Sei sulla pagina 1di 27

QuickStep Style Guide & Handbook

Intro to QuickStep 2

QuickStep Expectations 2

QuickStep Style Guide 3


Clean Verbatim 3
False Starts and Interruptions 4
Tags 5
Sentence Structure and Punctuation 7
Spelling and Formatting 8
Research 8

The Sessions Screen 9


Audio Player 10
Transcribing a Session 10
Timestamps and Speaker Identification 11

QuickStep Policies 16
Payments 16
Turnaround Time 16
Help Desk Extensions 16
Extension Guidelines 17
Session Drop Ratio 17
Policy for Lost Files 17

Understanding the Review Process 19


Probation Period 19
Review Form 20
Finding your Ranking 21

Reminders 22

Support 22
Yammer 22
Foot Pedal 22

FAQ 23

Edition 2.21.2018 Quick​Step​ Handbook 1


Intro to QuickStep
QuickStep​ is a team of high-accuracy single-step QAs who process general ​Clean Verbatim​ files using a
simplified​ ​version of TranscribeMe's General Transcription Style Guide. The purpose of this team is to
provide a faster turnaround time for TranscribeMe clients, while still meeting TM's ​guaranteed​ ​99%
accuracy ​standards. ​Single-step QA​ (Quality Assurance) means that the entire file is transcribed, proofread
for accuracy, and time-stamped with speaker identification by one person - you!​ Carefully adhere to the
following guidelines​ to produce quality transcription. Failure to do so may result in the loss of your
QuickStep status.

QuickStep Expectations
● Accuracy: ​You are expected to capture all speech relevant to the audio without editing,
paraphrasing, or otherwise changing what the speakers say. Every effort should be made to make
out even difficult audio.
● Consistency: ​Each file must be consistent within itself in regards to punctuation, numbers,
symbols, spelling, etc.
● Timestamps and Speaker IDs ​must be accurate and timestamped within two seconds of the
speaker's first utterance. Each speaker is assigned the same ID throughout the file, based on the
order in which they appeared in the audio. See p. 11 for tips on speaker identification.
● Research: ​Every name, company, and term mentioned in the audio should be researched to
determine the correct spelling and formatting. If there is more than one accepted spelling, choose
one and be consistent).
● Professionalism: ​You are expected to have a strong grasp of correct sentence structure and
punctuation. Be sure to thoroughly proofread for typos or formatting errors before submission. ​All
QuickStep files should begin with a capital letter and end in a period​ (or question mark), even
mid-sentence. They should also be broken up into logical paragraphs, even in single-speaker files.
● Guidelines ​located to the right of the workhub screen should be strictly followed.
● Activity​: You are expected to complete a minimum of 2 audio hours per month to remain on the team.
Please fill out an ​inactivity form​ to provide notice of short absences.
● Turnaround Time:​ At least ​75%​ of your QuickStep files should be submitted by the due time.
● Drop Rate:​ No more than ​10%​ of your sessions can be dropped.

Edition 2.21.2018 Quick​Step​ Handbook 2


QuickStep​ ​Style Guide
Clean Verbatim
QuickStep files should be transcribed in ​Clean​ ​Verbatim​ (CV). This means that:
● Stutters and stammers ​should be removed:
No​ → ​Re-​repeated ​s-s​-​sounds such as ​st-st-st​-​stammers and stutters should not ​be,​ be
tr-​transcribed. ​This​, this includes when a speaker ​says,​ says the same word twice.
Yes​ → Repeated sounds such as stammers and stutters should not be transcribed. This includes
when a speaker says the same word twice.
Yes → Yeah, yeah, yeah​. But it is ​very​, ​very​ important to include words that are repeated for
emphasis, because people ​do do​ that sometimes.
● Filler​ ​words ​-​ ​such as uh, um, hmm, mm-hmm - ​should be removed​. However, if such words ​answer a
question​ or ​add meaning​, you may use them - sparingly - using these spellings:
Affirmative​: uh-huh​,​ mm-hmm.
Negative: ​uh-uh​, ​hmm-mm​, ​mm-mm​, ​huh-uh​, ​nuh-uh​.
● Thinking​ ​sounds​ ​should be ignored​:
No​ → I want a ​buh, buh, buh, hmm…​ pizza
Yes​ → I want a pizza.
However, non-standard words that have meaning in the sentence can be included according to what the
speaker actually says. Please check Google and use the most common spelling you can find.
Yes​ → Then you do the same steps, ​da​, ​da​, and then ​da​. That's all it is.
● Informal​ ​truncations​ should be transcribed using the proper spelling, ​not​ based on ​pronunciation​:
No​ → ​gonna​, ​gotta​, ​wanna​, ​kinda​, ​sorta​, ​coulda​, ​'cause​/​cuz
Yes​ → ​going to​, ​got to​, w
​ ant to​, ​kind of​, s​ ort of​, ​could've​, ​because
● Pronunciation​ ​based​ ​on​ ​accents​ should be ignored, including words ending in -ing.
No​ → ​Pacifically​, I'm ​gonn'​ go ​ta​ that there ​stoar​ and all ​'em​ stores where they be ​sellin'​ milk.
Yes​ → ​Specifically​, I'm ​going to​ go t​ o​ that there ​store​ and all ​them​ stores where they be s​ elling
milk.
● Regional​ ​slang​ ​words​, however, ​should be included​, using spellings found in a dictionary. For
example, if the speaker says, ​Gotcha. Y'all ain't ​gonna​ believe this​:
No​ → ​Got you​. ​You all aren't​ going to believe this.
Yes → Gotcha​. ​Y'all​ ​ain't​ going to believe this.
If the speaker actually says, “Got you,” etc., then transcribe it as such.
● Incorrect​ ​grammar​ should be transcribed as spoken​; do not ​edit​ their speech for them. For example,
There's so much apples in them trees:
No ​→ ​There are​ so ​many​ apples in those trees.
Yes​ → ​There's​ so ​much​ apples in ​them​ trees.
● Contractions​ ​should be transcribed as spoken​. If the speaker says ​they're​, transcribe ​they're​. If they
say ​they are​, transcribe ​they are​.
● Conjunctions​ ​and​ ​interjections​ - ​and​, ​but​, s​ o​, ​or​, ​then​, ​well​, etc. - at the beginning, middle, or end of
sentences​ should be transcribed.
And​ I wanted to get a new dog, ​but.
Yeah, yeah, yeah. ​So ​why didn't it work out?
Well, ​it was like, “This dog won't work.”
Yeah. ​But​ that's too bad.
And ​he kind of smelled, ​so.
In other words, we correctly ​spell​ what the speaker says, regardless of ​how​ they say it - regardless of
accent​ or ​pronunciation​ - but we do not ​edit​ or ​alter​ ​what​ they say, incorrect though it may be.
Edition 2.21.2018 Quick​Step​ Handbook 3
● Crutch Words and Phrases​ - such as ​like​ and ​you know​ - can be excluded when they don't add any
meaning or relevance to the sentence. Do not remove too much, especially with phrases such as ​kind
of ​and ​sort of​ that often add meaning to a statement.
I'm​ ​sort​ ​of​ ​tired. It was ​like​ 1:00 AM when I went to bed. That's ​kind of​ l​ ate, ​you know​?
But, ​you know​, if the speaker, ​like​, talks like this, then, ​you know​, you can, ​like​, remove the crutch
words to make it, ​you know, like​, readable.
● Feedback Words​ - ​At TranscribeMe, we remove words that someone says while another person is
talking, when the ​sole​ ​purpose​ of these words is to let the speaker know they are still listening. ​DO
NOT interrupt the main speaker's speech like this:

The feedback words in the example above make the text difficult to read.
Instead, we remove the feedback, leaving us with a much cleaner transcript:

Feedback words can be removed when they:


A) Sit on a line by themselves
B) Don't answer a direct question
C) Aren't acknowledged by anyone
D) Add no meaning and are not directly commenting on the topic at hand

However, ​do​ ​include ​longer phrases that ​add meaning​ or serve to ​comment​ on the speaker's statement.
So that's pretty much the plan we've established.
Sounds great.​ ← Do ​not​ remove
And now we can get started.
Yeah. Right.​ ← Remove
Yeah. Right. Okay. That sounds great. ← Do ​not​ remove
● Background​ ​Conversation​,​ such as that in a restaurant, should be omitted. We transcribe all
speech/videos relevant to the file or anyone that ​interacts​ ​with​ the main speakers in the file. ​This
includes when an interviewer mutes the phone to have a side conversation with a coworker.

