Document Formats
The transcript documents can be either .txt, .rtf, .rtfd, .doc, .docx, or a mixture of these. There are many providers of transcripts, but many do not provide sufficient time stamps for Lumberjacks' needs.
Time Stamps
Most “industry standard” timecode-stamped transcriptions can be used, including files with timecode or time stamps at the start of every text range, or where the time stamps are mid-text range. The more time stamps in the transcript, the more accurate the lock between media and transcript will be.
The media can have either SMPTE timecode or count up from zero. Builder can recognize timecode or time stamps in various formats separated by “:” or “;” or “.” and optionally surrounded by square brackets []. For example: 01:23:45 or 01:23:45:21 or 1:23:45 or 1.23.45 or 1:23:45.420
Speaker
Builder also looks for named speakers in uppercase text in the transcript. These can be separated on their own line or at the head of the transcript as either SPEAKER NAME:transcript text... or SPEAKER NAME:transcript text... or SPEAKER NAME:transcript text... The colon after speaker name is important,
Download this example as a Microsoft Word document.
0:00:00
Speaker 1: Transcripts start with a time stamp representing the start of the clip. Then there are typically time stamps at the start of every paragraph. Anything between the time stamps becomes a Keyword Range in Final Cut Pro X named after the speaker, and with the transcript in the keyword's Notes field.
00:00:17
SPEAKER 2:
When a new person starts speaking they are often identified by all caps on their own line before the text of their transcript .
00:00:35
Speaker 1: But interview subjects can also be in-line followed by a colon and then a tab or space before their transcript. Either way is fine.
[00:00:44]
SPEAKER 2: There's also a lot of flexibility in how the time stamps are formatted. They can have square brackets around them, or not. The numbers can be separated by : or ; or . They can have one or two digit hours. They can be hours-minutes-seconds or hours-minutes-seconds-frames or hour-minutes-seconds-milliseconds. They can represent your clip's timecode or count up from zero.
0.01.29 The time stamps can also be in the same line as the transcript, or even in the middle of a transcript. The more time stamps there are in the transcript the more accurate the keyword ranges will be. At least one every minute.
[00:01:36;08]
SPEAKER 1:
Make sure that time stamps in the document progress. Lumberyard does a “dry-run” to check for out-of-order time stamps and will warn you if it finds problems.
00.01.52 The transcript document should have the same name as the clip it was generated from, so you'll have one transcript document for each clip in your Final Cut Pro X event. The document can be saved as .txt, .rtf, .rtfd, .doc, or .docx. Organize the transcripts for the event into a folder that you select in Lumberyard.
00:02:22.345
Speaker 2: Lumberyard looks inside your folder and matches the transcript name to the clip name in the event. That way the processing of transcripts can be done as a batch instead of one at a time.
[0:02:40]
It's not necessary to have a time stamp at the end of the transcript. If there's no final time stamp, Lumberyard will use the end of the media file as the last time stamp.
Take1
Take1 Transcription have the option for "Lumberjack Ready" transcriptions. Contact one of Take 1's production team to walk you through the process:
Sarah Ranfone
Tully Satre
Rebecca Simpson: [email protected]
or call on 855.958.2531 (US) or 0800 085 4418 (UK)
SimonSays.ai
SimonSays.ai have a dedicated format for Builder. In Export Options, select Lumberjack Ready.
SpeedScriber
From SpeedScriber export "Plain Text with Time stamps > Start of Sentence." SpeedScriber does not add the ":" at the end of the speaker name, so you will need to open the exported transcript into a text editor, and replace each "Name" with "Name:" Repeat for each name.
Rev.com
From Rev export Captions > SRT.
Speechmatics
Speechmatics is supported from within Builder. Select Add to Batch.
Trint
From Trint, export "Entire Transcript > Word Document." You will need to remove the title at the top of the transcript before importing to Builder. Trint only provides one time stamp per paragraph in this export, so we recommend creating shorter paragraphs in Trint before exporting.
Verb8tm
The Word output format would work for Builder. You would need to manually remove the heading/title information from the top of the document before importing it into Builder.
Generic
In the absence of a specific recommendation SRT (Captions) should work for Builder.