SuperMemo: Incremental video

1 What is incremental video?
- 1.1 Incremental video with YouTube
- 1.2 Incremental video with video files
2 Incremental video in SuperMemo
- 2.1 Incremental video with YouTube: Outline
- 2.2 Incremental video with video files: Outline
3 Hints: Incremental video

What is incremental video?

Incremental video is a technique for learning or watching video recordings with the view to forming lasting memories of the material viewed. Incremental video makes it possible to learn from multiple sources of video without neglecting any one of them. The processing of individual sources proceeds in parallel and time allocations for individual sources depend on student's priorities. Incremental video is to video as incremental reading is to learning from electronic texts. In incremental video, it is the user who decides which portions of a video are important to remember. Those portions are part of a standard learning process known from other applications of SuperMemo. Each portion of the video forms a separate topic that is reviewed at increasing intervals. Despite working with thousands of videos, you do not ever need to feel lost, ever get bored, or ever miss a valuable scene. You can also use SuperMemo as your video or music jukebox. You can work with individual videos for mere seconds. Just as much as is needed to set the new viewing point, determine the priority and determine the date of the next viewing. This way, you can process dozens of videos daily and work with thousands of videos in parallel. You shall feel overwhelmed with the richness of YouTube never again!

On the face of it, incremental video looks like channel-zapping on steroids. However, it is also a powerful learning technique that makes it easy to process thousands of videos in parallel without getting lost. It also makes it easy to learn individual video scenes for long-term retention. It can be used to learn sports, master musical instruments, understand biology, or learn fun dialogs in your favorite movies. Individual videos are processed in small portions. Viewing can be resumed at any time at the last viewing position. Best pieces are preserved for repeat viewing. Individual videos are prioritized and served on a daily basis in manageable portions in order of priority.

Although, incremental video can roughly be implemented with a pair of digital video recorders working in tandem, only SuperMemo provides a bona fide implementation with the whole set of incremental learning tools. SuperMemo makes it possible to learn from standard video files (e.g. MP4). It also makes it easy to learn from YouTube videos. At the moment of writing these words, no other software in the world provides the incremental video learning toolset.

Figure: Exemplary screenshot from an incremental video process in SuperMemo.

Incremental video with YouTube

Incremental video in SuperMemo can use YouTube videos. You can capitalize on the video streaming power of YouTube and speed up SuperMemo even though you may suffer video lags on weaker Internet connections. You will also dramatically save on hard disk space. Videos are notorious hogs of space, your YouTube collections will take a tiny fraction of space needed for video based on local files. YouTube collection will actually take less space than incremental reading collections while carrying lots of learning power. If your network is slow or frequently down, you should take comfort in the fact that this situation is likely to change for the better as networks improve worldwide all the time (and this progress is not likely to be slowed soon). Last but not least, SuperMemo will hopefully encourage you to upload your own educational videos to YouTube and thus share them with others.

Incremental video with video files

Incremental video in SuperMemo can use video files from your local drive. You do not need to limit your learning to videos available from YouTube that can be notoriously volatile (e.g. with accounts closed for copyright violations, embedding blocked by account holders, videos made private, etc.). You are not limited by the need to access the Internet. You can also process private videos.

Using incremental video with local files requires computers with significant processing power. On a strong PC, processing video files is fast and painless. It may be slower on weaker PCs. Your collections may grow to be terabytes large, which is the main disadvantage of keeping files local. Make sure you have a big dedicated external drive for a backup of your material. The time-scale resolution of local videos is higher than that of YouTube-based videos, which are limited to setting Start and Stop points in increments of 1 second. With local files, you can review fragments that last milliseconds. You can review your best football move at nauseam. Moreover, with local files, you can work when disconnected from the net.

Incremental video can use local files in several video formats (incl. MP4, WMV, AVI, MOV, Mpeg, and more). Not all video formats are supported. Before you commit to SuperMemo, give it a try to see if your files can be processed or converted to acceptable format with third party tools.

Incremental video in SuperMemo

SuperMemo is a pioneer in implementing versatile incremental video. For those who are familiar with incremental reading (also pioneered by SuperMemo), incremental video is an analogous technique. Instead of text extracts, you generate video extracts, i.e. portions of a larger video. Video extracts are viewed repeatedly in increasing intervals (as it is the case with other pieces of information in SuperMemo).

Individual videos and their extracts are treated in the same way as all other topic elements and enter the learning process according to the rules that are known from incremental reading. Currently, only passive review of extracts is supported. There is no equivalent of cloze deletion in incremental video. However, you can use videos as answers if you choose so.