False Starts and Interruptions


Some situations can be dealt with a number of different ways. The most important thing is that you are
consistent in how you treat these situations throughout the file. For example:
● False​ ​Starts​ -​ ​Speakers commonly make mistakes or change the direction of their thoughts, which can
prompt them to begin a phrase or sentence over. This is referred to as a false start. In QuickStep style,
you may indicate false starts however you feel is appropriate - with single dashes, double dashes,
ellipses, capital or lowercase letter for the next word, etc. - ​but be consistent throughout the file.
Some people might like to​-​ you could do this, for example.
And if they really​--​ If the sentence is starting afresh, why not use a capital?
And, I mean, there are times​...​ If they just trail off and then begin again, you could even do this.
● Cleaning Up False Starts​ - ​If the speaker corrects him or herself a few words into a false start, you
may omit the error and transcribe the correct sentence, removing the false start.​ ​QuickStep does not
have a word limit for false starts that should be omitted, but please use good judgment. Do not over
Edition 2.21.2018 Quick​Step​ Handbook 4
edit speech or leave out anything that changes the meaning of what was said. Remember, when in
doubt, type it out.
With false start: I went to ​the mall​-- the store, and I bought some milk there.
Corrected false start: I went to the store, and I bought some milk there.
● Interruptions​ - ​When speakers interject or talk over one another, the interruption may be formatted
with double or single dashes, or ellipses, etc., but be consistent. Note that the dashes are connected -
with no space in between - to the final word before the interruption and the first word after the
interruption - with no space between the dashes and the timestamp.

You could also use ellipses:

If the speaker is interrupted while asking a question, insert a question mark immediately after the double
dash or ellipses.

If the speaker continues the same question after the interruption, place the question mark at the end of
the completed question.

Tags
QuickStep uses the same tags as the general transcription pool, but ​DO NOT ​use the [guess?] tag.​ If you
are fairly certain of the word/phrase and it makes sense ​based on context​, transcribe it; otherwise, use
[inaudible] or [crosstalk].​ ​Tags are ALWAYS in lowercase letters. ​Make snippets for your tags to avoid
mistakes.
● [inaudible] ​←​ ​always​ ​lowercase
This tag ​takes the place​ of a word or words you cannot make out due to a difficult accent, poor audio
quality, background noise like a cough obscuring the words, or even a word that you couldn't find through
research.
● [crosstalk] ←​ always lowercase
This tag is is NOT used to “represent” ​where​ crosstalk took place. Like the [inaudible] tag, it is only used​ ​to
stand in for words that you cannot understand, ​due to another main speaker talking over them.

In the above example, stamp S2 at the utterance of “Yeah,” - ​do not wait​ for S1 to stop speaking to time
the stamp. The stamp matches the ​first​ utterance of its respective speaker, regardless if anyone else is still
talking.

Edition 2.21.2018 Quick​Step​ Handbook 5


If two speakers are talking over one another but you ​can​ understand one or both, then include the words
spoken, and there is no need for the [crosstalk] tag.
Did you get the papers and file them?
Yeah, I filed the papers, definitely.
If S2 says something in the middle of S1's speech but you cannot make it out, then you need to add the
[crosstalk] tag ​on a new line​ (new speaker = new line) with a timestamp for S2.

Please make every effort to make out all spoken words before resorting to using the [inaudible] or
[crosstalk] tags. ​The most important thing to emphasize is the need to listen to the audio to make sure it
makes sense. If something doesn't make sense, it is usually a mishear. If there are trouble spots, please
re-listen to the segment you are having trouble with. Do not immediately resort to [inaudible] tags. Also
make sure that sentence structure is clear and coherent. ​Do not include words that make no sense​ to the
context of the file.
Here are some useful extensions for dealing with troublesome audio:
Ears for Chrome Boom 2
FxSound Hear

● [foreign]​ ​←​ ​always​ ​lowercase


If one or more speakers are ​speaking​ in a foreign language, please use the [foreign] tag.
If, however, they are ​speaking​ English and use a foreign word or phrase that you can find with an internet
search, transcribe what they say.
So he turns to me, and he goes, "Whatever you want is fine. Mi casa es su casa, as they say."
She was speaking French, though, and so she used the word maison instead of house.
Is expelliarmus really a Latin word?
She put out a new book titled Mi Estas La Plej Bona.
If you can't find the term after searching, then use the [foreign] tag instead of [inaudible].

Nonmandatory Tags - [laughter], [music], [applause], [silence]


The following tags may also be used, but are not required.

● [laughter]​ ​←​ ​always​ ​lowercase


This tag is placed at the point where laughter ​occurs​, ​regardless of who is laughing. The [laughter] tag
should ​never​ ​appear on its own line​.
Sometimes there are small laughs scattered throughout a conversation. It is not necessary to capture
every giggle or chuckle, as long as the general mood is captured. In the following example, assume it is S2
who burst out laughing at the end of the joke.

● [music] and [applause]​ ​←​ always lowercase


These tags ​do not​ affect ​when​ you timestamp. The times on the stamps represent where ​human speech
began, not laughter, music, applause, or any other sound.
[music]​ ​–​ ​This tag is used to note music relevant to the recording, such as the introductory theme music to
a podcast. ​Do not​ use this tag for ​background​ music, e.g. something playing on the radio or at a restaurant
(unless the main speakers mention the music).
Edition 2.21.2018 Quick​Step​ Handbook 6
[applause]​ ​–​ ​If there is applause included in the recording, please use this tag. Like [music], the [applause]
tag does ​not​ affect the timing of the stamp.
For example, if your file begins with 15 seconds of music before any speech occurs, type the [music] tag,
then transcribe the first sentence, then place your cursor before the music tag and stamp at the first word,
“Good.” It will look like this:

● [silence]​ ​←​ ​lowercase, on its own line, NO punctuation OR timestamp


If there is no human speech, music, or applause for a period of ​ten or more seconds​, then use the [silence]
tag, regardless if there is other “noise” - cars, dogs, coughs, etc. - or not. This tag rests ​on its own line​ and
has ​NO timestamp OR punctuation ​at all.

● Other​ Noises
Non-verbal or non-human sounds - coughs, sneezes, cars honking, dogs barking, nuclear explosions, etc. -
should be ignored. We only transcribe ​speech​ and use the tags specifically mentioned above. ​DO NOT use
any tags not listed here.

