube concentrates on distributing video materials rather than audio materials If you want to share language materials via Y
24. ght Tips page and the Community Guidelines can help you determine whether Add services E Facebook t Twitter 3 Google Reader your video infringes someone else s copyright By clicking Upload Video you are renresentina that this viden does not violate Illustration 17 Selecting a video file to add to YouTube The video you choose at this point will be sent to YouTube This stage can be the 25 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox most time consuming especially if you are working on a slower internet connection it will take a while to send your video to YouTube To make the process take less time you should consider sending a compressed version of your video clip e g in mp4 format as can be produced by MPEG Streamclip to YouTube rather than the full original Digital Video DV file which you captured from your video camera If you edited your video in MPEG Streamclip you likely already have your video clip in MPEG 4 format which should be many times smaller than your original video and take much less time to send to YouTube Once you have selected your video file YouTube will begin to receive it and you will be presented with a number of fields in which to describe your video and customize how it will be seen by others on the internet These include the title of your video e g Cree Conversations CILLDI 2009 a brief description e g Conversations in Plains Cree
25. hat should be processed and the primary tier that should be presented in the output Remember the limitation in the current pre release version of CuPED that clickable stories may only have one speaker this will hopefully be improved in later versions to make full use of ELAN s more robust tier system Since these settings are required to produce output you need to provide a transcript and a tier name before the Flowplayer template can do anything else they are marked by default with a star Optional settings things which can be customized but can also be skipped over if you re in a hurry are marked with a heart instead When a setting is currently unavailable or invalid CuPED indicates it with a red icon In Illustration 24 since no transcript file and no primary tier have been selected yet both of their associated icons are red In both cases we need to make a selection in order to proceed Click on the Transcript file option in the left hand list of template 35 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox settings then click the Add button towards the bottom of the window to find an ELAN transcript to process If CuPED is able to load this transcript the icon next to the Transcript file option will turn green 800 CuPED 3 Preview Template Preview dzanagu diyi ts ika nakaasdat la 4 Along time ago this woman she was flirting around Dzanagu diyi ts ika
26. hniques Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox Introduction Introducing ELAN 1 What is ELAN ELAN EUDICO Linguistic Annotator is a software tool commonly used in documentary linguistics to produce time aligned annotations of digital audio and video recordings That is using ELAN it 1s possible to match up arbitrary pieces of text annotations with sections of audio or video media producing documents transcripts which permit fluid navigation between text and media Actively developed since 2002 and made freely available as open source software ELAN is a comparatively mature software tool having grown out of the European Distributed Corpus Project EUDICO and ongoing work in the analysis gesture and sign language at the Max Planck Institute for Psycholinguistics Nijmegen Netherlands Besides this institutional commitment to continued development ELAN has several notable strengths 1 ELAN is free open source multi platform software Pre compiled versions of ELAN are available for Windows Mac and Linux platforms and the entire ELAN source code is freely available for anyone to make changes to 2 ELAN produces non proprietary standards compliant output ELAN transcripts are stored as Unicode based XML conforming to a well documented technical standards thus encouraging long term reuse and archival of these resources 3 ELAN is flexible ELAN sets no h
27. ighlighted automatically as viewers play through the recording and the speaker icons next to each annotation may not be able 38 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox to play back sections of the recording To fix this you can run a small work around program from within the folder where the clickable story is found Versions of this work around program are freely available for Windows Mac OS X and Linux and can be downloaded from sweet artsrn ualberta ca cdcox cuped DVD Workaround zip Simply copy this zip file into the clickable story s folder uncompress it and run the version of the work around program appropriate to your operating system e g on Windows double click on Windows Setup bat on Mac double click on Mac Setup etc 39
