All About Tags!
Introduction
First, let's define tags. There are two types of tags: Speech Tags which replace words that you cannot make out, are unsure of or taking a guess, or are in a foreign language, and Nonspeech Tags which notate a specific sound or silence.
Tip: To help avoid typos, especially with tags, consider creating snippets or using a text expander program. For example:
ccc = [crosstalk]
fff = [foreign]
mmm = [music]
appl = [applause]
sss = [silence]
lll = [laughter]
Tag Placement and Punctuation
Speech Tags
[inaudible], Guess, [crosstalk], and [foreign] all take the place of speech that cannot be transcribed. As such, they should all be placed inside of punctuation, i.e., punctuated as you would if it were speech.
Do you know when [inaudible] should arrive?
I'm not sure why he [crosstalk].
She wanted to go to the [bank?], but I thought it was too late in the day.
My Egyptian mother-in-law always says to me, "[foreign]." I just smile and nod politely because I have no idea what that means.
Since the tags are taking place of speech that you cannot transcribe, you will have to use the surrounding context or the inflection in the voice to determine the appropriate punctuation.
[foreign]?
I'm not sure I understand [crosstalk]--
[inaudible].
Nonspeech Tags
[music], [applause], [laughter], and [silence] do not replace speech; therefore, they do not need to be punctuated as speech. The [laughter] tag should never be placed on its own line, while the [silence] tag always rests on a line of its own with no punctuation.
[music] Welcome, ladies and gentlemen! Thank you for coming tonight. [applause] I'd like to introduce tonight's guest of honor, Mr. Jean Deaux. [laughter]
Please turn over your exams and begin.
[silence]
Mrs. Smith, I have a question.
Speech Tags Up Close
[inaudible]
The [inaudible] tag is fairly straightforward. It simply replaces any word(s) that you are unable to make out. That said, only use [inaudible] when you cannot make a guess and it is not inaudible due to crosstalk. Please use the most appropriate tag to convey the situation.
She went to the [beach?] but forgot her [inaudible]. I don't think [crosstalk]--
She did what?
Well, she never sat down. She walked the entire time. She was [inaudible].
Although the [inaudible] tag can stand in for multiple words, please only use the tag for words that you cannot make out. If you are able to make out words between, please transcribe them.
She went to the [inaudible] but forgot her [inaudible].
Sarah did not [inaudible] for her final exam, but [inaudible] anyway.
For some great tips and tricks to help you squeeze every last word out of your file, take a moment to read Dealing with Poor Audio.
Guess Tag
The Guess tag demonstrates that you are taking an educated guess at the word based on what it sounds like and the surrounding context, but you are not certain. It can also mean that you are not certain of the spelling because your research failed to provide a definitive result (such as a name). This is the only tag that is modified, meaning you replace the tag with your actual guess.
- What are the disadvantages of [lithium ion?] batteries compared with other rechargeable batteries?
- Some are advocating that the Florida [scrub?] jay be the state bird as it is more prevalent than the Northern Mockingbird.
- Today's guest speaker will be Paul [Vasquez?].
When it comes to names, if you are not able to verify the spelling due to lack of context, etc., please use the most common spelling you can find. For some great tips and tricks to help your research, check out the article How to Research Like a Sleuth.
[crosstalk]
The [crosstalk] tag is used when you are unable to make out what someone is saying because another main speaker is talking over them or at the same time. This tag is NOT used to mark where the crosstalk occurs, but rather to replace words that you cannot make out. If you can make out what is being said, transcribe it. Let's look at some common crosstalk scenarios.
Please note: All speaker designations and use of curly brackets are for clarification purposes only. DO NOT label speakers (the QA will take care of that), and DO NOT use curly brackets in a transcription job. The use of curly brackets indicates where the crosstalk begins and ends within the speaker's line.
Scenario 1
Your main speaker (S1) is talking, and a second speaker involved in the recording blurts out statements in response.
S2: {What study are you referring to?}
S1: If we really want to understand the flow of ideas, we will need to limit the number of people who can talk at any one time.
Incorrect:
The study showed that the more people you have involved in the discussion the harder it becomes to keep track [crosstalk] to which person. If we really want to understand the flow of ideas, we will need to limit the number of people who can talk at any one time.
Correct:
The study showed that the more people you have involved in the discussion the harder it becomes to keep track of [crosstalk]--
What study [crosstalk] to?