Incremental video with YouTube: Outline

This is how you work with incremental video in SuperMemo:

open your favorite YouTube videos in Internet Explorer
choose YouTube import option to import videos to SuperMemo (e.g. Shift+Ctrl+Y)
use Learn to process individual videos
use Start and Stop buttons to mark interesting scenes
use Extract to generate new elements with scenes marked with Start and Stop
use learning tools in SuperMemo to prioritize, schedule, and organize videos and video extracts

Figure: An extract from the "Pirates of Silicon Valley" (yellow in the template is used to differentiate between extracts and parent videos). SuperMemo will play the fragment between the start time of 8 min. 43 sec. and the stop time of 9 min 8 sec. (of the original YouTube video). The checkmark near the Test button indicates that the fragment should be played in a loop. On the right, you can see the description of the video imported from the YouTube database. In pink, you can see references generated automatically (when importing videos with Edit : Import web pages : YouTube (Shift+Ctrl+Y)). Both the video description and the references are propagated from the original video element to all video extracts.

Importing videos from YouTube

To import videos for incremental learning do the following:

open YouTube videos in Internet Explorer
choose one of the following:
- press Shift+Ctrl+Y
- Import YouTube videos on the learnbar (right-click over the Import articles button)
- Edit : Import web pages : YouTube from the main menu
optionally, set import options (e.g. which videos to import, video priority, name of the import node in the knowledge tree, etc.)
click Import

If you see a video embedded in a webpage choose Copy video URL on the YouTube video context menu (right click), and paste the video to SuperMemo with Ctrl+N (as you would paste anywebpage).

If you receive a video forward from SuperMemo, you can copy its code from the body of the mail and paste it to your collection with Ctrl+N (as you would paste any text or article).

Viewing YouTube videos

Use Learn in the same way as when learning with SuperMemo. The videos start automatically. When you get bored or need to watch other videos, press Start to mark the point from which you will resume the video next time you see it. If you find an important fragment that you would like to learn or view again, press Start at the beginning of the fragment, and Stop at the end of the fragment. Use Test to view the fragment again. Press Extract if you would like to create a new fragment that should take part in the learning process as a separate element.

You can use Mark and Resume to set a bookmark that will not affect the point from which video starts (e.g. when preparing an extract).

You can generate extracts without interrupting the viewing process. Extract elements are generated only when you move on to the next element or when you press Alt+X.

If you get the same videos over and over again, and you would like to get some variety independent of video priority, use Learn : Sorting : Sorting criteria and increase Randomization (move the thumb from Prioritized topics towards Randomized topics, i.e. to the right).

If you are not sure how to handle the incremental learning process, incremental reading is a good introduction to understanding incremental video.

Forwarding video fragments to friends

If you forward a YouTube element in SuperMemo via e-mail to others, it will include the link and start:stop boundaries of the video. Forwarded videos can be viewed in a web browser or directly in SuperMemo with Start and Stop buttons set to make sure only the recommended fragment plays.

Your e-mail will look similar to this one:

Please have a look at this YouTube video:http://www.youtube.com/watch?v=WOzNAiOR7fY?rel=0&t=49&end=79See the fragment from 00:49 to 01:19.If you are using SuperMemo 15 or later you can also:(1) select this code (e.g. triple click):{SuperMemoYouTube:WOzNAiOR7fY,00:49,01:19,00:00}(2) copy the code to clipboard (e.g. Ctrl+C)(3) use Ctrl+N to paste the code to SuperMemo and play the recommended fragment.Sender ID: SuperMemo Research==================================================#Subject: Marissa Mayer: Why Google Maps is far ahead on Android#Author: Jean-Louis Nguyen#Date: Feb 27, 2013, Wed, 00:18 (updated: Feb 27, 2013, Wed, 00:18)#Source: YouTube#Comment: Uploaded 2012-09-27 (2669 views)#Link: http://www.youtube.com/watch?v=WOzNAiOR7fY#Collection: YOUTUBE [Element=6667]#Generated: Aug 01, 2013, Thu, 12:13:00#Software: SuperMemo 17 (Build 17.0, Apr 01, 2013)