28. in the Properties window which appears as in Illustration 12 below AD Properties Title 1 DVDStyler Untitle C Documents and Settings Christopher Cox DesktoptPlat T Duration 00 01 36 Audio aac stereo 44100 AC 46 kHz w TROT EA NTSC 720x480 do not remultiplex transcode Title pater fd pre commands Post commands call mgm menu 1 b Illustration 12 Adding subtitles to a video in DVD Styler For Mac users commercial software such as Apple s own DVD Studio Pro which is bundled with the Final Cut Studio http www apple com finalcutstudio dvdstudiopro 20 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox is also capable of producing video DVDs with multiple subtitles tracks and boasts excellent Unicode character support Unfortunately neither iDVD or iMovie are able to add multiple selectable subtitle tracks to a video being put on DVD For users of DVD Studio Pro a few changes are necessary to the procedure outlined above Since DVD Studio Pro is unable to load srt subtitles we have to have ELAN produce subtitles in another format stl Spruce Subtitles that this DVD authoring software understands As of ELAN 3 9 0 it is possible to produce stl files directly from ELAN through the following procedure 1 With the transcript loaded for which you want to produce subtitles select File gt Export As gt Subtitles Text to bring up the subtit
29. ing or boldface selectively within annotations Again this can be a good thing since it forces you to think about what you mean by using typographical effects You can certainly use other textual conventions to represent changes in pitch volume or voice quality but ELAN doesn t attempt to stipulate what those conventions should be In general however ELAN 1s capable of integrating with other software tools for media editing and redistribution on the whole making these restrictions much less serious The remainder of this document describes methods of using ELAN to help with various common tasks from segmenting words and phrases out of longer recordings Section 1 to distributing recordings on subtitled DVDs Section 2 via YouTube Section 3 or as web based clickable stories Section 4 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox Section One Segmenting audio recordings with Audacity In many language documentation projects it is common to be faced with the task of dividing a single longer recording into many smaller independent pieces This segmentation is often undertaken in order to include audio or video clips in standard Toolbox dictionaries rather than refer to an entire recording you produce many smaller independent audio or video files representing individuals words or phrases that need to be included with particular dictionary entries It goes without say
30. ing that this kind of segmentation can be a time consuming process reviewing an entire recording from within an audio or video editor selecting individual clips extracting them to separate files and assigning these separate files consisting names can take a considerable amount of time and is usually less than wholly engrossing Even once the desired clips have been extracted there are still a number of issues that remain 1 Performing segmentation like this risks losing the association between the original context in which the word or phrase was produced and any later citation of that same material 2 It can be more difficult to assign consistent metadata to a host of small audio or video clips than to one source recording If at some point your project decides to start keeping detailed metadata about the recordings it is working with having to label several hundred small files with the same metadata could end up being quite the inconvenience 1 This is perhaps not the most flexible way of including audiovisual materials in a Toolbox dictionary Margetts 2009 describes an alternative method whereby ELAN transcripts of longer recordings are imported into Toolbox as Toolbox texts and the Toolbox lexicon is then made to refer to individual annotations within those texts with a reference like rf Buluhagalagala_OlDU_OOO8 sound d toolbox_texts Buluhagalagala_O1DU wav 12 066 14 835 This technique avoids many of the problems
31. inherent in segmentation e g losing the association between source context and dictionary citation producing a host of easily disassociated audiovisual clips for the sole purpose of inclusion in a dictionary etc and may be worth considering for Toolbox based dictionary projects See Margetts 2009 http hdl handle net 10125 4426 for details Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox 3 Manual segmentation tends not to add much reusable information to your project The process of manually extracting individual audio or video clips from a recording would not only be difficult for anyone else to repeat exactly but also doesn t generally contribute any information about what s actually being said in those clips which is usually what we re interested in This isn t to say that segmentation isn t necessary at times or shouldn t ever be undertaken there are certainly cases where segmentation may be the most practical option even given the above issues Rather than invest time in manual segmentation however it s also possible to develop reusable transcripts in ELAN and to use the annotations within those transcripts to automatically segment a given recording This section presents instructions on how to automatically extract clips from an ELAN transcript of an audio recording using Audacity http audacity sf net to perform the necessary segmentation 1 Producing an Audacity label fil
