Sie sind auf Seite 1von 11

Rulebook

Introduction
1. Format of the transcription
2. Simple rules (Dresing & Pehl)
3. Extended rules (Dresing & Pehl)
4. Clean rules
5 Anonymization
6 Frequently asked questions about transcribing
Introduction

When transcribing, many questions and ambiguities arise, especially in the early
days. The purpose of this document is therefore to provide you with a set of rules in
which you can hopefully find answers to most of your questions.

The first section informs you about the basic formatting of our transcriptions. The
formatting is the same for all transcription types. You will have internalized it after the
first few transcriptions.

Under points 2, 3, 4 and 5 you will find detailed explanations of the individual
transcription types. We have included various questions that were asked by our
freelancers. If any of your questions remain unanswered, please let us know.

Should you have any further questions or suggestions regarding the expansion of
this rulebook, they are always welcome.

Have fun expanding your knowledge!

Your Transkripto Team


1. Format of the transcription
In order to be able to offer our customers uniform products, it is very important that all
our transcriptions are transcribed and formatted in the same way. Never forget that
you may not be the only person working on an assignment. Therefore, please
observe the following formatting rules:
● Title of the transcription: Name of the file in bold and underlined, centered.
● Save as rtf file under the same name as audio file, but without file extension
(without .mp3, .wav)
● Font: Arial
● Font size: 12
● Line spacing: 1.5
● Text alignment: Left-aligned and character style: Normal
● Abbreviation for interviewer and respondent with I: and R: (not R.: or R :)
● In the case of multiple respondents or interviewers, the abbreviation is R: for the
first respondent and R2: for the second respondent, as well as I: for the first
interviewer, I2: for the second interviewer.
● Time stamper after each speaker change, except for "Content rules":
#hh:mm:ss-m#, for example: #00:00:15-1#.
● The last number after the hyphen is always a single digit and stands for
milliseconds. If the audio software does not display this, you can help yourself by
adding -0 manually.
● For files with only one speaker: Please set paragraphs and time stamps at
sensible intervals, every 2-5 minutes, depending on the length of the file. There is
no speaker designation here. Start directly with the text.
● Each speaker gets his/her own paragraph. A blank line is placed between the
speakers. Short interjections are also transcribed in a separate paragraph.
● Incomprehensible words are marked with (uncl.); suspected words are put in
brackets, for example: (Jerusalem?). For longer paragraphs, additionally with a
time stamper at intervals of about one minute. (uncl. #00:00:15-1#) (Jerusalem?
#00:00:15-1#) (also for content rules).
● Numbers are written out up to and including twelve. Round numbers such as
thirty, forty, one hundred are also written out.
Numbers are written out in increments of ten from 10 – 100 (e.g. sixty, but 62 as
numeral), in increments of hundred from 100 – 1000 (e.g. nine hundred, but 960
as numeral), in increments of thousand from 1000 – 10,000 (e.g. three thousand,
but 3,900 as numeral). Numbers from 10,000 on are transcribed as numerals.
● Notation of times: 7:30 a.m., 12:00 p.m. or twelve o'clock
● Special characters that stand for a word, are written out: e.g. percent, paragraph
● Foreign languages in an audio file: Please mark as (foreign language)
➔ The following is an example of how the formatting of your transcription should
look like:

14563892- Interview 3

I: So nice to see you. I am glad that we finally managed to have this meeting today.
#00:00:03-0#

R: Hi Helena, thanks for the invitation. Yes, same here. I am super happy that we
finally meet. #00:00:09-0#

I: Great. Okay, reason for this meeting are two things. First one, you wanted to create
a freelancer account on our website Transkripto. And the second reason, we wanted
to talk about the experience with subtitles and translations. #00:00:25-0#

R: Correct. I had some difficulties to create an account because I could not install the
two-factor authentication. #00:00:32-0#

