 I wanted to ask you, what in your view is good quality captioning? What are the most common errors and how can they be avoided? Okay, yeah, thank you. That's a great question. You know, I don't know originally that I approached captioning in terms of yes or no. This is right and this is wrong. I mean, there are clearly wrong captions, misspellings or captions that are dropped or missed. You know, just errors in the captioning file. But I've tended to approach captioning in terms of choices that captioners are making. I've been really fascinated by non-speech captioning or sound identifiers, all of those captions in parentheses or brackets, that it seems like captioners tell me anyway that they have a lot of control over and I'm interested in those decisions that captioners make about what to caption and how to caption it. I still think there are some common errors that I want to point out, but I think I've been more interested just in why has it been captioned this way and not this way. And both ways might be right in their own ways, but they might create different meanings or different interpretations for viewers. And that's a little different for me than just this is right and this is wrong. But I think some common errors, things I've been interested in include missing speaker identifiers when somebody's speaking off-screen and it's not clear who's speaking or if two people are speaking at the same time and you've got both of those lines of speech at the bottom center of the screen with just a preceding hyphen, it may not be clear who's speaking which line. So I think speaker identification is something that needs to be closely attended to. I think confusing speaker identifiers or speaker identifiers that might not be exactly appropriate. Maybe I'll give an example in a minute or two, but confusing speaker identifiers. I mentioned bottom center captions, so captions that are not well placed or captioning interfaces that don't support robust screen placement. So captions that are just at the bottom center I think can create some problems for readers and viewers. Boilerplate captions aren't necessarily wrong, but I think at times captioners can rely on, they just go to those captions that are familiar to them when describing non-speech. Like in a horror movie it might be eerie music. Eerie music that just is repeated dozens of times over the course of the movie. These aren't necessarily wrong, but they might show kind of a lack of creativity or diversity in description. Captions that are too fast, of course. I mean, there's been a lot of work on caption timing. And I tracking studies, as you know, that try to figure out what speed is okay and what speed is too fast. These studies go back into the 90s, as you know, but captions that are too fast. Recently I've been just tracking reading speeds over thousands of captions, just doing this in Excel so you can kind of automate reading speed. And maybe 30% of the captions in a two-hour movie might be over 200 words per minute. I don't have exact calculation, but there are quite a few captions that are just way too fast. Missed sonic illusions, I think, are a problem. So maybe they're not just five notes, but maybe they have a kind of deeper meaning. The example I use in my book are the five-note motif in Close Encounters, which is like a 40-year-old movie. But other movies have drawn upon this famous five-note sequence. And captioners really need to understand that these are not just five notes, but they have a kind of deeper meaning. They need to be captioned in terms of the Close Encounters theme or the five-note motif from Close Encounters. So missed sonic illusions, I think, could be a problem. Language identifiers that don't include the transcribed speech. This seems to be a big problem on Twitter, where people are complaining rightfully so about just language identifiers like Speaks in Foreign Language. It's not clear what language that is. Or even Speaks in Spanish, but the Spanish words are not transcribed.