32. les text export window 2 In the Export as Subtitles Text window you can select those tiers you wish to export as subtitles just as before At this point however it is also important to select what kind of timecode your subtitles should use unlike srt subtitles which rely on minutes seconds and milliseconds for their timings the stl subtitles that DVD Studio Pro uses refer to actual frames within the video instead Since different countries use different video systems each of which may have a different number of frames being recorded each second we need to tell ELAN what system our video is intended for so that our subtitles stay perfectly in sync ELAN lists three different video systems as a kind of short hand for the number of frames per second in your video PAL NTSC DF and NTSC NDF If your DVD is being made for use in the Americas outside of Brasil Paraguay Uruguay Argentina and French Guiana Japan South Korea Taiwan the Phillippines or Burma your video likely follows NTSC timing and you can choose either NTSC NDF which in informal tests has had the best rate of success with staying in syne with NTSC video in DVD Studio Pro or NTSC DF If your DVD is intended for use anywhere else on earth you re likely using PAL or SECAM encoded sources and you can safely choose PAL in this window Illustration 13 shows this export window with a single tier selected for export and NTSC NDF as the timecode 21
33. n ELAN transcript Although this section concentrates on creating DVDs with multilingual subtitles under Windows it may be possible to adapt these procedures to DVD authoring software for Mac OS X and Linux as well 1 Installing DVD authoring software In order to create a video DVD that can be played on a stand alone DVD player you will need to install DVD authoring software Authoring software helps to automate the task of importing captioning and converting videos into the MPEG 2 format expected on standard video DVDs and may offer facilities for creating interactive menus as well For Windows users there are several freely available DVD authoring packages which are capable of producing subtitled discs In the instructions that follow we concentrate on using DVD Flick www dvdflick net for all DVD authoring since it provides a fairly straightforward interface to basic DVD functions and excellent support for customizing how different sets of subtitles will be displayed Another possibility would be to use DVD Styler www dvdstyler de which currently provides many more options than DVD Flick for customizing menus and graphics although its support for Unicode subtitles seems to be somewhat less robust 12 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox fi Setup DVD Flick Welcome to the DYO Flick Setup Wizard This will Install OY DF lick 1 3 0 7 on your computer lt iz recomme
34. n can be quite useful when subtitle text needs to be displayed somewhere else on the screen to avoid occluding important parts of the original video Once you have made the changes you want to this screen click on the Accept button to return to the Properties window for your video 4 Authoring and burning a DVD in DVD Flick Once you have added all the videos and subtitle tracks that you want you can choose where you would like to save your project by clicking on the Save Project button in the main DVD Flick window You can also customize the menus which viewers will see when they play the DVD by clicking on the Menu Settings button at this point but that is outside of the scope of this description By default DVD Flick will produce a DVD that launches directly into playing the video when started with no menu screen 16 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox We are now ready to compile or author the final DVD from our video sources and subtitle tracks into a format that can be burned to disc This authoring process can take some time especially when DVD Flick needs to convert much of the video into the MPEG 2 format used on video DVDs If you only intend to make a single copy of this video you can insert a blank DVD R disc at this point and burn away If you intend to make multiple copies of your DVD however it may be better to save the results of the authoring proce
35. nded that you close all other applications before Continuing Click Next to continue or Cancel to exit Setup Illustration 6 Installing DVD Flick To install DVD Flick download and run the installer from the DVD Flick website www dvdflick net download php Follow the prompts which the installer provides and DVD Flick should install itself on your computer adding an icon to your desktop and a folder in your Start Menu Once installation is complete you can start DVD Flick with the icons in either of these locations 2 Adding a video to DVD Flick To add a new video to a DVD simply drag and drop the video file that you have annotated in your ELAN transcript onto the DVD Flick window or click on the Add Title button on the right hand side of the DVD Flick window DVD Flick should be able to load most if not all of the same formats that ELAN is able to annotate under Windows if all goes well DVD Flick should display the video you have added in its main window along with information about its file name duration and associated audio and subtitles tracks 13 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox E DVD Flick Unnamed project 5 3 3 New project Open project Save project Project settings Menu settings Create DYD Guide About Update PlautCast46 Add title C Documents and Settings Christopher Cox Desktop PlautCast46 PlautCast46 mp4 Duration 10 01 minutes Edit