Silent and Foreign Files


If you come upon a session that is ​entirely​ silent or foreign - check the beginning, end, and several
selections from the middle to be sure - ​note the session ID, cancel out of the session, and send a ticket to
the Help Desk​ with the ID. If there are some small parts in English, please transcribe those and then use
the [foreign] tag where necessary. Just one [foreign] tag will be sufficient for multiple foreign speakers in a
row. If you do not report these types of sessions, you risk losing the full pay for that file.
Sentence Structure and Punctuation
You are expected to demonstrate a strong grasp of basic sentence structure and punctuation in your work
at TranscribeMe. Please see ​this guide​ for brief explanations and examples of the following concepts:

Run-on Sentences Comma Splices


Sentence Fragments Hyphens
Commas and Conjunctions Plurals vs. Possessives
Separating Verb & Subject Punctuating Quotations

● Quotation​ M ​ arks
Do​ use quotation marks for direct and hypothetical speech and “spoken punctuation.” (e.g., if a speaker
says ​quote​ or ​quote, unquote​. In QuickStep, you may also use quotation marks for media titles, slogans,
etc., if doing so aids readability and conveys the speaker's intent.
● Dashes
Single​ dashes may be used to set off additional information within a sentence - like this - or ​double​ dashes
-- like this -- providing the speaker's intent is conveyed and the sentence flows well.
● Ellipses
As previously mentioned, ellipses may be used to indicate a false start or a speaker trailing off. However,
DO​ ​NOT​ ​use them ​in the middle of a sentence to indicate that the speaker paused, regardless of the
length of that pause​.
No​ → I think I want​…​ that one.
Yes​ → I think I want that one.

Edition 2.21.2018 Quick​Step​ Handbook 7


Spelling and Formatting
In QuickStep, you may reference any well-known dictionary for spelling. Always check the file guidelines to
know whether US or British English is required (see Guidelines, p. 10). Also remember that your transcripts
are ​professional ​documents, not social media posts. Choose spellings accordingly, and be consistent.
Please make use of the following free resources:
Merriam-Webster​ - US ​English Cambridge​ - BE ​and​ US ​English
Oxford​ - ​BE​ and ​US​ English Collins​ - ​BE​ and U
​ S​ English
At TranscribeMe, we use the -ise endings for BE, NOT -ize.

Spelling Words Out


If a speaker spells a word out, you may follow any of the following formats. Just make it readable, and be
consistent.
My last name is Mehlhaff, M-E-H-L-H-A-F-F.
or​ M E H L H A F F ​ or​ ​m-e-h-l-h-a-f-f

A single letter mentioned on its own may be lowercase ​or​ capital.


The next letter after ​J​ is obviously ​K​.
Did you know that ​x​ plus ​y​ equals ​z​?

Email Addresses
● If they say, "rini@transcribeme.com," then we transcribe: ​rini@transcribeme.com​.
● If they ONLY say (spell out), "R-I-N-I at T-R-A-N-S-C-R-I-B-E-M-E dot C-O-M", then we transcribe:
rini@transcribeme.com​.
● If they say BOTH, "Rini, R-I-N-I, at transcribeme, T-R-A-N-S-C-R-I-B-E-M-E, dot com, C-O-M," then we
transcribe: Rini, R-I-N-I, at transcribeme, T-R-A-N-S-C-R-I-B-E-M-E, dot com, C-O-M.
● One more example: Email me at name, that's N-A, M as in Mommy, E, at gmail.com.

Numbers and Symbols


QuickStep does not have set rules for formatting numbers and symbols. Use your best judgment, convey
the speaker's meaning, and ​be consistent​. If you need additional guidance, check out any well-known style
guide such as the Chicago Manual of Style or the AP Stylebook.

Research
Always use Google to search and format given names correctly. ​It is your responsibility, as part of TM's
99% accuracy guarantee, to research every name/company/term mentioned in your file​. In the case of
company names, go by the name in the main text of their website, not the logo, as those are often
stylized, for example:
adidas / iPhone / athenahealth / MEDITECH / Toys”R”Us
Exception​: Do not use all caps for brand-name medications.
No​ → She takes ​OCREVUS​, ocrelizumab, for multiple sclerosis.
Yes​ → She takes Ocrevus, ocrelizumab, for multiple sclerosis.
Yes​ →I had no idea that Tylenol is actually just acetaminophen.

Check out these ​tips for Google searches​.

Edition 2.21.2018 Quick​Step​ Handbook 8


The Sessions Screen
Please reference the ​TranscribeMe TR Handbook ​for more information on understanding the
WorkHub. The ​Sessions ​tab will be added to your WorkHub once you have been promoted to a
QuickStep QA. The ​Jobs ​tab will still be there, allowing you to continue to work as a transcriber when
desired. You will also notice different folders that might contain QuickStep sessions. If the client picks a
category when uploading a file, it will show up in the folders. The ​Sessions ​tab contains all the QA files
from QuickStep or from any special teams you might be assigned. Files located in the ​Lower
Expectations folders ​are sometimes paid at a higher rate, as these files have been identified as difficult
audio. ​Please do not use the Lower Expectations folder while on probation.
When you are ready to work on a QuickStep file, click on the Sessions tab and this will open the work
area. A few things you need to note that are very important:

1. Session Rate​: ​This is the rate you will be paid for the audio, by audio hour. The default pay for
general QuickStep work is US$40 an ​audio​ hour. The amount paid will therefore depend on the
duration of the audio, not how much time you spent working on it.
2. File​ ​ID​: ​Always write this number down ​FIRST ​thing. If you have an issue with the file or need an
extension, you will need to reference this ID.
3. Priority​: ​The lower the number, the higher the priority of the file. Each session has its own priority,
ranging from 1-10, with 1 being the top priority. High priority files usually mean that the file is in
danger of being late or is a rush file for a client. We ask that you complete high priority files as
accurately as you can but note that they tend to be time sensitive. Please do not keep priority 1 files
overnight.
4. Duration​: ​This indicates the full length of the file.
5. Due​ ​date​: ​When you pick up a file, the due date and time of the file will be listed at the top. This
number ​does not​ indicate the amount of time the file will stay in your queue. The clock counts down
showing the time left until the file is ​due​ to the client, in days hours:minutes:seconds. When a file
becomes overdue, the text will be red. We know that sometimes you will pick up a file that is
already overdue. We ask you to do your best to expedite processing on those files, but you will not
be held accountable for the late nature of the file. See the section on Extensions below for more
details on what we expect in terms of file deadlines.

Edition 2.21.2018 Quick​Step​ Handbook 9


6. Description​: ​If the client inputs additional information about a file, it will show up as a description
beneath the due date. ​DO NOT follow any instructions given by the client in the description field.
Clients are instructed to contact the sales department with any special requests. The description
may give you some insight on the correct spelling of names and/or places, but we always follow the
guidelines listed for each session. If you are ever unsure about anything in the description section,
please create a ticket on the Help Desk and ask.

Guidelines
VERY IMPORTANT!​ Always check and expand the Guidelines for each file
before starting. This is where you will find out if the file is Clean or Full
Verbatim. The file might also have special instructions about speakers or a
different style that you might need to follow. Clicking on QA Guidelines will
open another box. You can then click on the different lines within the guidelines to expand each section.
This will give you more information about the style. This is especially important if the file does not follow
the standard default guide.
The guidelines may have letters behind them specifying which English spelling -
US or British - to use. ​US English is our default unless otherwise noted​.
● NA - Not Applicable: The client did not specify, so use US English.
● AE - American/US
● BE - British English
● UE - Aussie English: ​expand guidelines​ for ​Aussie exceptions
● KE - Kiwi English: ​expand guidelines​ for ​Kiwi exceptions
● IE - Irish English
● SE - Scottish English
You will need to pass the appropriate accent exam to receive files other than US English. These can be
found under the ​Exams​ tab.

Audio Player
The audio player works the same as
on the Jobs tab but with a few more
options:
● If a file was uploaded as an mp4 video, you will see an orange arrow. Click it to open the video.
● The ​Non-Interactive playing​ mode is most appropriate for QuickStep (and FV) files. If your text or
audio is jumping around or doing other strange things, first check that it is in this mode.

