InqScribe News

Guest Blog: How documentary filmmakers transcribe and subtitle with InqScribe

InqScribe has many different kinds of users. But the majority of our users can be categorized into three groups: university researchers, professional transcriptionists, and documentary filmmakers.

To highlight how our folks are using the tool, we would like to do an occasional feature where we invite our users to be guest bloggers.

Here’s our first guest blog from a pair of award-winning documentary filmmakers. We were particularly interested in how they might be using InqScribe as part of their translation and transcription workflow, so we asked them to elaborate:

—

DigAllMedia

By Dianne Griffin and Erica Jordan

Digall Media, a 501(c)(3) organization founded by Erica Jordan and Dianne Griffin is currently in production on their one-hour documentary Painted Nails – a Vietnamese immigrant story of exotic nail art, pampered clients, and the serious health risks that lurk beneath the brightly painted surface.

We’re excited about using Inquirium’s product InqScribe to transcribe dialogue and create English subtitles for our Vietnamese and Spanish speaking characters. We’re still fine-tuning the workflow of importing subtitles, generated with InqScribe, into Final Cut Pro. It took some time to figure out the importing and exporting specs, but it was worth trouble-shooting. We can now transcribe dialogue in InqScribe with timecode, export it as an XML file using a custom FCP XML template. When the XML file is imported into Final Cut, the subtitles (as text elements) magically appear on a new timeline. InqScribe’s support page offers to look at your files to help trouble-shoot subtitling issues. InqScribe works great with a foot pedal, saving valuable post-production time.

Filmmaking is hard enough; it’s great to find a product such as InqScribe to make it easier.

Please contact us at info@inquirum.net if you’re interested in highlighting your work.

New InqScribe video tutorials

InqScribe is so easy to use, who needs tutorials, right? Well, a little background can’t hurt, so we created four new introductory video tutorials that provide (1) a basic overview of InqScribe, (2) an introduction to shortcuts and snippets, (3) tips on using time codes, and (4) an introduction to subtitling. They’re all a part of our revamped home page.

Tip: Inserting the current time of day or Transcribing while videotaping

You can insert the current time of day into your transcript. This is useful for instance, if you want to take notes while you are recording during a meeting or video shoot. Later when you import the media, you can sync the start time to your video.

For example, let’s say you’re shooting an interview, and it begins at 1:00pm. During the interview, you can take notes in InqScribe, noting when a particularly interesting conversation happens by inserting the current time. For instance, if the interviewee says something interesting 12 minutes and 3 seconds into the interview, you can insert a time stamp next to your note about that with one keystroke, e.g.:

“[01:12:03.00] T didn’t know it at the time.”

Alternatively, instead of using time of day, you can also use a stopwatch synced to the start of the video recording.

How do you do this? Just set the Media Source to an Offline Media type and select “Use time of day” or “Use stopwatch timer”.

Here are detailed instructions:

1. Create a new transcript “FIle->New Document…”
2. Click on the “Select Media Source…” button
3. In the “Source Type” popup menu, select “Offline Media”
4. Under “Time Code:” select the “Use time of day” radio button.

(Or you can select “Use stopwatch timer” and select a start time and end time.)

Then just type away as you normally would, using Command-; or (Ctrl-; in Windows) to insert the current time. Instead of using the media time, InqScribe will now insert the current time of day, or the stopwatch time.

InqScribe and stenographers

We were recently contacted by a stenographer who wanted to use InqScribe with a stenotype machine for video and audio transcription. In theory this should be easy, right? A stenotype is just another input device, like a foot pedal. But in practice, there are a few challenges. Here’s our current understanding of how InqScribe and stenotype machines can play well together. If you’re a stenographer and have thoughts on this, please let us know.

Can I type from my stenotype machine into InqScribe?

Yes. In most cases, you’re going to be using CAT software that converts stenotype chording into individual key events. (Basically, the CAT software pretends its a keyboard sending keystrokes to the computer.) This approach works well with a lot of applications, including Word, TextPad, and others. InqScribe should have no problem receiving text input in this way. (A point of comparison: we know that InqScribe works well with AutoHotKey and other utilities that send artificial key events.)

Can I trigger shortcuts and snippets in InqScribe via my stenotype machine?

Currently, we doubt this will work. That’s because the stenotype machine is not normally listed as an available input device. The way that InqScribe tracks shortcut triggers is by checking all known input devices (keyboards, foot pedals, gamepads, etc.) for known shortcut combinations.

With a stenotype machine that is interfaced via CAT, the machine doesn’t show up as an available device. The key events just magically appear. And since key events by themselves don’t serve as triggers for InqScribe, we recommend instead a workaround like using a foot pedal with the stenotype machine, where the foot pedal provides the triggers.