2. Simple rules (Dresing & Pehl)


● It is transcribed verbatim, i.e. not phonetically or summarized. Dialects are to be
accurately translated into standard language. If there is no suitable translation for
a word or expression, the dialect is retained. For example: I ain’t going to school
today (British English dialect) --> I’m not going to school today.
● Informal contractions are approximated to written standard language. E. g.
“gonna” becomes “going to” in the transcript. Sentence structure is retained
despite possible syntactic errors. For example: "I had went to the store.” “She
gave it to Tom and I.”
● Discontinuations of words or sentences as well as stutters are omitted; word
duplications are only transcribed if they are used for emphasis (“This is very, very
important to me.”) Half sentences that are only missing the completion are
recorded and indicated by a slash /
● Punctuation is smoothed in favor of readability. This means that in the case of a
brief lowering of the voice or ambiguous emphasis, a full stop rather than a
comma is used. Units of meaning have to remain intact. Questions are marked
with a question mark.
● Pauses are marked by three ellipsis points in parentheses (...). Pauses are only
transcribed if they are particularly long or significant for what is said. However,
readability should not be impaired where possible.
● Comprehension signals from the person not speaking at the time, such as "Hm,
aha, yes, uh-huh, right" etc., are not transcribed. EXCEPTION: An answer
consists only of "Hm" without any further elaboration. Monosyllabic answers can
be transcribed in exceptional cases where it is significant. This is especially the
case when the comprehension signal indicates agreement. In this case, "Hm"
always requires an additional interpretation, e.g. Hm (affirmative). Hm (negative).
● Each speaker's contribution receives its own paragraph. In between speakers
there is a blank line. Short interjections are also transcribed in a separate
paragraph. Time stamps are inserted at the end of a paragraph. Example of a
time stamp: #00:00:32-5# - #hh:mm:ss-m#
● Speaker overlaps are marked with // with a space before and after the double
slash. The beginning of the interjection is marked by // . The text that is spoken at
the same time is then within this // and the interjection of the other person is in a
separate line, also marked by //.
See example below, “very good” and “to this topic” are spoken simultaneously.
● Emotional non-verbal utterances of all parties involved that support or elucidate
statements (laughter, sighs) are transcribed in brackets if of significance to the
content.
● Unintelligible words are marked with (uncl.). If possible, longer incomprehensible
passages indicate the reason: (uncl. cell phone ringing) or (uncl., microphone
rustling). If you assume a certain word but are not sure, put the word in brackets
with a question mark, e.g. (Xylometazoline?). Generally, all inaudible or
unintelligible passages are given a time stamp if there is no time stamp within one
minute. Often quick researching helps to find an incomprehensible word based on
the word fragment and context.

➔ Example of a verbatim transcription (Simple rules - verbatim):

I: Thank you very much for taking the time for an interview. #00:00:03-0#

R: Sure, that's no problem. I also want to serve science (laughs). No, seriously, it is
very, very close to my heart to be able to contribute // to this topic. #00:00:10-0#

I: Very good. // Then we would now start with the first question. What do you expect
from this interview? #00:00:14-0#
R: (...) But now you really surprise me with this question. #00:00:22-0#

I: We can start (uncl., microphone rustling) with something else. #00:00:25-0#

3. Extended rules (Dresing & Pehl)

These transcriptions are often needed for scientific analyses. The purpose of the
extended rules is to reproduce the conversation as realistically as possible.
● Word and sentence breaks are marked with /: "I was worri/ concerned." Word
duplications are always transcribed.
● Speaker overlaps are marked with // with a space before and after the double
slash. The beginning of the interjection is marked by // . The text that is spoken at the
same time is then within this // and the interjection of the other person is in a
separate line, also marked by //.
See example below, “very good” and “to this topic” are spoken simultaneously.
● Pauses are marked according to their length by ellipsis dots in parentheses. Here
(.) stands for approximately one second, (..) for approximately two seconds, (...) for
approximately three seconds and (number) for more than three seconds. Example:
pause of 6 seconds = (6)
● Comprehension signals and filler sounds of the interviewer ("Hm, yes, aha, um"
etc.) are always transcribed if there is no overlap. All utterances of the interviewee
are transcribed. This also means filler sounds such as "Hm" and "Um". In this case,
"Hm" always requires an addition as to how it is meant. E.G.: Hm (affirmative). Hm
(negatively). Hm (thoughtful).
● Emotions are transcribed: (laughs) (laughs) (hesitates) (inhales audibly) (sighs)
(clears throat) etc.
● Transcribe all non-verbal actions: (knocks on the table) (child's voice in the
background) (phone rings) etc.
➔ Example of an extended transcription:

I: Um, thank you very much for taking the time for an interview. #00:00:03-0#

R: You’re welcome. Hm (thoughtfully), that's not a problem. I have/ would also like to
serve science (laughs). No, seriously, it is very, very close to my heart to be able to
contribute // to this topic. #00:00:10-0#

I: Very good. // Um, well, then we would start with the first question. #00:00:12-0#

R: Hm (agreeing). #00:00:12-0#

I: So, what do you expect from this interview? #00:00:14-0#

R: (5) Um, you really surprise me with this question. I would like to, um/ #00:00:22-0#

I: We can start (uncl., microphone rustling) with something else. #00:00:25-0#

4. Clean/smoothed rules

In clean rules, the transcription is not verbatim. Instead, the focus is set on good
readability. These texts are often used for editorial purposes.
It is transcribed into correct and good standard language, especially regarding syntax
and grammar. It is especially important to understand the context to transcribe
smoothed rules, so that you can correct the syntax and grammar without changing
the meaning.
Keep in mind: In spoken language, people often correct themselves. You correct
these sentences by choosing the sentence structure that makes more sense in the
context. Therefore, it is extremely important that you understand what is being said
and that you can recognize the sentence structures.

Below are some rules that can be applied to smoothed transcription:


● Dialects are translated into standard language
● Confirmation receptors of the interlocutor can be bypassed (refers to signals like
"Hm", etc.)
● Filler: simply leave out. Fillers are words without which the sentence does not lose
its meaning. For example, in some context, the word “well”, "like", or “so”.
“Well, I, like, really don’t know what to say.”
● Write out currencies (e.g. Euro, Dollar)
● Set time stamps as usual. Example of a time stamp: #00:00:32-5#
(#hh:mm:ss-m#)
● Non-verbal utterances are not transcribed
● Unintended word duplications are not transcribed
● Fragmentary words or sentences without content meaning are not transcribed.
Fragmentary sentences with content-related meaning will be expanded or rebuilt
into complete sentences if possible.
● Incorrect syntax must be corrected.
● Colloquial words and phrases are replaced by standard language synonyms
● Contractions are smoothed out: “isn’t” to “is not”, “you’ve” to “you have” etc.
● Word abbreviations are smoothed out: "’cause" to “because”
● Pauses are not marked
● Syntax: Omit sentence breaks or word breaks. Sometimes speakers need several
attempts to formulate a sentence.
➔ Example: If a sentence is not completed but still contains information or
grammatical elements that are important for what follows, please insert "..." and then
continue in lower case.

➔ Example of a clean transcription:

I: Thank you very much for taking the time for an interview. #00:00:03-0#

R: You are welcome, that is no problem. I like to serve science. It is very close to my
heart to be able to contribute to this topic. #00:00:10-0#

I: Very good. Then we would now start with the first question. What do you expect
from this interview? #00:00:14-0#

R: You really surprise me with this question. #00:00:22-0#


I: We can start (uncl., microphone rustling) with something else. #00:00:25-0#

5. Anonymization

Simple anonymization 
Customers who order transcriptions from us can request simple anonymization free
of charge. This means that information that points to a specific person is
anonymized.  
Names of persons are replaced by [person], places by [place], and organizations and
companies by [organization] in square brackets. 
Important: In general, only information that gives an indication of a specific person
will be anonymized.
Example: Mister Schmidt has been working for Transkripto for ten years. 
Mister [Person] has been working for [organization] for ten years.

