Live Session (2025-26)

Back to Overview
Speaker 1
00:02
Hello everyone, welcome back to uh week eight. Week nine. This is going really quickly. Um I know that there are some inconsistencies in the uh uh Of course, because like we sometimes move freaks around so it fits our schedules better. So this is week nine already, uh which we called audio and video. Um and uh there were a lot of uh of videos for the topic video, how how fitting. Um and yeah, are there any questions, anything I should like explain in more detail? Anything that you you know might want like more information about this is this is a good chance to ask.
Speaker 2
00:48
I have I have a question more of your opinion. Sure. When you have a video, when let's say you have a a lecture, somebody speaking and you are going to offer a text alternative to this. Do you think I should write a a summary of a more factual text or do I write it just as The person is saying it is quite difficult to read when somebody's using speaking language.
Speaker 1
01:22
Yeah, I mean there are always levels of like how you make transcripts. So there are uh transcripts that are very close to what is said that's important for things like You know, if you're in a uh courtroom or something like that, you want to know like intonations and the ums and as and this that person hesitating, stuff like that. For a lecture like we do, like here, that doesn't really matter, right? I'm I mean I'm I'm not lying to you, I hope Um so you know it's like you d you don't know you don't need this additional information. So I think um for stuff that uh that is more casual, uh it's totally fine to uh To um yeah, do a more generalized version, but like you should try to be as close as possible to what has actually been said.
Speaker 2
02:18
Okay. Because it uh the e the the the one fits the people with cognitive problems and the other one is more inclusive for people who might not be able to watch or Yeah, and it might it it might just be like something that is like, you know, audience dependent.
Speaker 1
02:40
Um, you know, I mean There are probably rare occasions where you want to have two types of transcripts. One that is like, you know, f full like the verbal what what was that? Like I just corrected me or like went into another direction in a mid sentence. There are people when they do transcripts cut that together even. So uh so uh Or or like take out duplicate duplications like I did right now. Oh, you you can't you can't l really make yourself listen to your own voices. uh as closely as uh as you should. But like there s some people like cle clean that up, other people don't, and it really depends on your target audience. For example, if you do a transcript that you would use like with the goal of using it in a like newspaper article, then you would probably clean all of that up immediately. But if it's just like for people who uh have the video maybe on in the background and can't read uh and can't hear all of the sounds, um they might want to have it like very detailed and like Like, you know, just what you hear is also what you see on the screen.
Speaker 2
03:52
So there is no clear answer.
Speaker 1
03:56
But no, that as so often, like the context is is important, right? So that's uh that that's it. Oh I personally I would not usually transcribe like filler words or ums and arms and stuff like that, but I would keep it as close to the original because that everyone can use that. Right? If you cut stuff out, then people might be in like, oh, is that important information that I wanted? Is that cut out? Or I heard this like sidebar and it's not in the transcript. What's going on? So I wouldn't I would keep it as close as possible usually.
Speaker 2
04:30
But that's very hard. I I just did this on my video and it was a l um a lecture and she was uh she was talking half sentences and uh she used these uh Swedish uh words like mmm um and so on. But when I put it in a text it was very hard to read. So I used the the captions to to more uh take all these these details he was saying. Right. And then I used a texted alternative like more fact Textual text.
Speaker 1
05:07
Yeah, I think you can do that. Like that that's totally okay. And like not not every like video is really suitable for like direct transcription, like if you if you explain certain like language things and stuff like that, it gets really, really difficult. Having a proper like separate article might be even better than like just having like the words on on a page. But it's it's difficult and it's like there are many different ways to do it Yeah, thanks. No problem. Uh Sean, is that uh the right way to say it? Go ahead.
