The projects I have been working on as a volunteer for Divine Truth and God’s Way involve creating computer programs to automate tasks that would otherwise be time and labour intensive to perform. As data consistency is essential to computer programs running successfully, I am also involved with developing standards for information being produced within the organisations.
Though this work was set in motion years earlier, I came on board as a volunteer in September 2016 and began to work on the development of projects which had already been planned for by Jesus, but had not yet been undertaken due to the limits of his time and lack of available volunteers with the required technical skills. Jesus now manages and oversees the projects, analyses organisational workflows and develops long term plans, while I assist in realising the plans through developing the required code and computer applications.
Our aim is to support the Divine Truth and God’s Way teams by automating time consuming tasks where possible, allowing volunteers to focus their time and energy on core activities relating to creation and sharing of Divine Truth information, which require special skills and cannot be automated.
Much of the information generated within the organisations is in the form of Microsoft Word documents, and under Jesus’ guidance, I have been developing customised computer programs to reduce the time involved in repetitive document processing tasks like formatting, editing and modifying style, format and structure in order to create other types of documents based on the same content.
Volunteers involved in document creation are being trained to produce standardised documents which are of a high quality, accuracy and consistency, and this enables the automated transformation of these initial documents into other formats. Source documents like Presentation Outlines and Transcripts can now be converted into a variety of forms including eBooks, printable booklets, subtitles for videos and computer scripts used to automate tasks like video clipping – all with very little human effort and in a fraction of the time that it has previously taken for a person to produce these outputs manually.
We aspire towards many more future outcomes for increased productivity and standardisation of processes, and I desire to continue supporting Divine Truth & God’s Way in bringing more economy to everyday tasks and enabling volunteer time and effort to focus on the key activities that allow God’s Truth to be freely shared worldwide.
Donations support me to continue this work, and I’m extremely grateful for the gifts I have received. I love these projects, the opportunity to work with this team and to contribute towards organisations close to my heart. Since October 2017, this has been my primary work and source of income.
The Transcript documents, produced within the Records Team managed by Barb, contain all the information required for Video Subtitling: the written transcription of the presentation, time codes at each paragraph in the document which correspond to the video recording, non-spoken information which is required for language translation (eg: the presenter points to the word ‘Love’ on the whiteboard) and summary of parts of the presentation which are not directly relevant to the topic (eg: the presenter gives information about where to find bathrooms at the venue of the presentation). This allows for Video Subtitles to be automatically generated from a Transcript document, based on a series of logical rules encapsulated in a computer program.
The program for converting a Transcript into Subtitles takes into account a number of factors to create subtitles optimised for reading. This includes splitting text from one subtitle to the next at punctuation in text where possible, creating subtitles which are consistent in length, separating subtitles evenly over two lines and ensuring that each subtitle appears on screen for a sufficient time frame.
Using these files, subtitles can be made available on YouTube videos, as well as loaded into a media player when watching a video offline. This provides accessibility for non-English speakers and those with hearing impairment.
Prior to using this automation, it has taken volunteers from the Records Team approximately 40 hours to convert a Presentation Transcript into Subtitles, which involved manually splitting the text up into small portions of an appropriate length, sourcing and adding the additional time codes for each individual subtitle and then running the data through another program in order to convert the document into the required SRT file format. This time and effort then needs to be repeated for each language the Transcript has been translated into.
Using the programmed automation, Subtitles for a Transcript or Translation are generated in under 1 minute per document.
Computer generated Subtitles are currently being tested and refined for the English language, and the project will reach its full potential in future, when used to create subtitles from Translated Transcripts.
Further programs have been created to automate the conversion of Transcripts into various Book formats.
Previously, each book format represented more manual effort for volunteers of the Records Team and took around 4 hours per book to produce. By automating the re-formatting required, it now takes just minutes to convert a Transcript document into each type of book.
A Publishing Format Book is a version of a Transcript intended for reading and printing, which is made available in digital format as a PDF document. Formatting changes are made to the Transcript to improve readability: time codes and off-topic content are removed from the document and the small paragraphs of the Transcript are combined to make fewer paragraphs of longer, even lengths.
