Flow Transcriber: a transcript for your media


Home » Help Center » Flow » Flow Transcriber: a transcript for your media



With Flow Transcriber, you create transcripts for your audio and video. The transcripts are synced word-for-word with your media, so a text selection in your transcript will correspond to setting an in- and outpoint in your media, making it extremely easy to e.g. pull quotes out of an interview!

How to open the Transcript application

Find the Go to transcript action on a clip thumbnail and click it.

thumb-action-transcript

You can also select multiple clips and click the Open transcript icon in the toolbar:

tooltip-transcript-app

If you open the Transcript app for multiple clips, you’ll have a navigation element at the top of the page which you can use to go to the next or previous clip:

clip-details-mo-nav

The tab buttons in the top right corner of the screen can be used to toggle the visibility of the player, the transcript, the subclip details and the clip details:

Interface elements

The left side of the screen is taken up by the player. Read more about the Flow player here.

The Transcript application has two states: View and Edit. The View state can be used to mark quotes, while the edit state is used to… well, edit the transcript. The player controls at the bottom of the page change when toggling between view and edit mode.

Editing the transcript in Edit mode

If you want to edit the transcript, click the EDIT button.

The bottom part of the screen shows the controls which control the player and can launch tools for editing the transcript.

controls
Each control is linked to a shortcut.

F1 Play / Pause the player.
F2 Play / Pause the player. When the text cursor is inside the transcript, playout will start from that position. Pausing the player using F2 will move the text cursor to the word corresponding with the player position.

When a text selection is active in the transcript, only that selection will be played out.

F3 Mark a time range for text input, as explained here.
F4 Rename the speaker of the current paragraph
F5 Selecting a range of text in the transcript and then pressing F5 allows you to change the in and out points of that selected text.
F7 / F8 Rewind / Forward the player one second. The combination with the shift key will move only a single frame.
F9 / F10 Change playback rate

Writing a transcript without using speech recognition

Note: we use F1 -> F11 shortcuts. On a Mac, you might have to hold down the fn-key as well. So F3 becomes fn+F3 etc. If you find this annoying, you might want to configure this as shown here: https://support.apple.com/nl-be/HT204436.

Play your video with the F1 control. When you hear something you want to transcribe, press and hold down F3. The following dialog should pop up:

F3-down

Hold down the F3 key and try to remember what is spoken out loud. Release the F3 key after a few spoken words. Now the dialog changes to

F3-released

You can start typing the text you heard and press the return key (or press Confirm) when you have finished. The text you entered now appears in the transcript:

transcript-part

What happened behind the scenes is we took the in-point (where you start holding down F3) and the out point (where you released F3) and assigned that to the start of the first word and the end of the last word you typed. Then we interpolated the timing so all words get linked to a position.

Try it out! If you click a word while holding down Ctrl (Cmd if you’re using Mac), the player will seek to the position where that word is spoken out loud!

Now continue using the F3 tool to transcribe more text. When you confirmed a piece of text, you can press F3 again immediately. This will start recording again from the last out-point.

In the F3 dialog, you can adjust the in- and outpoint if necessary. You can also set the in or outpoint to the current playhead position by pressing the green dot next to the timestamp. Press the large Preview button to preview the selected range.

Use automated speech recognition

Find the Start Speech Recognition action on a clip thumbnail and click it. Note that it might be collapsed under the three dots  button.

speech-start-action

You can also select multiple clips and click the Start Speech Recogntion action in the Other Actions menu of the toolbar:

selection-start-speech-action

The speech recognition will now start for this item / these items. The workflow takes some time to run (it runs about real time, so 1h of material will take 1h to process). You can monitor the progress in Workflows section of the Clip Details Screen.

Wait for the speech recognition to complete. Then open the Transcript app like explained here. You will now have a starting point for your transcript, based on speech recognition. Go into edit mode by clicking the EDIT button above the transcript text. Each word is synced to a position in the video. Clicking a word while holding down the Ctrl key (Cmd key on Mac) will seek the player to that position.

transcript-after-speech-recognition

You can simply edit the transcript like you would do in a text editor. The software will keep the link with the timecode positions in you media intact.

Mark quotes in the View mode

If you want to use the transcript to mark interesting parts and create quotes, click the VIEW button. Also make sure that the subclips sidebar is visible by clicking the subclips button in the top right corner:

subclip-tab-button

The interface now contains a player at the left, the subclips sidebar at the right and the transcript in the middle.

Note: The player controls and the subclips sidebar are explained in detail here. This article is about using the transcript text to create subclips.

Creating a new subclip using the transcript is as easy as marking the text in the transcript, and then clicking the ‘new subclip’ text input, as shown in the animation below. The text selection is instantly converted into a timeline selection. Naturally, you can also add a description, tags and mention team members.

create-quote-anim

 

Edit speakers

If you use automated speech recognition, the speakers will have names like “Speaker 1”, “Speaker2”, … To change these to meaningful names, you use the ‘Edit Speakers’ tool. The tool is launched using the button at the top of the transcript text

edit-speakers-button

The following dialog appears

With the tool, you can navigate through all speakers using the big gray buttons on the left and right side of the dialog, and give them a name. After giving a speaker a name, don’t forget to press the ‘Save Speaker Name’ button.

To help you identify the speaker, a paragraph of text spoken by this speaker is highlighted in the transcript. You can press the ‘Listen’ button to start playing the paragraph in the video player. Press the ‘Next paragraph’ button to listen to another paragraph spoken by the speaker.

speaker-tool-animation