Transcribing a Session

When you open a new session (that hasn't been worked on yet by another QA), you will see an orange box
with white dots. Click on the box, then start typing, and the text area will turn blue. You may timestamp
QuickStep sessions as you transcribe, or transcribe the file then go back and timestamp, or break the
process up into sections. Find what works for you​.

Edition 2.21.2018 Quick​Step​ Handbook 10


Q: How should I start my session?
A:​ With a capital letter. Yes, even if the speaker starts mid-sentence. If the speaker starts mid-word, then
you can leave out that partial word or, if you can hear almost the whole word and are certain what it
should be, you can include it.
Q: How should I end my session?
A:​ With punctuation, namely a period or a question mark. If the speaker ends mid-word and you are
reasonably sure what that word should be, then please include it. Otherwise, you can leave out the partial
word at the end and simply close with punctuation as though it were the end of their sentence.

Timestamps and Speaker Identification


Inserting timestamps and speaker IDs is one of the most important jobs of the QA. Timestamps​ identify
speakers​ and ​mark the time​. You insert timestamps by pressing the Ctrl key along with the speaker
number. For instance, if you are inserting a timestamp for the first speaker, you hold Ctrl and then press
the number 1. The second speaker will be Ctrl +
2, and so forth. The time on the stamp will match
the time you press the number key. You can click
on Hotkeys under QA Guidelines if you need a
reminder.

Timing the Stamp


Timestamps are inserted when the ​first​ ​word​ is uttered, even if they begin with false starts and filler
words - um, uh, etc. - that are omitted from the transcript.
When you are first learning to timestamp, it may be helpful to ​pause the audio to place timestamps​ to
ensure accuracy. As soon as the speaker starts to make a​ vocal​ ​noise - ​not​ a laugh, sneeze, inhale, etc. -
press Ctrl + / to pause. If you miss by a little bit, press Ctrl + / again - to play - and the audio will skip back a
second or two when it resumes. Once you get it paused at the right point, you can insert the
timestamp/speaker ID. This is especially useful when speakers interject at virtually the same time.
Q: Do I insert a space after a timestamp?
A: ​No space is needed after a timestamp. The text can be snug up against the stamp.

Speaker IDs
Speaker IDs are always inserted in the ​order that they speak​ in the file. The number you use for the
stamp will correspond with their position within the file.
Only the first nine speakers can be inserted directly by pressing Ctrl and a number. If your file has ten or
more speakers, then you will need to first insert a timestamp using one of the normal hotkeys, and then
click on the orange timestamp box and select or type the correct speaker ID for each timestamp.
Note how a file will look with different timestamps and speaker IDs.

Edition 2.21.2018 Quick​Step​ Handbook 11


Identifying Multiple Speakers
Multispeaker files can be tricky, but they become easier with practice. Create a ​voice map​ in a notebook or
on a separate program opened next to your screen. As the different speakers are introduced or speak in a
file, note the speaker IDs, names, places, and any identifying information you can about them, such as:
● Gender
● Distance​ from mic - This gets tricky if the speakers move around throughout the file
● Age​ (approximate) - You can tell age by speech habits, vocal quality
● Accents
● Vocal quality​ - Loud, quiet? Deep, high? Scruffy, clear? Resonant, nasally? Sound like a celebrity?
● Speech habits​ - Any words or phrases they use a lot? Fast, slow talker? Stutter or stammer?
● Personal information​ - Kids, activities, location, companies, schools, pets, hobbies, etc.
● Personalities​ - Outspoken, soft-spoken, rude (interrupts others a lot), polite, etc.
For example:
S1 Alex, young male, deep, resonant voice. Says “like” and “right” a lot.
S2 Casey, middle-aged female, bright/upbeat voice. Professor at UCLA. Sons named Jason and Todd.
S3 Todd, Casey's son. Medium voice, outspoken. Guffaws while talking. Backpacking fanatic.
It can be difficult to make absolute identification just by voices, but keeping notes goes a long way to at
least making an educated guess. Make sure to destroy any type of personal information you have written
down once you have completed the file.
Q: Is there any way to check whether I got the right speaker?
A: ​Click on a timestamp to bring up the purple speaker's window. You'll see a little play triangle within
each speaker's row. This will play a snippet of that speaker's voice to help you pick out the correct one.
You can use the forward and back buttons in this window to cycle through all of the timestamps you've
marked for that speaker, to compare the audio at each point and verify that you've picked the correct
speaker ID.
Q: What if I can't figure out who is speaking?
A: ​As is the case with making out difficult audio, you should make every effort to correctly identify the
speakers in your session. If for some reason you are not able to identify the speaker (such as for a short
whisper or an inaudible comment), you can insert ​S?​. This is done by using ​Ctrl + 0​. Use this only as a last
resort.
S?​ is also used in [crosstalk] situations where so many people are talking that it's impossible to make out
any one thread of conversation. In that case, you would put the [crosstalk] on its own line and stamp it
Edition 2.21.2018 Quick​Step​ Handbook 12
with S?.
Correcting a Timestamp
As you progress in a file, you might realize you have made a mistake on the speaker ID. You can correct the
number of the speaker without having to delete the full stamp. If you are unsure of a voice and want to
listen to a certain speaker, click on the little arrow. It will play the sections you have assigned to that
speaker.

Adding Names to Timestamps


There will be some files that a client has requested a
name be used for the speaker instead of S1 or S2. When
this request has been made, the timestamp editing box
will display a bit different than the normal. You will have
an extra box that will display as ​edit names​.

When you click on ​edit​ ​names​, another box will open that will allow you to insert a name.

Once you have input the names, remember to click ​Save​.

In the session, you will now see the speaker names instead of S1
or S2.​ Remember to note which speaker ID belongs to each
name​; to insert the IDs you will still press Ctrl + 1 or 2, etc. So
Tom would still be Ctrl + 1, Mary Ctrl + 2, etc. If, for some reason,
you do not know the name then you can use the S1, S2 and so
forth.

Timestamp names are only used if noted in the Guidelines, NOT in the Description at the top of the
session. ​Remember to always check the guidelines to make sure you are inserting the speaker IDs required
for the file.

Edition 2.21.2018 Quick​Step​ Handbook 13


Paragraphing
You already know that each change of speaker gets a new line or paragraph. Yet an integral part of the
QA's job is to break up long ​single-speaker​ monologues​ into paragraphs. Clients DO​ ​NOT​ ​like reading
solid blocks of text. Break up long-winded speech at or around:
● A clear ​change of topic
● 10 to 15 lines​ ​of text​ ​in Preview​ mode
● 1.5 to 2 minutes​ of the same speaker
Proofread the document in Preview mode and find the most logical place to break up overly long
paragraphs - based on the above criteria - even though the separate paragraphs may not be even in
length.
Q: Do I need to timestamp where I break the paragraphs?
A: Yes​. Every paragraph needs a timestamp, even if it is the same speaker as the paragraph before.

Saving and Extending a Session


Q:​ Can I look through the queue and pick out a file to work on later?
A:​ ​No​. ​Never​ pick up a file until you are ready to begin working, as we are on a tight schedule with our
clients. ​You must transcribe ​five minutes​ of audio before saving​ and extending your session, to give you
time to decide whether or not you wish to complete the file.

