Annotate and Analyze Online Video with OTTO

This post will briefly trace the evolution and application of a powerful new video annotation platform that enables collaborative viewing and encourages cooperative, crowd-sourced criticism. This new tool, called OTTO (Open Text Tool for Online video), was first used in conjunction with a video annotation assignment in a Film Noir MOOC that generated close readings of key moments in film noir, but its potential uses extend much further: the tool could change how films and videos are taught with its capacity to simply and powerfully join critical analyses to film moments, and relate discreet analyses by source and by runtime across films that share stylistic or generic traits.  While other online video annotation tools can support participatory and collaborative viewing and reading exercises in the classroom, there are particular functions unique to, or uniquely powerful in the context of, OTTO. To present a fuller picture of this tool and its potential, I would like to step back and trace OTTO's evolution from an idea to its current form and function.

Winter 2011: Shannon Clute and I co-authored The Maltese Touch of Evil: Film Noir and Potential Criticism (Dartmouth College Press). The book explored exemplary moments in film noir (which we dubbed 'noiremes') through the lens of Oulipian constraint and recombinatorics. The final section of The Maltese Touch of Evil (MTOE) suggested the idea of developing a "participatory, procedural (and ultimately, perhaps, encyclopedic) database dedicated to film noir" (262). The MTOE project "would allow readers to add their own investigative notes, to tag or annotate the filmic material with further information and metadata, to author new noiremes, and to suggest or create constraints for resequencing existing noiremes (percent of a frame in shadow, number of times 'Baby' is uttered in one scene of a screenplay, etc.)…" (263). At the time the book was published, we weren't sure how or evenif the MTOE project would develop, but we were hopeful that such a project was possible. In terms of its potential uses, the MTOE project could help students pursue close readings, share viewing notes online, and curate annotated entries in an ever-growing digital archive.  Moreover, in keeping with the Oulipian roots of the project, such a tool would be able to generate surprising insights leading to new critical observations or new avenues of investigation through mathematical constraints and algorithmic processes.

Summer 2012: OTTO moves from an idea into an active development phase through the confluence of two events. First, I am appointed as the Executive Director of iLearn Research at Ball State. Second, I am asked to teach a MOOC on the topic of film noir. As a researcher and an instructor, I begin to think about the learning outcomes and design of the MOOC, and return to the MTOE Project idea. Over the next few months, Shannon and I have in-depth discussions about designing and developing a tool in the spirit of the MTOE Project. Such a tool will need to enable student-generated annotations on specific shots in film noir, support detailed student commentary and analysis, and exist as part of a database structure with algorithmic capabilities. Based on these initial design considerations, an iLearn Research team, led by developer Chris Turvey, starts prototyping OTTO in late 2012. As the development process begins, Chris proposes that OTTO supports the annotation of YouTube videos, opening up many new possibilities and uses.

Spring to Summer 2013: OTTO is available in an early beta version and was deployed in a student assignment in Ball State University's Investigating Film Noir MOOC. As we developed OTTO, my team explored other video annotation tools including Indiana University's Film Annotators' Workbench, Vertov (an annotation plug-in for Zotero), and But these tools didn’t provide all of the functions we wanted in OTTO. For example, OTTO contains two timelines for annotation purposes. One timeline tracks the remaining runtime of the film. But a second timeline correlates any filmic moment to the film's running time percentage. An algorithm in OTTO automatically and mathematically generates this second "segment timeline" by dividing any YouTube video into 100 evenly spaced time increments. As part of the recombinatory and constrained spirit behind OTTO, the segment timeline facilitates comparative analyses of annotated entries across multiple films in a collection through mathematical constraint. 

In the Classroom: My MOOC students used OTTO to annotate key film moments from the weekly film noir screenings. To avoid copyright issues, OTTO's film noir collection contains only public domain films such as Detour, The Strange Love of Martha Ivers, and D.O.A. All these public domain films are available on YouTube.

Each annotated entry in OTTO is connected to a specific time stamp on the streaming video. Therefore as students annotate a film, they automatically share their clips and their analyses with other students. You can click on any annotation in OTTO to jump directly to that moment in the video. Moreover, OTTO annotations function like mini-essays: they reveal how students are making sense of these films and how these moments can contribute to an overall understanding of film noir. Click here to visit OTTO (or go to the following URL: and explore how students annotated a film like Edgar Ulmer's 1945 noir masterpiece Detour. FYI, you need a Google account (such as a Gmail address) to log into OTTO.

As a teaching aid, OTTO provides a template for students on how to write concise and close analyses of film moments. First, students have to select an annotation category. There are ten categories in OTTO such as dialogue and screenplay, performance/acting, photographing, staging, etc., to help students limit their focus to a particular aspect of a shot or scene. Second, there is a description text box for students to provide a context or a simple overview of a particular filmic moment. Third, there is an analysis text box that enables students to reflect on how the moment they are tagging and annotating contributes to their understanding of film noir and connects to larger theoretical, cultural, or thematic issues.