36. ning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox hand list By selecting the Create ISO image checkbox and clicking on the Accept button DVD Flick will produce a standard iso file for this DVD rather than burning it to disc immediately Encoding Unnamed project DVD Flick Abort Prepare files Minimize bo tray Encode video Encode audio Entertain me Combine sources Add subtitles Author OVD When done Finalize l Bo nothing k ae Title 1 of 1 video 1 of 1 Time elapsed Encoding 8 seconds Illustration 11 Encoding a DVD in DVD Flick Once you are ready to create your DVD click on the Create DVD button in the main DVD Flick window This will begin the DVD authoring process with DVD Flick automatically converting your video s and subtitles into formats needed and either burning or saving the final compilation to an iso file for you Again this may take a while to finish up if you want to go get a coffee this would probably be a good time to do so Alternatively DVD Flick also includes a built in Tetris game for those desperately in need of a diversion 2 Within the Project settings window it is also possible to change the video system i e NTSC for most of North and South America Burma Korea Japan and several other countries and PAL for most of Europe Africa Australia and Oceania southern Asia and Brazil with which this DVD will be compatible
37. o using ELAN June 22 July 2 2010 A Berez C Cox subtitles text file that should be included DVD Flick Unnamed project a Pe J E New project Open project Save project Project settings Menu settings Create DYD About Update PlautCast46 Add title Edit title S Remove title Move up Properties of title 1 PlautCast46 Move down Add General C Documents and Settings Christopher Cox Desktop PlautCast46 PlautCast46 eng srt Format SubRip Chapters Language English a Edit 9 Compact list Total duration Remove 10 01 minutes 1 titles video sources Audio tracks Alt bitrate 9000 Kbitis Move up Harddisk space required 1686 Mb Move down 1726519 Kb Browse lt Previous title Next title gt Illustration 8 Adding subtitles to a video in DVD Flick We can repeat this procedure to add multiple sets of subtitles to the video each of the sets of subtitles that we add here will be accessible to the viewer on the final DVD as separate subtitle tracks Again if we had a set of subtitles in the original language e g PlautCast46 Original srt and a set of subtitles that represent translations e g PlautCast46 Translation srt we could add them both at this point and they would both be available to viewers later on Not only does DVD Flick permit us to have multiple subtitles tracks but it also allows us to change how and where each of these sets of subtitles will be displayed It
38. ompatible we need to 1 Select the tiers whose annotation we would like to have extracted 2 Select Exclude tier names from output 3 Under Include time column for select only Begin Time and End Time 4 Under Include time format select only ss msec Aligning text to audio and video using ELAN June 22 July 2 2010 AOO Export as tab delimited text Export tier s as tab delimited text Select tiers Tsuut ina Eee Notes Al _ Output options __ Restrict to selected time interval __ Add master media time offset to annotation times vi Exclude tier names from output _ Separate column for each tier Include time column for Include time format M Begin Time hh mm ss ms M End Time v ss msec Duration msec _ SMPTE Timecode hh mm ss ff N ISC taroD Trame S ae meme OK Close Illustration 2 Exporting Audacity labels in ELAN A Berez C Cox Once this is done click on OK and save the resulting text file This file now contains the text of all of the annotations on the tiers that you selected presented on separate lines and preceded by the start and end time of each annotation in milliseconds This is all the information that Audacity needs to extract audio clips automatically if you want to make changes to this file e g doing a search and replace to get rid of particular characters you can do so now and Audacity will pick up these