Save​:​ After you have completed ​five minutes​ of transcription, ​you may save and extend your session.
Clicking ​Save​ will save your work and show you the amount of time the session will be assigned to you.
The ​default​ amount of time a session will remain in your queue is ​one hour​.
Self-Extension:​ ​If you open up a session that is not yet nearing
its due time, you can self-extend ​up to 10 hours ​by clicking on the
Keep Session for​ dropdown menu at the bottom left-hand side of
the screen. Once you have ​completed ​five minutes​ of
transcription ​and selected the desired extension time for the file,
press ​Save​. A pop-up box will display the time remaining until
your self-extension expires.
DO NOT self-extend​ a file until you have completed at least ​FIVE
MINUTES​ of transcription and intend to finish the file.​ ​Extensions
are to allow you a break​ AFTER you have already started working
on a particular file, not to set one aside to work on later.
As a file nears its due time, the amount of time available for self-extension will decrease until, eventually,
no self-extension option is possible. If you need more time on the file, create a ticket on the Help Desk and
request an extension according to the policy below (Help Desk Extensions). Only ask the Help Desk for an
extension if you can complete the file in the allotted due time.
Save & Exit​:​ ​This will save your session and take you out of the Sessions​ ​screen (this option ​does not
show an accurate amount of time remaining - always click ​Save​ first to see the amount time remaining). To
get the session back, click on the Sessions​ ​tab on the WorkHub and it will reopen your saved session.
Again, the session will only be assigned to you for ONE hour unless you have requested an extension. If
you log back into the WorkHub after one hour, and have not requested an extension, you will see a
different session come up and any work that you have done before will be lost.
Cancel​:​ ​If you open a session but do not wish to work on it, click ​Cancel​. The session will then go back to
Edition 2.21.2018 Quick​Step​ Handbook 14
the queue to be picked up by the next available QuickStepper. You may then again click the ​General
button under your Sessions tab to browse for more sessions
Reject this Session​: ​When you click on ​Reject this Session, ​the rejection reason box will display. If you
come across a file that, for whatever reason, you absolutely do not want to work on, rejecting a session
will cause the file to be removed from your queue. If you have rejected all the files available, the no
session available message will display when you click on the Sessions tab.
● Jargon​: ​This means that you are not able to make out any words in
the file. The voices/words can't be heard or it sounds like a foreign
language.
● Low​ ​Quality​ ​Audio​: ​You feel the audio quality is too low for you to
complete the file.
● No​ ​Sound​: ​The audio player is moving but no sound can be heard.
Please make sure that you troubleshoot your computer. 99% of the
time clearing out your cache or checking your mute button will help.
● Terrible​ ​transcription​: ​Only applies to 2-step QAs; see QA Handbook
● Won't​ ​Play​: ​The audio player is not moving and no audio will play.
Please make sure that you troubleshoot your computer.
● Audio​ ​Length​: ​The file length is longer than the time you have
available. Please, do not fear those longer files. We will always give you as much of an extension on
the file that we can, up to 40 real hours per audio hour. Create a ticket on the Help Desk to request an
extension as needed.
If you have a reason for rejecting that is not listed, please choose Audio Length.
Submit​: ​Once you have completed the session and are satisfied that the transcript is in perfect shape,
ready for the client, you can click ​Submit​! A pop-up window will be displayed, asking you to confirm. Once
you click ​Submit​, the transcript will be sent to either the client or to a Reviewer. When you click Submit,
the page will reload and you will receive your next session for QA. ​Create a Help Desk ticket immediately
if you accidentally submit a session so they can return the file to you. ​If you did not write down the
Session ID (which you ​should have done first thing​), you can retrieve the ID from your Work History page.

Edition 2.21.2018 Quick​Step​ Handbook 15


QuickStep Policies
Q: What is the minimum quota of files to be completed?
A: ​For QuickStep, we ask that you aim to complete one audio hour a week, but we only require​ two audio
hours a month​. This is to ensure that we can keep a good estimate of how much work we need to set
aside for the team on a daily basis. If you need to be away for a short period, please notify the team of
your absence using this form: ​https://transcribeme.wufoo.com/forms/z14bo9ah08e29v7/

Payments
The rate of your QuickStep file will always be displayed at the top of the file. Sometimes different teams,
groups, and files will have different rates.
TranscribeMe does not pay for partially completed files. You must complete the full file to receive
payment.
All work submitted is subject to review prior to being approved and payment being made. The standard
expected quality level is 99% minimum. Work must be submitted in accordance with provided
instructions, style guides, and format. We reserve the right to reject work, with or without pay, at our
discretion, for work that does not meet the 99% quality threshold.
Work may also be reviewed by the admin team after the submission in complete. Work that does not
meet the 99% quality threshold in this admin review may result in additional probation files or removal
from the team.

Turnaround Time
The expected ​turnaround​ ​time​ for QuickStep files is ​48 hours​. This time frame starts from the time the
client uploads the file to the time the completed transcript is submitted. ​As a QuickStepper, at least 75%
of your sessions should be submitted on time.
After picking up a file, that file will stay in your Sessions​ ​tab while you are actively working on the file. If
you stop working on the file for longer than ​five minutes​ - this includes leaving the session screen and
going to another tab on your computer - the hub will save your file for up to ​one hour ​if you have NOT
self-extended or asked for an extension from the Help Desk. If you do not come back and continue working
on the file within that time, or if you have not added an extension to the file, you will lose the file.

Help Desk Extensions


If the Self-Extension option does not display, this means that the Help Desk will need to manually apply an
extension. A manual extension can only be requested ​after you have transcribed at least ​five audio
minutes​.​ This will give you time to make sure that you can commit to the file.
Create a ticket on the Help Desk, giving them the Session ID. The Help Desk will manually extend the file,
and it will stay in your sessions tab the length of the manual extension that was granted. You can view the
time remaining in your extension by clicking ​Save​. You can then safely save and exit without losing the file.
You must manually keep up with the extension time you have been granted. Once the extension time has
expired, the file will ​drop ​from your account and be released to the QA queue. Make sure you track your
time.
All manual extensions override the self-extensions, and if you are still within the time frame the manual
extension was granted, then your file is safe. The self-extension box may not appear on your session while
the manual extension is in place.
Please remember if the file is due to the client and you are requesting time over the due date of the file,
we may not be able to extend it for long periods of time, but we will grant as much time on the file as we
can. After your probation period, we also limit the total amount of time you can hold onto a file, based on
the length of that file. When you've reached the maximum time allotted, then you can continue working
on the file, or you can drop it, but you will no longer be able to get a manual extension.

Edition 2.21.2018 Quick​Step​ Handbook 16


Extension Guidelines
For audio files ​less than 30 minutes ​in length:
● Average processing time: 4 hours or less
● Maximum total extension: ​10 hours
For audio files ​between 30 and 90 minutes ​in length:
● Average processing time: 4-12 hours plus breaks
● Maximum total extension: ​45 hours (​ given up to ​12 hours​ at a time)
For audio files ​over 90 minutes ​in length:
● Average processing time: 8 hours per audio hour, plus breaks
● Maximum total extension: ​40 hours per audio hour ​(given up to 24 hours at a time)
Note​: These limits do not apply to our probation QAs. Experience shows that most probation QAs will have
no problem completing sessions within these times, but I don't want anyone to worry about their speed.
Just put out the best sessions you can.

Session Drop Ratio


Once a file has been on your account for ​more than one hour​, we expect that you will follow that file
through to completion. If you then fail to complete the file, either by canceling or by allowing your
extension to expire, it is considered a ​dropped ​session.
Any QA may have an occasional dropped session. Sometimes something comes up suddenly and
unexpectedly in life, and you must release your file to deal with it. Sometimes you start on a file and then
find that it is too difficult for you to complete and you'd prefer to let it go. This is fine. However, every
hour that you hold onto a file that you are not completing is one hour less that the next QA has to work on
it before that file is due to the client. For this reason, we track the number of dropped sessions you have
each month versus the number of sessions you complete. The percentage of sessions dropped - ​whether
you worked on it or not - ​of those you take on is referred to as your ​drop ratio.
QuickStep QAs are expected to maintain a drop ratio of ​10% or lower each month​. This means no more
than ​1 out of every 10 ​sessions you keep for more than an hour can be dropped each month. QuickStep
QAs must also maintain a drop ratio of ​10% or less overall​. This is due to the QuickStep teams goal of
reducing our average turnaround time for general files and to also surpass our client's expectations where
possible.