Self-Print Booklets are a format of a Transcript which a reader can download, print and bind into an A5 booklet, using their own home printer or a professional service. The automated process adds cover pages to the Transcript, and increases the font size of text throughout to ensure comfortable reading on a smaller format page.
The eBook format is optimised for reading on a digital eBook reader device, such as the Amazon Kindle. Formatting is simplified to comply with eBook standards, time codes are removed, a title page is added, font size is increased and headings within the document are transformed into clickable links to the table of contents page, which also becomes a list of chapter links that allow the reader to navigate to the desired section of the book.
A video clip is a short video extracted from a full length presentation. Presentations are of 1-2 hours in length on average, and cover a more general topic, where a clip will be a shorter in length and focused on a specific topic. Videos can be viewed on a website such as YouTube, as well as offline on a personal hard disk drive.
The benefit of being able to view clips as well as extended presentations is that clips can provide more immediate answers on a specific topic, and being shorter to view, can be a more practical way of receiving information. Keywords in the name of the clip allow for easy searching of content, both online and offline on a local file system. For example, a search using the keyword ‘cancer’ will show all clips on this topic, which may have been clipped from a number of different presentations, and information is quickly accessible.
A Transcript or Outline document contains all the details required for cutting a full length presentation into a series of shorter clips. Headings within the document become titles for clips, heading indentation levels allow the creation of clips and sub-clips which are increasingly specific in topic, and time codes determine the start and end times for the clip and points at which cuts need to be made in the full length source video.
A number of computer programmed processes have been developed and used in combination to achieve the result of automating video clipping. While a Transcript or Outline document can be used, the benefit of generating clips from an Outline is that this document is created early in the production process, where the Transcript cannot begin until after footage of the presentation has been produced by Lena in the Production Team. By clipping from an Outline, clips can be made publicly available at the same time the extended presentation is published.
Previously, the process of video clipping was performed independently from Outline and Transcript documentation. A volunteer performing video clipping would need to review footage of the full length presentation, decide where to create cuts and come up with an appropriate title for each clip. A Transcript of the presentation would often not be completed at this time, and in the past, Outline documents may have either not existed or lacked detail that corresponded closely with the video, so volunteers creating clips did not have the benefit of the headings and time codes from a document to guide them. As this work was somewhat subjective, the clips would then have to be reviewed and checked by another volunteer. Specialised video software was also required to perform video clipping. Taking into account all steps, this process took around 4 hours per clip produced, and this time and effort quickly compounded, as each extended presentation of 1-2 hours can be broken into around 10 clips at minimum.
Under the new process, Jesus & Mary do the work upfront to create a thorough and well planned Outline document, prior to a discussion or presentation being given. An automated process transforms the Outline into a format which can be loaded into an iPad app, and during live filming, Production Team volunteers Lena or Eloisa use the app to tag the times at which discussion of each topic in the Outline begins. After filming has taken place and video editing is complete, time codes linking topics in the original Outline document to the video are imported back into the Outline via another automated process.
From here, a computer script for cutting the full length video into a series of clips can be generated from the Outline document. Free software called FFmpeg is used to automate the task of clipping, and the whole process requires very little human input. A series of commands required for the FFmpeg program are created from a completed Outline in less than 1 minute, and using this script, each clip takes approximately 2 seconds to be cut from the full length video and saved as a separate video file. A volunteer just needs to generate and start the script on a computer or server, and does not need to be present while it runs.
Additional computer programs have been created for generating other scripts to automate processes like uploading video files and associated data onto video-sharing websites YouTube and Vimeo.
Computer scripts can be used to automate repetitive tasks and reduce the time and effort required to perform data and file handling manually. As anyone who has managed a YouTube channel will know, uploading and setting up details for each of your videos is time consuming, and keep in mind that a single day of filming for Divine Truth can result in several extended videos and dozens of clips which all need to be uploaded and configured with descriptions and details to allow for keyword searching.
Jesus currently manages the Divine Truth YouTube channel, and he will soon be able to generate video upload scripts from Outline or Transcript documents with just a few button clicks, and then be able to sit back and relax while the script runs and does the work. ☻