Exemplary YouTube videos

Incremental video is a video equivalent of incremental reading. However, there are many things you may wish to learn that are best mastered with video and cannot be substituted with reading. The list is truly endless. However, here are just a few examples taken from YouTube to give you the first sense of why incremental video is an important component of incremental learning:

learn a foreign language with the help of movies, speeches, lectures, etc.
learn recent history with archive footage
explore the student's video paradise: Khan Academy
learn to play a musical instrument using video tutorial produced by experts or best artists
learn to play along or sing along your favorite songs
use Mike Phelps videos to learn how to swim butterfly
recover from injuries by learning physiotherapeutic exercises
watch movies incrementally
listen to your favorite music
watch historic sports events
watch video lectures from reputable universities
enjoy or learn the best jokes or comedy pieces by Leno, Connan, Jon Stewart, and others
master speechmaking with videos of MLK, JFK, Reagan, or Obama
seek motivation from Tony Robbins or other self-help gurus
learn to cook
relive memorable moments in the lives of your family and friends
learn physiology, geology, physics, or history, wherever a video explanation is needed
learn sign language, volleyball, soccer, etc.

Incremental video with video files: Outline

This is how you work with video files using incremental video in SuperMemo:

put your video files into a single empty folder
use File : Import : Files and folders to import the contents of that folder to SuperMemo
use Learn to process individual videos
use Start and End buttons to mark interesting scenes
use Extract to generate new elements with scenes marked with Start and End
use learning tools of SuperMemo to prioritize, schedule, and organize videos and video extracts

Figure: An incremental video extract from the 2013 men's singles US Open final between Novak Djokovic and Rafael Nadal (yellow background in the template is used to differentiate between extracts and parent videos). On the right, you can see the import parameters of the video, which should ideally be converted to annotations that will help you locate the video and assist in the training (or learning). In the picture, the 9th game into the 3rd set, the Spanish challenger, down love 40, defends the first of the three consecutive break points with a topspin serve into the ad court to win the point with his trademark forehand down the line upon the Serb's otherwise excellent aggressive return.

Video deletion

Incremental video is currently most suited for processing video or audio material passively with the most interesting portions extracted for passive review as topics. Currently you cannot "cloze delete" portions of video or audio files with a keystroke, however, you can easily use video extracts as answers to text questions, or video questions. For that purpose, mark your sound or video components with the Answer attribute. You can also define templates that will make that process easier.

Viewing video files and learning

Use Learn in the same way as when learning with SuperMemo. The videos start automatically if you have AutoPlay checked on the element menu. When you get bored or need to watch other videos, press Start (on the Extractor panel). Start marks the point from which you will resume the video next time you see it. If you find an important fragment that you would like to learn or view again, press Start at the beginning of the fragment, and End at the end of the fragment. Use Test to view the fragment again. Press Extract if you would like to create a new element with the extracted fragment that should take part in the learning process.

If you get the same videos over and over again, and you would like to get some variety (even though the videos might be of top priority), use Learn : Sorting : Sorting criteria and increase Randomization (move the thumb from Prioritized topics towards Randomized topics).

If you are not sure how to handle the incremental learning process, incremental reading is a good introduction to understanding incremental video.

Exemplary video files

For a list of exemplary videos that may be used in incremental learning see: Exemplary YouTube videos.

Note that not all video material may be learned with the help of YouTube. There might be copyright or privacy issues. There is also one pesky problem with YouTube that can ruin any long-term learning process: videos can be pulled at any minute (unless you upload them yourself for public viewing with full respect to copyright). Last but not least, the time resolution of local file incremental video is higher that that of YouTube-based video. You can set your starting point with millisecond precision (as opposed to 1 second resolution of YouTube videos).

Material that might work better with files on your local hard disk include:

lectures or interviews that are available for download but not published at YouTube
lectures or interviews that are available for download and you wish to keep local (e.g. to ensure speed, precision, persistence, or offline viewing)
copyrighted material from iTunes
movies
home video
language courses
all other files that you want to learn but cannot upload to YouTube (or just you do not want to be bothered with the hassle)

Hints: Incremental video

Hints: Incremental video with YouTube

Define "YouTube Extract" template to have it used for extracts from YouTube videos (as long as you use the "YouTube" template). Otherwise, "YouTube" template will be used with the background changed to yellow. The simplest approach is to save your "YouTube" template with the new name ("YouTube Extract") and then redefine it (e.g. by changing the element color). Differentiating videos from their extracts (e.g. by color) is important for your strategies in optimizing the learning process.
Incremental video is less incremental than incremental reading. Incremental video was designed on the assumption that, unlike in incremental reading, you go through the main body of the video linearly. This is because of multiple context problems when splitting videos into smaller portions and watching them in unpredictable order. This way, once you make an extract, the assumption is, you do not come back to the once viewed portion of the video as it is already part of the extract. When you return to process the video, you start from the last position extracted or marked with Start. In other words, as you do not extract without viewing, there should be little needed for sub-extracts