Policy for Lost Files


Sometimes a file will be dropped ​accidentally​ by a QA. If this happens to you, please follow the below
procedures immediately:
● Post the file ID​ on Yammer in the ​Lost Files g​ roup letting all QAs know not to work on this file.
● Create a ticket on the Help Desk​, providing the ​file ID ​of the file you have lost. We can't find your file
if we don't have the ID.

If you pick up a file that has already been stamped up to - or beyond - the first five minutes:
● Check the Lost Files ​group to see if anyone has posted the file (using the ID) as lost. You can use
the Yammer search feature to search the ID
● Create a​ ​Help Desk ticket ​and let them know you have found a file that could be lost, .
Edition 2.21.2018 Quick​Step​ Handbook 17
● Notify everyone on Yammer in the ​Lost File ​group and ask if someone has lost the file, listing the
ID number
● If no one has reported the file lost within the hour, you can go ahead and work on the file.
● If it has been timestamped to ​less than five minutes​, then go ahead and work on the file.

Released/Canceled Files
Sometimes a file is not lost, but a QA has started the file and decided they did not have time or
something came up, and they cancel out of the file. This puts the file back into the queue for others to
pick up. Please remember that ​TranscribeMe does not pay for lost or incomplete files​. Pay special
attention to the due times on your files. If you decide to release a file that you have started, please post
it in the ​Lost Files ​group to avoid unnecessary delays for the next QA to pick up the file.

Edition 2.21.2018 Quick​Step​ Handbook 18


Understanding the Review Process
TranscribeMe offers clients a 99% accuracy guarantee. This means that all files completed by a QA
need to be client ready. Your completed sessions will be subject to review. As a new QuickStep QA, not
only are all your files reviewed, you are also under a probation period. Please make sure you read your
probation information carefully.

QAs must reach and consistently maintain a quality ranking of 4.5. Falling below this quality will be
grounds for removal. When reviewing your ranking, we look at: overall average (all files completed), last
month average (files completed last month), and current month average (files completed in current
month). Falling below the quality standards for any of these can be cause for loss of QA status. We need
QAs that can consistently reach and produce high quality files for our clients. It is very important that you
read all of your reviews and look at all the changes made to any of your completed files. This can be seen
under your work history.

If your file goes to a Reviewer, they have the option to ​Decline ​your work. A decline means the
Reviewer has deemed your work unsatisfactory. You will not be remunerated for work that fails to pass
the review stage. If this occurs, you will receive an email advising you that your work has been declined
with reasons given by the Reviewer. In such instances, you will have one hour to claim the file and start
correcting any errors the Reviewer has noted in the file. If you do not claim the file within an hour after
a rejection, the file will go back into the QA pool for another QA to complete the file. So please make
sure that all work you submit is of high quality, and review your work before submitting a file.

Note: if the file is a priority 1 file or is overdue by the time the Reviewer accesses it, then instead of
declining your work, they may choose to correct the errors themselves. If the Reviewer considers your
work unsatisfactory but corrects the errors themselves due to priority or time, this will be noted on the
review feedback form. In these circumstances, an administrator will review the session to validate the
Reviewer's decision. After this, you will be allowed to keep a portion of the pay earned for your work,
and a portion will be transferred to the Reviewer to compensate them for the extra time spent.

Please understand that Reviewers​ are reviewing to see that the file is client ready when you submit it.
Your rankings are based on this.

TranscribeMe does not pay for partially completed files. Please make sure your file is completed before
submitting.

Probation Period
New QuickSteppers will undergo a probation period of no more than ​1.5 audio hours​. This probation
period will be paid at a rate of $35/AH, as all files will undergo a detailed review with our dedicated review
team. Please choose sessions that are around 15-30 minutes in length when possible for your probation
files. If you are having trouble finding sessions of this length, please raise a ticket with the Help Desk and
ask for assistance locating acceptable QuickStep probation files. Also, please be sure to complete your
probation sessions one at a time, carefully reviewing the feedback from each session before selecting the
next. This will maximize your opportunity to learn from each new session and ultimately succeed in your
probation period!

After reaching 1.5 audio hours, you must ​stop working​ and wait for the next step in the process. If you
have work on the Jobs tab or for another team, then you may continue with that while waiting for your
final session to clear review. Once all probation sessions are complete and have passed through review
successfully, you must submit ​this form​ to request an admin review. At that time, if your probation
sessions have averaged at least 4.50 in ranking, then you will become a full member of the team and move
to the normal $40/AH rate.
Edition 2.21.2018 Quick​Step​ Handbook 19
Future files may undergo full reviews on a random basis, or may pass through smaller admin reviews to
receive feedback. If your sessions are found not to meet the required 99% accuracy for QuickStep, then an
additional probation period of 1.5 audio hours at $35/AH may be required to maintain your position on
the team.

Review Form
When a file is reviewed, you will have access to a PDF Review Form that looks like the below. A score of
five indicates that you had ​no errors at all ​for that category. Even one error can lower your ranking for a
category to a four, so please check your work carefully before you submit the session.
1) Please rank the audio quality for this session. ​The Reviewer will rank the audio quality. We
totally understand that poor audio quality can affect the quality of the transcript.
2) Should file have been rejected but due to priority Reviewer redid? ​Most times if your file is of
low quality, the Reviewer will send you examples and a reason why they are rejecting the file. You will
have one hour to pick up the file and start working to make the corrections. There are times when a
Reviewer must go ahead and make the changes because the file is due to the client. If the Reviewer
determines the file should have been rejected, they will tick this button. This means a manual review of
your file will be completed. If it is determined that the file is of poor quality, half of the payment might
be removed to compensate the Reviewer for their time. The Reviewer's job is to make sure that you
have submitted client ready files. It is not their job to QA a file.
3) How many speakers were present in the file? ​Total​ ​number of speakers in the file.
4) Were the timestamps accurate? ​This section includes not only the timestamps but also the
speaker IDs. Timestamps CANNOT be off more than two seconds in either direction.
5) Were all words spelled correctly? ​This also will include the spelling of names that should have
been looked up on Google, and if research has been conducted. This would also include if names were
spelled consistently throughout the file. Did you spell Linda or Lynda different ways? Research needs to
be conducted for those harder to understand words.
6) Are sentences/paragraphs structured properly or do sentences/paragraphs stop and start in
odd places? ​Does the file have a lot of ellipses, double dashes, false starts, or filler words that could have
been left out? If so then the sentences do not flow naturally. Have the sentences been structured
properly? Do the sentences include the full thought, or are they short and choppy? Are the paragraphs
overly long or short and choppy? You need to make sure the speakers' thoughts are not broken. Use
your best judgment here.
7) Has punctuation been used correctly? ​When thinking about punctuation think about all of it,
including the double dashes, single dashes, ellipses, not just commas, periods, quotes, etc. While we
would like commas to be accurate, there are always arguments as to different ways. If it is blatantly
wrong and changes the meaning of a sentence, then, yes, points will be removed for incorrect comma
usage.
8) Are there any Misheard/Missing Words/Letters? ​If the file has misheard words, it has misheard
words and will not have a ranking of five. The Reviewers do not have any special equipment that allows
them to hear the words better than you. Please, listen carefully and always take context into mind.
When thinking of rankings between one and five, think of it this way:
● No misheard words would be a ranking of five, which would be A+ and excellent job.
● One to two misheard words would be a ranking of four, meaning a grade of A- or B+ and very good
job.
● Three to four would be considered something between a B & C grade. You did not do horrible, but the

Edition 2.21.2018 Quick​Step​ Handbook 20


file did contain errors.

