An original fictional interview performed by synthetic voices. Use this text to check the exercise’s names and numbers.
WEEKLY EDIT NOTES — SOURCE SCRIPT Practice file 01 | Original fictional dialogue | Synthetic voices This is the unedited source. H is the host and G is the guest. Cue IDs are reference markers, not spoken words. P01 is a one-second silence. The recording is an exercise, not an actual interview. Pronunciation: Qadir = kah-DEER; Maya Chen = MY-uh CHEN. Source note: G06's opening may sound like “I listened” rather than the intended “I listen.” Accept either in this exercise; do not count this ambiguity as an editor error. [H01 / H] Today we're talking about a small weekly interview show and the work that happens after recording. Maya Chen, where does your editing process start? [G01 / G] It starts before I open an editor. I put the guest's name, the recording date, and a plain English episode title in one folder. For Episode 17, that folder is called East Pier with Qadir, because I can recognize it six weeks later without opening every file. I keep the original recording there, then make a working copy. That gives me a safe place to try cuts without wondering which version is the source. [G02 / G] Um, I also make a little list, uh, before I listen. [G03 / G] I write three things in a short note: the promised topic, any name or number I need to verify, and one question I expect a listener to have. The note keeps me from polishing a tangent just because it sounds lively. After the first listen, I add only the moments that need a decision, such as a restart or a long search for a word. I do not mark every breath as a mistake. [H02 / H] What happens when a speaker starts an answer twice? [G04 / G] I take the first one and, um, I just sort of start by taking the first one, because I think that's the one that... [G05 / G] Let me restart that. [G06 / G] I listen to both takes before cutting. In one recent test, the first take had a good example but no clear point. The second said the point first and still kept the example. I kept the second take, then listened across the join with my eyes off the transcript. A sentence can look clean on a screen and still sound clipped. If the join steals the end of a word or makes the speaker jump in volume, I move the cut or leave a little more room. [P01 — intentional one-second pause] [G07 / G] The pause after a useful point is part of the rhythm. I leave that one in. I remove a restart when it repeats information, but I do not run a blanket command that takes out every quiet space. A weekly show needs a repeatable process, not a perfectly flat voice. [H03 / H] How do you check the finished episode before it goes out? [G08 / G] I use a short checklist and listen to the exported file, not just the editing timeline. First, I check the opening, every cut I made, and the last thirty seconds. Then I check captions against a small list of likely errors. This episode's list has Maya Chen, Qadir, East Pier, Episode 17, 3.5 seconds, and $12. A transcript can turn a proper name into a common word or drop the decimal point, and a clean waveform will not tell me that. [G09 / G] Um, so then I save it, I mean, I, uh... [G10 / G] I keep one export labeled review and one labeled final. The review file is for checking; the final label comes only after I fix any audible seam or caption error I found. I record how long the edit and the corrections took. If a new tool saves ten minutes on the first pass but adds fifteen minutes of repair, that is useful to know before I pay for it. The same checklist makes next week's episode easier to compare. [H04 / H] That's a practical place to stop. Thanks, Maya. [G11 / G] Thanks.