We’re looking into ways that we might support snippet insertion (or shortcut triggering) via normally text entry. There are a couple of ways to do this. MS Word has one model for its auto-complete tool, where common abbreviations and misspellings are corrected or expanded after you type them. Another approach relies on an explicit triggering keystroke which follows the snippet abbreviation, so you’d type something like “tc`” and the ` character is the trigger to look up the “tc” snippet and insert it into the transcript. Since our goal is to let you transcribe more efficiently, we want to play around with these approaches to see what feels most comfortable. But if you’re got an opinion, let us know.

Power User Tip: Snippet Autocomplete

If you have text that you are repeatedly typing in your transcript (e.g. analysis codes, or long words), you can create snippets to insert them quickly.  Here are some tips:

• You don’t have to assign a trigger/keyboard shortcut to snippet
• You can activate the trigger window with <cmd-1> (Mac) or <ctrl-1> (Windows)
• When the trigger window is activated, you can select snippets by typing the first few letters, or using the arrow keys to select the snippet
• Hit <return> or <enter> to insert the snippet

Perian gives InqScribe for Mac access to Flash & more

Perian 1.2 is out, and highly recommended for users of InqScribe on MacOS.  A free, open-source QuickTime component, Perian gives QuickTime– and thus InqScribe– access to many popular media formats, including Flash video (.flv), DivX, 3ivx, and many others.  Alas, there is no comparable tool for Windows.

Workaround for Final Cut Pro 7 Importing Bug

Update: this bug is fixed in the 2.1 beta. Get the beta here.

If you’re using Final Cut Pro 7 and importing InqScribe-generated FCP XML files, you’re probably pulling your hair out. As of FCP7, any imported subtitles beyond the two minute mark show up with a duration of only one frame.

The problem is twofold. First, InqScribe is setting incorrect values for the in and out points for each generated subtitle. In prior versions of FCP, these values were essentially ignored on import, because in and out points for a static subtitle don’t really mean that much. (The start and end points for the subtitle, which determine where the subtitle goes in the sequence, were and are correct.)

Unfortunately, FCP7 is interpreting those values differently, and any subtitle with an in point greater than the subtitle’s stated duration ends up with a frame length of 1. Since InqScribe was setting every subtitle’s duration to 3600 frames (because this value shouldn’t really matter: effective duration of the subtitle is based on the start and end values), most users will find that subtitles that start at the two minute mark or later are affected.

We’re working on a fix for the next beta release. In the meantime, there is a workaround.

Here’s an excerpt from an InqScribe-generated FCP XML file:

<generatoritem id="Text">
    <name>Text</name>
    <duration>3600</duration>
    <rate>
        <ntsc>TRUE</ntsc>
        <timebase>30</timebase>
    </rate>
    <in>7540</in>
    <out>7610</out>
    <start>7540</start>
    <end>7610</end>

Note that InqScribe sets the duration to 3600 (regardless of the actual duration, which is based on the start and end values). InqScribe also sets the in and out points to the start and end values. The problem is that in and out values should technically never be greater than the duration.

So the fix is to change every instance of in and out to this:

<in>100</in>
<out>3600</out>

With this change, in and out stay within duration’s range, and FCP7 won’t clip the resulting subtitle.

To make this change easily, use a tool that supports regular expressions to find all instances of the in and out tags. Here’s a solution that uses sed, which comes installed on OS X.

1. Export the FCP XML file from InqScribe as usual (let’s say it’s called export.xml).
2. In the Terminal, navigate to the directory containing export.xml, and issue this command (which is one long line, make sure to copy the whole thing):

sed -e 's_\(<in>\)[0-9]*\(</in>\)_\1100\2_g' -e 's_\(<out>\)[0-9]*\(</out>\)_\13600\2_g' < export.xml > export_fixed.xml

3. Import the resulting export_fixed.xml into FCP7.

If you want to dig into sed so you understand what that command is doing, here’s a solid sed tutorial. It’s a very powerful tool.

We’re Tweeting

InqScribe is now on Twitter. We think this will be a great way to share quick tips and information about using InqScribe and related media tools. You can track our tweets on the sidebar of this blog, or follow InqScribe on Twitter.

Dive into Video

If you’re at all interested in where video and audio for the web is going, this chapter by Mark Pilgrim is required reading. Mark does a great job providing background on the core cross-platform codecs that are in use today (with the cross-platform bit ruling out WMV) and talks about how HTML5 will offer native support for video playback.

His focus is primarily on h.264 and Ogg video, since offering video in those two formats will cover all modern browsers. He also describes methods and tools you can use to encode your videos in those formats (including Firefogg, which was new to me, and FFMPEG2Theora for Ogg, and Handbrake for h.264).

InqScribe 2.0.5 is compatible with Snow Leopard

While we do have a new version of InqScribe expected soon, the current version, 2.0.5, works fine with the new Snow Leopard Mac OS (10.6).

In general, if you do have problems with InqScribe, please report it to us using the InqScribe Help menu’s “Report a Bug…” feature.  This provides us with a little more information to help us more quickly troubleshoot the problem.