39. on process or on a Mac let you drag CuPED to your Applications folder 2 Applying CuPED to ELAN transcripts Once you have downloaded and installed CuPED you can begin applying CuPED to your ELAN transcripts The CuPED user interface has been made to resemble a software wizard a set of dialogue windows which gently guide users step by step through the process of converting their transcripts Double click on the CuPED icon to start the application If all goes well you will be presented with the CuPED welcome screen As elsewhere in the CuPED interface you can advance to the next dialogue by clicking on the Continue button at the bottom 33 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox of the CuPED window By advancing past the welcome screen CuPED will present you with a list of output templates that you can apply to your transcript ASO CuPED 1 Choose Template Please select a CuPED template from the list below or add a template to the list using the Add template button The template you select will be used later to process your ELAN transcripts l panii Basic Fempiate A template to test out basic CUPED functions filled in CuPED Basic Template 5 filled in testing A template to test out basic CuPED functions 3 CuPED Basic Template A template to test out basic CuPED functions fe CuPED Flowplayer lA template which produces HTML JS Flash output u
40. oned YouTube videos using an ELAN transcript of a short video clip as our starting point 1 Creating a YouTube account In order to add your own videos to YouTube you first need to set up a YouTube account If you don t already have a YouTube account you can create one by following the Sign Up link found in the set of links located in the top right hand corner of the YouTube homepage www youtube com i YOuaTIT canada English Sign Up QuickList 0 Help Sign In Home Videos Channels Community Search Illustration 15 The header of the YouTube homepage YouTube will then guide you through the process of choosing a user name and password as well as configuring your preferences for use of the site You will only need to sign up for an account once when you have your new YouTube account set up you will be able to log into it by following the Sign In link which is also found in the top right hand corner of the YouTube homepage 2 Adding a video to YouTube Once you have created a YouTube account you can begin adding your own videos to the site To add a new video select the gold Upload button again found in the upper 24 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox right hand corner of the YouTube homepage This will present you with two options one to add a video from a file on your computer and another to create a video using a webcam Since we will be working
41. ouTube that only include audio you will need to convert your audio recording into a YouTube compatible video format first Thankfully there are a number of tools that can help you to accomplish this task o Under Windows and Mac OS X It s possible to load still images and audio tracks into Windows Movie Maker under Windows or iMovie under Mac OS X and to save the resulting image audio combination as a wmv or mp4 video that can then be uploaded to YouTube Although an introduction to either piece of software is outside of the scope of this discussion instructions are available online e g http www youtube com watch v wFQUWa6cN14 o Under Linux Unix The common multimedia processing tool ffmpeg http ffmpeg org can be used to turn an audio file and a still image into a video that can be uploaded to YouTube Assuming your audio file is audio_file wav your still image is picture jpg and you want the final video to be called output flv the corresponding command to invoke ffmpeg would be as follows all on one line ffmpeg f imagee loop_input i picture jpg aspect 4 5 i audio_file wav ar 44100 ac 2 ab 192k shortest output filv 31 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox Section Four Producing clickable stories with CuPED Another way of making materials that have been transcribed in ELAN accessible to a broader audience is by producing so called clickable s
42. sing Template Flowplayer to interact with media Note Adobe Flash Play Add template Remove template O Details Pa hy i z k Go Back Continue e O a Illustration 23 Selecting a template in CuPED You can think of templates as providing different ways of presenting your transcript each template sets out one way of transforming the transcripts it is given into a particular output format Even though CuPED currently only produces clickable stories at the moment it s possible to imagine additional templates that would transform an ELAN transcript and its associated media into say a printable Word document or a LaTeX manuscript or any number of other output formats To access the clickable story functions available in the current beta version of CuPED select the Flowplayer template then click on Continue 34 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox AO CuPED 2 Customize Template Template settings Description Transcript file The transcript file to be turned into a web ce Frimary tier page 2 Convert media J Title of page Current setting ada Edit Remove GoBack Continue a ae Illustration 24 Customizing the Flowplayer template in CuPED CuPED allows users to customize certain settings for the templates they select In the case of the Flowplayer template these settings concern the name of the ELAN transcript t