If information cannot be directly attributed to a person or if information is publicly


accessible, it does not have to be anonymized. 
Example: In Berlin there are many universities, e.g. the FU Berlin and the HU Berlin.
The president is Sabine Kunst.

Extended anonymization, including detailed breakdown (table/list)

Extended anonymization is about making anonymization more specific for our


customer. 
For simple anonymization, all names and places are always marked the same, i.e. as
[person] or [place].  
Especially in the case of transcriptions with a lot of information about people or
places, it is difficult to understand which person or place is involved if everything has
been anonymized in the same way. To simplify this for our clients, persons, places
and organizations are numbered.
In an additional document, the names and numbers are noted accordingly. In this
way, the customer has the possibility to check which number refers to which person
or organization
You can find the template in the download area in your account.

6. Frequently asked questions about transcribing

1.) What software can I use for transcribing?


➔ There are many different offers, some of which you can download for free, for
others you have to pay. Programs that we suggest are Express Scribe and f4.

2.) How important are the milliseconds or the correct format of the time
stamps?
➔ It is very important that you insert the time stamps exactly as we have shown
them in our examples in this file. If the time stamp does not correspond to the
required format, data analysis programs such as MAXQDA cannot read and analyze
the time stamps correctly.

3.) How do I know when to deliver a file?


➔ You can see the deadlines when you select the file, but also when you have
already downloaded it. Please respect the deadlines. We calculate them in such a
way that we still have enough time to check your file. It is very important that you let
us know if there is a delay. The deadline is CET.

4.) Control of audio and characteristics. What do I do if I don't understand


an audio recording well? Can I return it?
➔ Please check all audio files immediately for language, accent, quality, etc.,
especially if you take on more than one file at a time.
Should you ever catch a file whose accent you don't understand well or which you
simply can't understand because of the audio quality, you can return this file within
one hour. After this time, the file can only be removed from your account by us. When
you download a file, please listen to it immediately, even though you may not start
transcribing right away. It is very difficult for us to assign a file that is very close to the
deadline. You are welcome to write the file anyway and we will adjust the price after
consultation with you and the client.

5.) What do I do if the information in the file (e.g. number of speakers, audio
quality, language) is not correct?
➔ Please write us in good time. It is important that you write us if any of the
information in a file is incorrect. This way we can inform the client early enough and
adjust the price for the transcription both for the client and for you. Deviations in the
number of speakers by one or two people do not have to be communicated by email,
in this case it is sufficient to provide details when uploading the file.

6.) How do I know how good the quality of my transcripts is? Are my
transcriptions checked every time?
➔ In the first period, your transcriptions will be checked or corrected. As soon as
we have the impression that you get on well with transcribing or have good ratings,
checks will only be carried out on a random basis or depending on the type of file or
transcription.
Please do not upload a transcript if you are not convinced that the quality meets the
customer's requirements. In these cases, please send us an email in advance.

7.) What kind of transcriptions can I take on at the beginning?


➔ We recommend that our "beginners" take on shorter files and simple rules at
the beginning. In the beginning, there may be more mistakes, which can be corrected
gradually and with the help of our control sheets. Once you feel more comfortable,
you can also take on longer files, extended rules and express files.

8.) What happens if I get sick and can't finish a transcription?


➔ You are ill or have to take care of a sick family member? No problem, it
happens to all of us. The important thing is that you let us know early enough so that
we can find someone else to transcribe the file for us.

9.) What can I do if my question has not been answered here?


➔ Send us an email! We will continuously update this questionnaire in order to
answer all questions and to make transcribing as easy as possible for you. Please
get in touch if something remains unclear. That's what we're here for!

Have fun and success with Transkripto!

Your Transkripto Team

Das könnte Ihnen auch gefallen