FD Stamper Guide

Download as PDF

We're so glad to have you! You've mastered timestamping and, harder yet, identifying speakers! The base rate for Stamping is $7/AH; however, some projects may pay less or more. Barring specific circumstances (see Stamper Guidelines below), Stampers DO NOT edit the text. You are not QAing the file; you are just stamping.

Stamper sessions appear on the Sessions tab, so if you're on another QA team, you'll see these files in your List of Sessions with everything else. Since Stamper files come in a variety of styles, this makes it especially important to carefully read the QA Guidelines to the right of each file.

Stamper Guidelines

Q: Do I begin the file with a capital and end with punctuation?

A: Yes! Please make sure that the file starts with a capital letter and ends with punctuation (unless the session ends in a [silence] tag) like a normal session.

Q: So we aren't even cleaning up the transitions between chunks?

A: Nope. Except for certain special guidelines, you are JUST stamping! Always expand the QA Guidelines for each file to adhere to the proper style.

Q: Do we need to break up monologues into paragraphs?

A: Nope. If the transcriber broke them up, then you DO want to stamp each paragraph, but if they left it together, then it stays together. The idea is that Stampers do not interact with the text.

Q: What if the transcriber didn't create a new line for a change of speaker?

A: Put the stamp in the middle of the TR's paragraph, right where the speaker changed. The hub will automatically kick that stamp down to a new line in the final output (you can also see this in Preview Mode). The same goes for speaker changes at chunk boundaries - you don't have to move the new speaker down to the next line to stamp it.

Q: Can I use S? if I'm not sure who the speaker is?

A: Not if you can avoid it. For First Draft Stamper files in particular, we ask that you make every effort to identify all speakers and use the S? ID as sparingly as possible. Be sure to take notes on speaker characteristics and use the purple speaker box to compare voices to ensure the highest accuracy in your speaker identification.

To Edit or Not to Edit… (Usually, DON'T Edit)

Q: Okay. But this file is riddled with errors. I should fix that, right?

A: No! The files you're timestamping are First Draft files. Clients have paid for this service knowing that it will be a rough draft, not reviewed by a QA. Tempting as it is to fix one or two little errors here and there, if First Draft files become polished QAed products, there is no incentive for our clients to upgrade to a QAed service. And upgrades are good for all of us!

Q: But... All. The. Errors! What do I do?

A: Difficult though it may be, please fight the urge to correct the text unless the errors fall into one of the following categories. If the text does NOT fall into any of these categories, just smile and stamp. You are not QAing this, and we will conduct reviews for TR feedback separately.

Missing Text / Incomplete Chunk

If you come across an incomplete chunk - e.g., it's missing at least 30% or so of text, or it only contains a single tag but you can make out what's being said - then please:

1. Reject the chunk,
2. Transcribe the missing text, and
3. Fill out the Chunk Reporting Form so you can be compensated for what you have to transcribe.

Egregious Errors

If a chunk contains numerous typos (say, five or more separate words) or the majority of the chunk is sheer gibberish, i.e., the chunk is just really bad, please do:

1. Correct those errors,
2. Reject the chunk, and
3. Fill out the Chunk Reporting Form.

Invalid or Misspelled Tags

If the transcriber has included an invalid or misspelled tag, then please reject the chunk and correct or remove the tag. You should NOT fill out the Chunk Reporting Form for these types of rejections.

Q: But why NOT edit? What harm would it do?

A: Well, funny you should ask. The following is an actual conversation Michelle Moll, our Support Manager, had with a client after they downgraded from Standard to FD Plus due to their files being overdue (CS stands for Customer Support, Michelle in this case):

Types of Stamper Files

Most Stamper files are of the following varieties. Remember, check the QA Guidelines before beginning. An 🆔 emoji next to the name indicates that you will use the Speaker Names window to adjust the speaker IDs. See the Sessions Handbook: Speaker Names - Special Styles for details.

• First Draft - The Stamper provides full speaker IDs and timestamps with no unnecessary text edits.
• FD Punctuation Edits - Unlike other Stamper files, these files DO require light text edits to polish up the transcript. Always read the QA Guidelines to know the requirements for each file.
• First Draft RUSH - The client wants a faster turnaround time, so they are high-priority files. Because of this, the QA rate is higher. Please expedite.
• First Draft 8+ Speakers - The client has indicated that you will encounter 8 or more speakers. Because of this, they pay at a higher rate.
• 🆔 First Draft Moderator/Participant - The Stamper must identify all speakers as either Moderator or Participant, along with accurate timestamps and no unnecessary text edits. You may also see a numbered variant, identified with # or NUMBERED in the style name. For numbered Moderator/Participant files, you will need to number the participants, but only number the moderator if there is more than one. For example, Moderator, Participant 1, Participant 2, Participant 3, etc.
• 🆔 First Draft Male/Female - The Stamper must identify all of the speakers as either Male or Female, provide accurate timestamps, and no unnecessary text edits. Like with Moderator/Participant, there may also be a Male/Female # (numbered) where each Male and Female is numbered based on the order they speak in the file. For example, Male 1, Male 2, Female 1, Male 3, etc.
• 🆔 Schools Project - The client is only concerned with identifying the speakers as Student or Teacher. Time is not critical, so the timestamps do not have to be accurate.