--belongs to which person. If we really want to understand the flow of ideas, we will need to limit the number of people who can talk at any one time.
Explanation:
In scenario 1, the incorrect version fails to transcribe the second speaker who is talking over the main speaker. Remember, everyone involved in the recording has to be transcribed, and each new speaker must be placed on a new line. Placing just the [crosstalk] tag to note missing words is not a complete transcription if you can hear even part of what the second speaker is saying, and in any case, the second speaker needs to be placed on their own line, not in the middle of another speaker.
In the correct version above, we assume that you cannot make out everything said and so used the [crosstalk] tag to stand in for that speech. However, if you CAN understand everything said by both speakers, then there is no need for the [crosstalk] tag:
The study showed that the more people you have involved in the discussion the harder it becomes to keep track of which idea--
What study are you referring to?
--belongs to which person. If we really want to understand the flow of ideas, we will need to limit the number of people who can talk at any one time.
Scenario 2
Your main speaker is talking, and two speakers involved in the recording then interject statements on top of each other in response. Words of the main speaker, plus sets of words for each of the speakers that interject, are muffled by all three voices.
S1: The study showed that the more people you have involved in the discussion the harder it becomes to keep track {of which idea belongs to which person.}
S2: {What study are you referring to?}
S3: {The one we read about yesterday.}
S1: {If we really} want to understand the flow of ideas, we will need to limit the number of people who can talk at any one time.
Correct:
The study showed that the more people you have involved in the discussion the harder it becomes to keep track of [crosstalk]--
What study [crosstalk] to?
[crosstalk] we read [crosstalk].
--belongs to [crosstalk]. If [crosstalk] want to understand the flow of ideas, we will need to limit the number of people who can talk at any one time.
Explanation:
In scenario 2, as all the speakers are there for a meeting and are therefore engaged in the recording or part of the main purpose for the recording. As such, they need to be transcribed and acknowledged if you can hear even one word they say. Failure to transcribe the secondary speakers could result in a rejection for missing words, so it is best to include them rather than exclude them.
Remember, the [crosstalk] tag IS NOT used to indicate that there are multiple people speaking. It takes the place of words that can not be heard or transcribed clearly. Again, if you can hear everything said, there is no need for the [crosstalk] tag:
The study showed that the more people you have involved in the discussion the harder it becomes to keep track of which idea belongs to which person.
What study are you referring to?
The one we read about yesterday.
If we really want to understand the flow of ideas, we will need to limit the number of people who can talk at any one time.
Scenario 3
Your main speaker (S1) is talking, and an undetermined number of speakers (N/A) all start talking at once. Out of that large number, a few voices are identifiable and a few words can be made out here and there.
S1: The study showed that the more people you have involved in the discussion the harder it becomes to keep track of {which idea belongs to which person.}
S2: {What study are you referring to?}
S3: {The one we read about yesterday.}
S4: {That is a bogus study!}
N/A: {mumble, mumble, first, mumble, time.}
S1: {If we really want to understand} the flow of ideas, we will need to limit the number of people who can talk at any one time.
Correct:
The study showed that the more people you have involved in the discussion the harder it becomes to keep track of [crosstalk]--
What study [crosstalk] referring to?
The one [crosstalk] read [crosstalk].
--belongs to [crosstalk].
That is [crosstalk] bogus [crosstalk].
[crosstalk].
If [crosstalk] want to [crosstalk] the flow of ideas, we will need to limit the number of people who can talk at any one time.
Explanation:
Like in scenario 2, all the speakers are engaged in the recording or part of the main purpose for the recording and need to be transcribed. You do not have to acknowledge all the other voices for which you cannot hear individual words. Again, remember, the [crosstalk] tag is not a "there are several speakers here" interpretation. It is a "I cannot hear this word because another main speaker is talking" tag. If you can hear a second voice and they do not prevent you from hearing all the words the main speaker is saying, you do not need the [crosstalk] tag on the main speaker's line, but you do need to transcribe the second voice on their own line. With this many people talking, the chances you'll be able to pick out every word are diminished, but if you can pick it all apart, then the only [crosstalk] tag in this scenario is for the multiple speakers mumbling incoherently:
The study showed that the more people you have involved in the discussion the harder it becomes to keep track of which idea belongs to which person.