9) Are there any Style Guide errors? ​Style guide errors can include a variety of things such as tags,
speaker IDs, contractions. It is also very important to also ensure that the file follows the posted style
guide that is on the right-hand side of the session.
10) Has there been a reasonable attempt to transcribe all words? ​A reasonable attempt to
transcribe all words means that there are not [inaudible] left in that can clearly be made out. If the
Reviewer can make them out, then you as a QA should be able to understand the word. Did you use
proper research on Google for those harder words? Was an attempt made to guess as any of the words
that were hard to hear? Do the words match the context of the file?
11) If this was a two-step file, were all words in their correct audio chunk? ​Applies only to 2-step
QAs.
Finding your Ranking
As your completed sessions go through review, your overall ranking for each session will appear in your
work history. Please click the ​View​ link to view the reviewer's comments and notes on your file.
At the top of your work history, you can also view your average ranking across all reviewed sessions, your
average for the past 30 days, and your average for the past 90 days.

Q: This says I need a 4.5 rating to stay a QA. How do I know my rating?
A:​ Your QA rating is simply the average of all the ratings you've been given by reviewers. As a new QA,
most your files will go through review. (Occasionally one will get fast-tracked to meet a client deadline,
so be sure you always submit client-ready files!) When the review is completed on a file, you will get an
email with the review sheet attached, detailing your rating in each of eight categories as well as
providing comments and feedback to help you improve. You can also look up your most recent rating in
your work history.

Q: I've been a QA for a long time now. Why did my last file go to review?
A: ​There are a number of reasons that a file may go to review. It could have to do with the particular
project needing extra high accuracy, or you might be set to receive a spot-check for any of the
following reasons:
▪ Probation ​- you're on probation. All of your work is being reviewed!
▪ New QA ​- you've just cleared your probation period, and we want you to get some more
feedback on your early work.
▪ Inactivity ​- you've been gone for a while, and we want to make sure you remember the style guide
well!
▪ CSI File ​- the client complained about one of your past sessions. Whether justified or not (and we
do check), this will usually garner you a spot check on your next session, just to be sure everything's
in order.
▪ Declined Session ​- if you have a file sent back by the reviewer, or if the reviewer marks the "should
have been rejected" box, then you will likely continue getting reviews longer than you otherwise
would have.
▪ Accuracy Concerns ​- an administrator noticed something in your work that made us think you may
need some additional feedback.
▪ Low Ranking ​- your overall ranking is below or close to the 4.5 line, or you have a lower overall
ranking than we think you deserve, and we want to help you bump it up!
▪ Reviewer Check ​- you have been doing excellent work, and we are considering you for a
promotion. First, we need to see what the current review team thinks...
▪ Random Review ​- there's no real reason, but you just haven't had any feedback in a while.
Edition 2.21.2018 Quick​Step​ Handbook 21
As you can see, reviews can happen for a good reason, a bad reason, or no reason at all. The best thing
you can do is just keep putting out the highest quality work you can, and take whatever feedback you get
as one more tool to help you continuously improve your craft. If you are consistently pulling high ratings
on your reviews, you will be back to normal very soon.

Reminders
● Timestamp the first speech, even if it's filler words.
● Consistency is key, e.g., name spellings.
● Timestamp every new paragraph, even if it's the same speaker.
● Capitalize the first letter of the session, and end with appropriate punctuation.
● Check for updates on Yammer.
● Never copy and paste from MS Word into the WorkHub.

Support
If you have any issues with a file or have a question, please create a ticket on the Help Desk. You should
already be registered at ​https://helpdesk.transcribeme.com/portal/home

Yammer
Yammer is mandatory for all transcribers and QAs. Make sure that your name matches your WorkHub
name. The ​QuickStep Team ​on Y ​ ammer ​is a treasure trove of important notes posted at the top and on
the right-hand side of the group. Please make sure that you always post any questions you might have in
QuickStep​ ​Team​ ​and ​not​ All Network, as this may confuse transcribers.

Foot Pedal
For those of you who use foot pedals, we have a Google extension that is used on our WorkHub. QAs
who use INFINITY foot pedals can use this on the WorkHub. You can find the extension using this link:
https://chrome.google.com/webstore/detail/transcribeme- workhub/ekchlkccgoacdfccmicachmhgijgdkhi
● The Center is play/pause button and it will auto-rewind by 2 seconds if you toggle the pedal.
● The left is rewind and it will take you back 4-5 seconds if you press and release once.
● The right pedal does the forward function and will take you 4-5 seconds forward.

Edition 2.21.2018 Quick​Step​ Handbook 22


FAQ
Q: I love being a QA. Can I get a bonus for inviting my friends to QA with TranscribeMe?
A:​ Yes, you can! Whenever you invite someone to join TranscribeMe, fill out this form with their
workhub email and/or full name before they are activated as a transcriber. If they become a successful
QA, then you will both be eligible for a $5 bonus after their first 5 AH completed following probation!
https://transcribeme.wufoo.com/forms/zyt3oou03f1tb3/

Q: What happens if I don't get this RUSH session completed on time?


A:​ If you are on the hub and working, the session is not going to be taken away from you. If you need to
take a short break, please ask the Help Desk for an extension of a couple of hours so you won't lose it by
mistake. Just please don't keep these files overnight, as we need to get them to the client ASAP.

Q: How do I handle sessions that have foreign language being translated throughout the file?
A:​ There are two possibilities here, depending on the details of the file. In a file which has a primary
foreign speaker, a primary English speaker, and a translator actively working between them, it might
look like this:
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.

In a file which has mostly or only foreign speakers and there is a translator speaking over them to
provide English for the recording:
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 QA working for TranscribeMe.
Note that in this case, we create a new paragraph for each foreign speaker, even though they all have the
same speaker ID, since it is the same translator speaking English.

Q: If I have 10 seconds or more of silence, but one of the speakers says "um" during that time, is it still
considered silence?
A: ​Yes. For clean verbatim, you can ignore a filler word that comes in the midst of a long stretch of
silence. Mark it [silence] and pick up again with the next utterance. (For full verbatim, of course, you'd
transcribe that filler!)

Speaker IDs and Timestamps


Q: What if a bunch of people all respond with the same words at the same time?
A: ​If you can make out the individual speakers, then please insert the text for each speaker and
timestamp each one separately. In the case of a single group response where you cannot make out the
individual speakers, you can insert the group response once and timestamp it with the S? tag.
Q: What if I have more than nine speakers?
A:​ Speaker IDs can continue as high as needed, but the hotkeys are only available for speakers 1-9. To
insert a speaker ID higher than this, please press Ctrl+0 to get an S? tag and then click on it. You will be
able to type any number you need into the console.
Q: Isn't there an easier way to put in a two-digit speaker number?
A:​ Yes and no. You can't directly insert a speaker ID higher than S9. However, for high-speaker-count
Edition 2.21.2018 Quick​Step​ Handbook 23
files, when you only have a few speakers active at a time, you can use the Rename All button to speed
up your stamping.
● Take two of your earlier speaker numbers that are no longer in common use, and use Rename All to
convert them to a high number. For example, you could rename S1 to S51 and S2 to S52. This "holds"
those speakers while freeing up the Ctrl+1 and Ctrl+2 hotkeys.
● For a new speaker (let's say it's S14), use the S1 tag to mark it. S15 could be marked using S2 at the
same time.
● When those speakers have finished, use Rename All to convert all of your S1 tags to S14 and all of
your S2 tags to S15.
● When the file is complete, use Rename All one more time to put your S51 back to S1 and S52 back to
S2.
● It's a complicated system, but for certain files, it can be well worth it in the time saved on individual
stamps versus the time spent shuffling tags. Just be careful not to accidentally rename two speakers
to the same tag, as you will be unable to separate them again except on a stamp-by-stamp basis.