43. ss as an ISO disc image 1 e an iso file which preserves the contents of an entire DVD ina standardized file format That way if you decide to make more copies of this DVD you don t have to wait for all of the video to convert again the entire DVD will be ready to go and can be burned to disc with any DVD burning software that can read an iso file even on a computer or operating system where DVD Flick is not available DVD Flick Unnamed project SS S Gy G New project Open project Save project Project settings Menu settings Create DYD Guide About Update PlautCast46 Add title C Documents and Settings Christopher Cox Desktop PlautCast46 PlautCast46 mp4 Duration 10 01 minutes z Edit title 1 audio track s 2 subtitle s SS Remove title Project settings Move up General Create ISO image Move down video Burn project to disc Audio 9 Compact list Disc label PlautCast 46 Total duration Playback 10 01 minutes Drive 1 titles Awerage bitrate g 3000 Kbitis ere Harddisk space required Automatically erase disc if it is rewritable 1932 Mb 1979338 Kb Delete 150 image after burning Verify disc after burning rOWSe Eject tray when done Illustration 10 Changing project settings to create an ISO image in DVD Flick To create an ISO image for your DVD Flick project click on the Project Settings button in the main DVD Flick window then select the Burning option from the left 17 Alig
44. t F Mono 44100Hz 24 bit PCM ajae Tack Le i T Project Rate Hz Selection Start End C Length Audio Position 44100 C Snap To 00 h 00 m 00 880 s 00 h 00 m 01 640 s 00h00m 00 000 s7 Click and drag to resize the track Actual Rate 44100 F Illustration 4 An audio recording and ELAN produced labels in Audacity Now that the original recording and the labels have been loaded we are ready to have Audacity segment our audio for us From the File menu select Export Multiple This will bring up a small dialogue box which allows us to specify where the audio clips should be saved Export location what format they should appear in Export format and how the resulting clips should be named Although many other options are available to us we will choose to export our audio clips here as regular WAV files WAV Microsoft signed 16 bit PCM into a separate folder from the one where our original recording is stored using the labels to define the boundaries of our audio clips Split files based on Labels and using label names as the names of our clips Name files Using Label Track Names as in Illustration 5 Once these settings have been made click on the Export button to have Audacity begin extracting audio clips 10 Aligning text to audio and video using ELAN June 22 July 2 2010 A Berez C Cox ANON Export Multiple Export format WAV Microsoft signed 16 bit PCM ETIESE Options
45. title 1 audio track s 0 subtitle s Remove title Move up Move down Compact list Total duration 10 01 minutes 1 titles Average bitrate 9000 Kbit s Harddisk space required 1686 Mb 1726519 Kb Project destination folder C Documents and Settings Christopher Cox Desktop new_dvd Illustration 7 Adding a new video to DVD Flick 3 Adding and customizing subtitles in DVD Flick To add subtitles to this new video we will first need to export each of the tiers we wish to include as subtitles text via File gt Export As gt Subtitles Text just as we did to produce captions for YouTube If for example we wanted to have subtitles available both in the language of the recording and in the language of any translations we have provided in ELAN we could export two separate subtitles text files one with just the tiers in the original language e g PlautCast46 Original srt and another with just the tiers of the translation e g PlautCast46 Translation srt DVD Flick is able to associate multiple subtitles tracks with a single video and permits us to customize how and where these subtitles are displayed To add subtitles to a video select the video in the DVD Flick window then click on the Edit Title button on the right hand side In the Properties window which appears select the Subtitle Tracks option on the left hand side then click on the Add button to locate the 14 Aligning text to audio and vide
46. tories using a tool called CuPED Whereas YouTube concentrates on delivering video on line and imposes the restriction that all video clips must be ten minutes in length or less CuPED allows for both audio and video materials to be made available in a web browser whether on a local computer e g on CD on a laptop on a library workstation etc or over the internet and places no limit on the length of the materials being presented Although the resulting presentation requires a web browser to view it provides a more interactive interface than either YouTube or a standard video DVD and may thus be more interesting for some audiences This section presents instructions on how to apply CuPED to an ELAN transcript and its associated recording s to produce a clickable story which 1s ready for use on the web or on a local computer For those who would like further information on how to use CuPED a user manual is also available from the CuPED homepage see below 1 Installing CuPED CuPED 1s a free open source software tool which provides a user friendly interface to the media and transcript conversion facilities commonly needed to prepare ELAN annotated audio or video for distribution over the web Although CuPED is still in early development it is capable at the moment of turning most simple single speaker transcripts into clickable stories with plans to handle multiple speakers and other output formats in the near future 32
Download Pdf Manuals
Related Search