What study are you referring to?
The one we read about yesterday.
That is a bogus study!
[crosstalk].
If we really want to understand the flow of ideas, we will need to limit the number of people who can talk at any one time.
A Word About Crosstalk
Remember that we don't expect perfection or miracles. In bad audio with a cacophony of voices, it's not going to look pretty, and you're not going to be able to pick apart everything said. We just ask that you do your best to capture the main speakers to the best of your ability. For what you can't accurately capture, use the [crosstalk] tag. It'll become easier with time and practice, we promise!
[foreign]
Fully Foreign Files: Transcribers, if a chunk has a mix of foreign speech and English, please transcribe the English and place a [foreign] tag where appropriate. If your entire entire chunk is in a foreign language, please tag with [foreign] and report the file to the Help Desk within 10 minutes of submitting.
QAs, please cancel out of fully foreign files and report to the Help Desk.
Interactive Translator
In a file which has a primary foreign speaker, a primary English speaker, and a translator actively interacting between them, it might look like this (DO NOT label speakers on the Jobs tab; the speaker IDs are only shown here for clarity):
S1: [foreign].
S2: She says, "Please tell him everything I'm saying in English."
S3: Oh, great. I'd appreciate that. So my first question is, what is her actual job?
S2: [foreign]?
S1: [foreign].
S2: She says she works for TranscribeMe.
Translator Commentating
In a file which has mostly or only foreign speakers and there is a translator speaking over them to provide English for the entire recording, while not actually interacting with any of the people he/she is translating, we create a new paragraph for each foreign speaker, but they will all have the same speaker ID, since it is the same translator speaking English (DO NOT label speakers on the Jobs tab; the speaker IDs are only shown here for clarity):
S1: Okay, so now we are speaking Swahili, and Jude will translate so that it can be transcribed.
S1: That sounds great, thank you.
S1: First of all, what is your actual job?
S1: I am a transcriptionist working for TranscribeMe.
Nonspeech Tags Up Close
[music]
We transcribe the lyrics to songs when they are relevant to the file. How do you know if they are relevant? Well, as a transcriber, you may not have enough of a reference to determine whether it is or not. Here are some situations you may come across and how to handle them.
If the Speakers Discuss the Song
If the speakers discuss the song, it becomes relevant to the file and should be transcribed. Note: (singing) is NOT an acceptable tag and is only used for clarification purposes.
What's that song about all the lonely people? I can't think of what it's called.
Are you talking about Eleanor Rigby?
(singing) Eleanor Rigby picks up the rice in the church where a wedding has been. Lives in a dream.
Introductory Music
If the song (which may or may not contain lyrics) is an introduction to a podcast or seminar and is not specifically discussed, just use the [music] tag.
[music] From Preservation Maryland Studios in the historic podcast district of Baltimore, this is PreserveCast.
Music with Lyrics
But what if my entire chunk is a song with lyrics?
Since you do not have enough of a reference to make that call - you have no idea if the speakers are going to be discussing it in another chunk - you should transcribe the lyrics.
Note: If the file goes through QA and your chunk is rejected because the QA determined that it was not relevant and replaced the lyrics with a [music] tag, please fill out the Job Rejection Inquiry form.
There are some projects that specifically ask that all lyrics are to be transcribed. Always check the Style Guidelines for each job/file.
[applause]
The [applause] tag is considered an optional tag, but please do tag any relevant applause heard in the file.
- [applause] Thank you. Thank you. Please take your seats and let's get started.
- Next month, we will have Miss Terry Speaker. Thank you and good night. [applause]
[silence]
The [silence] tag is used when there is 10 or more seconds of silence, noises, or words (such as filler or feedback) that we do not transcribe. This can include coughing, sighs, background crowd noises, shuffling papers, etc.
What college courses are you registering for this semester?
[silence]
Well, I was thinking about Art History with Professor Snape.
[silence]
What is your major?
[laughter]
The [laughter] tag denotes when and what caused the laugh but NOT necessarily who did the laughing. Remember it is not necessary to tag every laugh, chuckle, chortle, guffaw, snicker, and giggle as long as the general mood is captured. In the below example, S2 was the one laughing, but what S1 said was what caused the laughter.
Correct:
So what do you get when you cross a joke with a rhetorical question? [laughter]
I don't know. What?
[silence]