Tips and Tricks


Q: How do I run a spell check using Preview Mode?
A: ​When you've finished doing your QA work, save, then click the ​Preview ​button at the bottom of the
screen. It will open a new window with the document as the client sees it. In this window, press CTRL + A
to select all the text and then CTRL + C to copy it.
● Open your word processing application (eg: MS Word) and paste the text into a document. Then
select the correct spell checker option for your application (ie: UK or US English) and then run a spell
check. That application will highlight any errors it finds.
● If you are using the free Grammarly extension for Chrome, it will ​not ​automatically check your text
on the Sessions tab. Instead, you will need to paste the text from your Preview window into
Grammarly to check it manually. If you click on the green G for Grammarly, you can select New
Document at the bottom to open the Grammarly Editor and paste your text in.
● Back in your browser, click the Back to ​Session ​button and your normal QA screen will come back.
Then jumping from the MS Word/Grammarly to the WorkHub, make whatever corrections your
spellcheck identifies. Make the corrections on the WorkHub and not in the Word/Grammarly editor!
This will help you pick up any spelling or typo errors. This needs to be done with every file before
submitting it. Never paste from MS Word into the WorkHub.
● Note: Do not run a grammar check because people seldom speak in a way that is always
grammatically correct and you are doing a transcription, not an edited piece of writing.

Q: What if I can't make out the words?


A:​ First, try changing the speed (up AND down) and playing with your ​DFX/Ears for Chrome ​settings
(including OFF). If you still can't make it out, then you can use the [inaudible] tag. Just remember that
any audible portions still must be transcribed, even if the file has poor audio quality. This will help the
client remember the conversation and fill in any parts that we can't hear.

Q: Any other tips or suggestions?


A: ​Stop the audio when you are making an edit. When you're done, go back and listen to that part
again. If there was one error, there are most likely several.
● Don't put commas after conjunctions (and, so, but). Commas go before conjunctions when joining
two independent clauses or in lists, but never after.
No​ → I love you and​,​ I think you're swell.
Yes​ → I love you, and I think you're swell

Edition 2.21.2018 Quick​Step​ Handbook 24


● This also goes for conjunctions at the beginning of sentences.
No​ → I wanted coffee. So​,​ I went to Starbucks.
Yes​ → I wanted coffee. So I went to Starbucks.
● When dealing with a very large number of speakers, you can use Ctrl+1 instead of Ctrl+0 to insert the
timestamp. This way, the 1 will already appear in the number box, and you only need to add the
second digit. Of course, please be very careful with this tip not to let any incorrect tags slip through
as speaker one!

Troubleshooting
Q: How do I stop the text from jumping around? I keep losing my place!
A: ​Try disabling autocorrect, using the checkbox at the top of your session textbox. Also make sure you
are in Non-interactive playing mode.

Q: I messed something up. How do I undo?


A:​ First, try pressing CTRL + Z. That is the universal shortcut for Undo and might help. (CTRL + Y is the
universal shortcut for ​Redo ​just in case you go too far...) Next, think about when you last saved. If it was
recent, then you can use the ​Show History ​button to jump back to the last manual save. Finally, try
opening your session up again in a new tab. This new tab should show you the last automatic save. If it is
recent enough, then you can close the original tab and Save from this new one. If all else fails, you can
always revert the session to its original transcriber text using the ​Regenerate Session Text ​link.
Unfortunately, this will affect the entire file, so it is really a last resort.

We always recommend that you clean the Chrome history and cache before working on a session.
https://support.google.com/accounts/answer/32050?hl=en
Since our clients upload their own files, there will be times when these files might have errors or
incorrect language selections. There are also certain files that you need to notify the Help Desk about.

● Early submission: You have accidently submitted a file before it was completed.
Please raise a ticket with the Help Desk and include the ​Session ID​, found on your work history page. The
administrator will be able to preemptively reject that file and send it back to you.
● Silent file: Your full file is silent.
Please raise a ticket with the Help Desk and include the Session ID, found on your work history page. The
full file will be checked to ensure there was not an error during upload.
● Foreign files: Your full session is not in English.
We at times will receive files that are bilingual. The style guide will usually note this. If your full file is any
language besides English, please raise a ticket with the Help Desk and include the Session ID. Never
transcribe the non-English.
● Audio will not play: Sometimes the audio will not play.
If you are having any issues with the audio not playing, always clean out your browsing history and your
cache. ​https://support.google.com/accounts/answer/32050?hl=en

Connectivity problems and Saving / Submitting Sessions

The Save, Save & Exit, Submit, and Cancel options above are only functional if you are connected to the
Internet. ​If you lose connection and press any of these buttons, you will lose your work!

IMPORTANT​: Please save your session every five [5] ​AUDIO Minutes​, rather than waiting until the end
of your session. By saving your work frequently, we can try to recover what you have done in case you
lose your work. Contact the Help Desk if you experience any technical difficulties - always include your
Edition 2.21.2018 Quick​Step​ Handbook 25
session ID number. We have an auto-save feature, but it is still a good practice to save your work.

Our system checks your connectivity every 60 seconds, and if we find that you're
no longer connected to the Internet, a red pop-up warning message will appear
in the top left corner of your screen. ​If you see this message appear, you no
longer have an internet connection. You have ONE HOUR to contact​ your network provider to restore
your connectivity, or otherwise you will lose the session.

Submitting Sessions
Before submitting a QA session, please take the following steps:

● Write down Session ID number.


● Click ​Save ​(as long as you have a good internet connection)
● Click the ​Preview ​button
● Copy the text from the ​Preview ​to a word document
● Save the Word document locally.
● Close Preview
● Press ​Submit​.

This will help us recover your work and to have visibility of the work that you have completed. Viewing
your work in ​Word ​will also help you view any mistakes, which you can go back and correct. Remember,
never copy & paste from MS Word into the WorkHub.

Error upon submission of Session

If you see an error message upon submission of Session, please take the following steps:
● Make sure all the procedures above were followed
● Take a screenshot of the error message
● Click the ​Send Error Report ​every time
● Note the date and time that this occurred.
● Attach the screenshot and create a ticket on the Help Desk.

Clicked Submit but nothing happens

IMPORTANT: Please do not press ​Submit ​more than ONE time, even if nothing happens.
Wait for just over 60 seconds to see if the red pop-up window appears on the top left corner of the
screen, advising you that your internet is no longer connected.

In almost all cases, this means your internet connection is lost. Please call your ISP to get support and
re-establish your connectivity. ​Never ​press refresh or back if you have not saved the session.

Audio Play doesn't work

When the Audio Player no longer plays or it stops responding to the hot-keys/navigation buttons:
● Wait for at least one minute
● ONLY if the red pop-up DOES NOT appear, then save the session and Reload the page.
● If this still doesn't help, please, clean your Chrome history and cache.
https://support.google.com/chrome/answer/95589?co=GENIE.Platform%3DDesktop&hl=en

Edition 2.21.2018 Quick​Step​ Handbook 26


Helpful Tip: ​Open another browser window and go to any website to check if you are connected to the
internet before you submit or save any session.

TranscribeMe will not reimburse you for Sessions which are lost due to connectivity problems of your
workstation, or sessions which are lost due to not following the procedures outlined in this handbook.
TranscribeMe does not pay for partly completed files.

Thank you for your interest in becoming a member of the QuickStep team! There are real opportunities
to have a rewarding career with our fast-growing company, with positions from Transcriber all the way
up to Project Manager. Best of all, you can enjoy this challenge from the comforts of your own home.

We look forward to seeing you online!


TranscribeMe Operations Team

Edition 2.21.2018 Quick​Step​ Handbook 27

Potrebbero piacerti anche