Hints: Incremental video with local files

Define "Video Extract" template to have your template used for extracts. Otherwise, The "Video" template will be used with the background changed to yellow. The simplest approach is to save your "Video" template with the new name ("Video Extract") and then redefine it (e.g. by changing the element color to the desired extract color). Color differentiation is important for optimizing the strategies in processing videos
If you want your extracts to run in an infinite loop, make sure you check Continuous on the component menu in your video extract template
There isn't much performance penalty for processing large video files. You can import whole DVDs too and process them incrementally
Not all file formats are supported, however, most formats can easily be converted to MP4 (e.g. with XMedia Recode).
Note that some MP4 files may not work correctly in SuperMemo. Please test a trial SuperMemo before you make a major commitment

Problems with YouTube-based incremental video

Here are some problems you may encounter when learning with YouTube-based incremental video:

YouTube videos can be removed at any time. They can also have embedding disabled. This is a major problem for the present implementation of YouTube-based incremental video in SuperMemo. There are no API tools that could allow of legal localization of YouTube contents (at the moment of writing). However, the problem mostly affects copyrighted material that shows up on YouTube illegally (e.g. movies, music, etc.). Occasionally, video authors themselves remove content. This should never be a problem though if you use your own videos. Hopefully, there will be more and more persistent content uploaded for educational purposes.
Alt+Left arrow shortcut is used by Internet Explorer to return to the previously visited page. As a result, it also works like an Undo for your Start:Stop setting. To return to the previous page in SuperMemo, use the Back button () on the navigation bar
it may happen that your network is down or the Internet is slow when working with YouTube videos. You may therefore prefer to first experiment with a dedicated incremental video collection to learn how to handle videos and prevent interruptions to learning in cases of network problems. You can always stop working with your incremental video collection, return to your regular learning, and resume video learning when the network returns to normal functionality
if you would like to use SuperMemo as your YouTube jukebox, you should also keep music videos in a single collection (or mark them all with a specific keyword for fast subset learning). You can later hide your jukebox collection in the system tray. Your jukebox will play only the selected fragments of individual videos in order of priority as specified in your sorting criteria. As in incremental reading, using intervals and priorities is a good remedy against getting bored with a given song or video
videos cannot be played in SuperMemo if their embedding is blocked (you will see a message Embedding Disabled By Request). Those videos you can only watch on YouTube. This means that those videos are pretty useless in your incremental video learning; however, you can still use prioritization and scheduling tools to choose which videos should be played in which sequence and on which day. Note also that many of those videos are uploaded multiple times by various users, and you only need to search for an alternative that can be embedded
for technical reasons, some videos (roughly one in a thousand) cannot be watched incrementally (they will ignore extracts and always play in their entirety); some can show large rounding errors for extract precision (around 10% of videos can only be watched in large increments). This problem seems to have largely been resolved after 2010.
some videos may display loading errors on first entry. Right-click over the video and choose Refresh to reload (or use Alt+Left arrow followed by Alt+Right arrow to revisit the element). If the error reoccurs, use the link in references on the right to visit the original page at YouTube. This problem seems to have largely been resolved with the newer versions of YouTube API
there are bugs in YouTube API, flash player, and Internet Explorer that can interfere with your work. Your script can use version=2 or version=3 of the player. Both have ups and downs. Version 2 tends to hang on very short video pieces that play over and over again (due to a memory leak in Flash software). Version 3 may hang for 10-20 seconds when loading videos preceded with advertising (without actually showing the advertising).

Your own incremental video script

Videos are handled with the help of YouTube Player API in HTML components using a small JavaScript program. When you first run SuperMemo, it writes this script into a file stored in the [BIN] subfolder of the folder in which you installed SuperMemo. The name of the file is yt.htm. If you know JavaScript, you can substitute your own incremental video script in that file (e.g. to change the layout, size of buttons, or even add new functions for processing videos). Here are the only components of the script that you need to preserve:

the order of INPUT fields (these are used by SuperMemo to collect extract boundaries)
SELECT and OPTION fields for generating extracts
INPUT field with YouTube video ID substituted by SuperMemo (id="videoid")
INPUT field that determines the position of video start and end (id="startvideoat" and id="stopvideoat")(must be "0:00:00")

The most important change you may wish to introduce is to decide between version=2 and version=3 of the player. Both have their bugs and disadvantages. For more see Player Version Issues. The version of the player you choose will be picked by default in all your YouTube elements.

If you would like to share your own script with others, you can upload it to SuperMemoPedia.

Incremental video

Contents