OTTO becomes a living archive that exists both inside and outside of the classroom. It is an act of open scholarship and knowledge sharing. It can begin in a classroom setting but easily opens itself up to other scholars and fans via the Internet. In fact, OTTO is an open text tool, as its name indicates, so anyone can leave an annotation in the archive. We are currently in the process of launching our second annotation project, "The Dude Meets OTTO: Annotating The Big Lebowski Mashups."

Students in my MOOC suggested other uses for OTTO including annotating video lectures. It could be a very useful tool to assess student understanding of a lecture, or to leave questions and comments for instructors based on specific moments. There are likely many other uses for OTTO since it can annotate any YouTube video.

OTTO is in active development, so we welcome any feedback or suggestions. Also, what are your thoughts about video annotation projects and the types of assignments that can be built using these kinds of tools?


I really appreciate seeing how the process of this project developed over two years. I see that the MOOC was the first chance you had to test OTTO with an audience. I have never developed a piece of software for a class. What was your process for developing what OTTO needed to do as far as specificity of annotation, easy of use, and discussion structure? 

Have you thought of sharing the code?

I can also see this being advantageous to course lectures. I am currently taking an asynchronous class in statistics and the ability to annotate the lecture when I had questions would be useful and a way to engage recorded lectures and make the relatively passive experience a bit more interactive. 

Good questions, Jamie.

In developing OTTO for classroom use, I thought through how to create a crowd-sourced, scalable, and online assignment. Ease of use was essential due to scale: it would have to be a learning tool that was easy to use, otherwise it would create lots of problems with the large number of students enrolled in the MOOC. Basically, we wanted a program that was fairly intuitive to use and that leveraged existing ideas around how to comment on a video. When we rolled out OTTO in the MOOC, we had very few issues with usability. I think most students had no problem using OTTO. I got many more questions about how to write annotations about film, which leads to your next question.

Annotation projects interest me because there really aren't a lot of standards out there right now for how to annotate video. So for OTTO, the bigger issue was explaining to the student why they would want to annotate a video in the first place. I focused on the idea of identifying key moments in film noir that help advance understanding of this particular film style. In other words, what moments really stand out as evidence of how the noir style creates meaning? To answer this question, a group of students takes on the role of active research and demonstrates how well they understand what the noir style is by identifying it themselves by tagging critical film moments in the movies. I saw this use of OTTO as a hands-on and participatory demonstration of their ability to read and discuss films noir.

For the specificity of the annotation, I created 10 pre-defined film analysis categories. Mainly, I wanted students to write short but detailed entries about only one formal aspect of any filmic moment to keep these annotations focused. Therefore, the drop down category menu was intended to limit the scope of their writing. The choice of category should get the students to ask key initial questions: about the moment I am about to annotate, am I writing about the dialogue? Or the editing? Or the sound design? Or the acting? Or the staging? etc.

I also wanted to help students recognize the difference between plot summary and critical analysis when they are annotating a film. One of the more common writing problems encountered by film instructors in introductory courses is that students just want to write about the film's plot and recap key moments without pushing their analyses in any new direction. Therefore, I created OTTO with two text boxes for each entry. In the first text box, the student is just asked to describe the moment they select and in a second text box the student is asked to analyze that same moment. By clearly separating these two types of film writing, I sought to push students in the annotation template to recognize the difference between describing or summarizing a shot or scene and actually analyzing it.

Regarding sharing the OTTO code: my team and I are actively looking for the best way to do that. I see this as something I would like to connect to other on-going annotation projects. I am hoping I can reach out to others working in the field of video annotations to see how we can go about getting this code out in the development community.

I suppose this is more of a pedagogy question. In the classroom when we lead discussion, we have the ability to move conversations that are going the wrong way and kind of put them back on track. I am assuming you have the power to delete comments and or reply to them afterwards. Were any of your students a little apprehensive about annotating in a relatively public space. Was there a concern over the consequences of sharing evolving ideas in a static space?

There is the ability to flag annotations in OTTO for review by the site's editors, so that is one way to keep erroneous annotations out of the database. You can also reply to an annotation, so that allows for any moment of annotation to become a threaded discussion in its own right.

There is always apprehension about public scholarship, and I like to use "opt-in" strategies in this regard. In the MOOC, we allowed students the ability to keep their username private, if they wanted.

I would think, in future versions of this project, I will explore more deeply how sustainable peer review and the history of peer review of these annotations would work. I think these become a different kind of knowledge object than the comment fields in YouTube. I see annotations, and the growth of annotation standards, as a way to curate and analyze video content on the Web. I would love to build more crowd-sourced tools into each annotation to gain confidence that as users read and react to the annotations there are additional validation methods that will improve OTTO as a knowledge resource. These are definitely areas I want to pursue as we continue to develop OTTO. 

Add new comment

Log in or register to add a comment.