I'm relatively new to CW work. Several HITS today were focus groups. I returned one tonight after getting fairly far along because I couldn't untangle the Style Guide about this: "If there are already two or more other speakers in your audio, don't worry about telling the audience...apart. Each one will just be Audience Member...with no numbers or gender needed. However, if there is only one main speaker, then be more detailed in specifying the first two group members who speak..." (by adding a number.) So, if there is one male main speaker and if you have 3 or 4 female Audience Members who speak, do you label them as Man 1, Audience Member 1, Audience Member 2, and the rest just as Audience member?? And if that is the case, if Audience Member 1 starts speaking again after the third and fourth, is she again labeled with the number, or does she then become an unnumbered Audience Member. Also would like confirmation that if you have male speaker and one audience member, you would use Man 1 and Audience Member as speaker tokens. As an aside, it's sometimes very challenging to tell different audience members of the same gender apart, especially when they're not close to the mic.
Mike from Castingwords got right back to me! So posting his response here if any other newbies need it. Q: So, if there is one male main speaker and if you have 3 or 4 female Audience Members who speak, do you label them as Man 1, Audience Member 1, Audience Member 2, and the rest just as Audience member?? A: Identify a single main speaker by (in order of preference): A) their name if provided. B) a description that best fits their role, e.g., "Teacher," Lecturer," etc. Only identify their gender if it's necessary within the context of the audio. The all female audience in this case would be "Audience Member 1", "Audience Member 2", etc. Only add the gender if both genders speak. Q: Also would like confirmation that if you have male speaker and one audience member, you would use Man 1 and Audience Member 1 as speaker labels. A: Yes, you can use "Man 1", however a more descriptive label per above is preferred. You could use "Audience Member 1" in this case. Alternatively, you could use "Man 1" and "Man 2" if both are male, or "Man 1" and "Woman 1", if audience member is female.
Group Interviews i thought I knew how to do these. Now I'm not sure and even I am insecure about it. i had 3 or 4 middle aged black women speaking often over each other, etc. i labeled them all by number but wasn't really happy with that either. So I changed them all to Participant. There were also side talking between ladies and restaurant servers I presume. Fingers crossed. I'll post the results if anyone is interested.
oldbones, if you're insecure, imagine how I feel! I asked a question today about another HIT where the Transcribe Now tab told me to follow the Group Format Style guide but the names of the people were given. Disconnect! I emailed CW support who told me, for that particular HIT, I should use speaker labels as if it were an interview. My takeaway from the last couple of days was, if in any doubt, ask before submitting the hit. I have been impressed about getting answers back so quickly from CW support.
Remember, CW = Frustration. Remember, CW = frustration. Maybe a human should actually listen to these multiple speaker transcripts and tell us what labels to use to remove doubt and render consistent HITs will which will be put together by the editor for final edit. It's in everyone's best interest to do so. Putting careful and well meaning transcribers in a no-win position is not good, CW. BTW are the CW people reading these forums? They did read and occassionally respond to the old forum.
Group transcription It may be clear to Mike but it's not clear to me. The labels I was using were a panel of about 4 middle age black women that sounded very much alike. One of the problems is no one tells you how many people are in the audio. Are there three or are there four? These were political women talking about their experience in Washington. not audience members at all. Other transcription services charge a lot extra the more speakers you have. That might be why we get more than our share of these. IF CW were smart they would, too, and they may be, just not passing it on to the workers, since there are almost an unlimited supply of people willing to sweat these things out. on mine i could be relatively certain of a couple of the women, but didn't even know for sure how many were there. They didn't introduce themselves. They mentioned names but not in reply to something so we could label the speakers from the audio. Another one I did, which may have been what you were working on: again, did not know who was on. There was a Dave referred whose name was not in the directions.Then there was an unidentified to and a mysterious man. Dave? Yeah, I labeled it and crossed my fingers. i got a 9 but that's probably just because my work has een consistent for so long. The editor, as you may know, can choose to lower the score when it is in final edit. i'm still now 100% out of danger. Walking on egg shells... This will get CW in such a mess that, mark my words it will go back. They're obviously outsourcing the edits now but nevertheless I don't envy the editors on these. on the ones I improve and grade I'll be giving some slack, as in, lots of slack. .
Multiple Speaker Labels I just saw the final edit on a 20 minute job with 3 speakers. Two of the male speakers sounded very similar. I was given the names of the speakers in the job information, so before doing my second pass, I googled their names and watched about 15 minutes of you tube videos to see if there were any speech idiosyncrasies I could use to help identify one from the other. One of them pronounced words like "little" with a funny rolling r, sound. So I tagged the speakers accordingly. This transcription was Improved, which was interesting. The Improver changed a few of the speaker tags, but not many. However, in the final edit, they basically eliminated all but a few tags for the third speaker altogether. In reading the complete transcript, I did get better idea of the context of who was contributing what to the discussion. But with a 3, or 8, or 20 minute part, I don't know how you could do that, unless the audio is in stereo and there is stereo separation (which happened with one of the focus group jobs I previewed). My grade from the Improved and Final was 8 which I was ok with because this was the longest transcription I'd done so far and there were a couple of typos. I don't know how much the changes in speaker tags affected the grading but I am certainly going to be very cautious about taking on a transcription with multiple similarly sounded speakers again.
You and me, both. i got a really bad grade on an edit the other day. One of the spurious comments was that i messed up a label as in one label in several hundred. I still had it so I listened to it again. No, I didn't. Not to my ear. Also accused me of being sloppy and paraphrasing. Said I left out words. Said i left out punctuation. Said I put the wrong time on an inaudible tag. Big deal, i guess. Who knows. I hate to say it but there have been times when I have had more than a sneaky suspicion that they just wanted their fingers in my paycheck. The more the edit or HIT pays the more likely I am to get an 8. At one point we caught them giving false and fictitious feedback. They blamed it on programming. This was before Mike came on board. That put me on auto 8 for a while and cost me probably $100 a week. Eventually things got back to normal. it's easier to return and find another HIT to do than to try to figure out what other graders/editors and CW staff consider correct or incorrect. And time is money. [Hope CW isn't reading this] As an editor/improver, i haven't been counting off much if at all for this. if it's not accurately transcribed and does not follow style guide that's all I take into account when grading. If you don't mind typing up a comment and sending it and then waiting for an answer, that's fine. I guess I'm too impatient. I want to know how to do it, do it and then be done with it.
If I had the experience you did, oldbones, I would be right there with you. But I am so new with CW (less than 20 transcriptions under my belt), I am still considering my time OJT. What slows me down the most right now is triple checking everything against the style guide. I don't envy you\editors your jobs. And yeah, I expect frustration and differences of opinion by the editors about what I've chosen to leave in and take out. I'm getting better, cause I've been getting mostly 9's lately, but I'm not taking anything for granted. I can see that doing work for CW is always going to be subject to second-guessing. I've gotten more pissed off and frustrated doing speechink's verbatims. They've dinged me for putting commas before dependent clauses or putting in a paragraph break in a different place than the editor thinks it should be.
i can't do Speechink transcriptions. Why bother when there are always CW jobs and they pay better. i think you have the right attitude towards this. I've been triple checking myself. I could go on at length about the style guide. But you don't want o hear, er, I mean, read me ranting on and on. It's their business and they can do what they want to with it. Being half blind, and with two grandkids to support, I'm just glad to have a job at all and doing it at home is such a blessing.