Speaker 3
05:48
I'm building on Annika's question. So mine is about automatically generated captions. So for example, one of your videos you said double A and then it translated into a boule, which was really funny. But uh how like what is there a success criterion on that and how should you think
Speaker 1
06:12
Yeah, good question. So this is this is one of the uh big pitfalls of like having these automatic. transcription services because they are often not acc accurate. And you know, unfortunately here in this type of course we don't have the um budget. to be really frank and and and direct, uh to like go through every transcript and like correct it. And I'm like I'm I I'm not super happy with that, but I rather have some transcript and something out there. than nothing. In general, the uh caption or transcript is only a valid caption or transcript if it's accurate to what was said. So, you know, it doesn't need to be word for word, but like it can't like say something completely different. I remember when I started using uh Whisper First uh for transcripts, it uh it had this um thing where it just put at the end of every video please like and subscribe my video although this was a lecture you cannot subscribe to this you already have subscribed subscribe to it. But it was like, oh, this is something I I I tr transcribe from a lot of videos, so it must be a core thing that every video does. And I'm like, no, it's not. Like, don't do this. So uh so yeah, it's like uh I I think there's a lot of good stuff with automatic transcription and like the models have improved a lot over the last five years uh for sure. Um and if you look at like YouTube they are still using I think a fairly old model uh or like one that does not do a lot of stuff. I'm I'm not an expert for that for those things. So so it's often like a little bit wrong, more wrong than if you like do it locally with like a specialized like transcription model. So um yeah you can you can improve the quality there, but there's always the need uh of a person to go through and like fix fix some issues. And usually that's that's pretty straightforward. But someone has to do it.
Speaker 3
08:26
Okay, so uh use it, but then look into what it has generated and kind of fix the problems.
Speaker 1
08:33
Exactly. And I can I I actually have something that I wanted to show you anyway about that. So uh Gonna share my screen. Uh da da And there we go So in 2020 I started a YouTube channel like so many people. Don't don't blame me. Um It's not the most the the the most thing, but like you know, how do you do how do you make sure that like when you do transcripts um and um uh and and and uh captions, how do you make sure that they are good? And so one thing that I did is I wrote out my script word for word and then use a teleprompter to uh read every line. Which made it much more difficult to make videos because I'm not a trained, you know, teleprompter person. But oh we're going we're getting there uh eventually. But basically what I did, I wrote all of this out, like all of the text and put it in teleprompter and let that scroll by me. So I already had the text of what I was going to say. Um and then when I uploaded the video to YouTube, I could go to where the subtitles they always change where the stuff is. To the subtitle menu and do edit as text and just paste it in, like the whole text that I did. And then when you click done, it automatically adjusts the timings. to where your text goes. So this way you you already know what you said, right? And it does not need to go and like extract the text from the video. It just takes the text that you have. compares that to the automated transcript and then makes the timings correct. And that works actually pretty well. And you can like if you um If you click here you can see how how much text there is. So you know I did change some of the texts so they don't they are not too long. um and also not too short. So that's that's a thing that you can do. And that's really, really straightforward. I think this is one of the best user interfaces for captions for doing this alignment stuff. But I'm sure that's there's desktop software as well that does that. But yeah, that's that's one way to do it. The other way is you know you upload it to YouTube and if you have like a better transcript you can upload that as an SRT file as well. I think here you go plus caption No. Oh, because I already have one discard changes, that's important. Oh, it it just gives you a new field here. That's not what I wanted. Uh discard. So you can usually under languages you can add subtitle languages, which is the only way to show this. Let's go to German. And then you can add your manual subtitles here and you basically do the same thing. You can upload a file, type them manually, or auto-translate. So that would be a way to do it. And then autosync is what I just did. And that only works for the original video language. So select subtitle file to publish and then with timing this would be out of your Mac Whisper software, you know, whatever you do, and then you can upload that uh really easily and then also edit it and see you know where are the the timings and such so it's I think that's pretty good. Um Yeah, uh let me see if I can give you a view of Mac Whisper that doesn't look like uh I'm uh giving you all the details of all our clients. I think they would appreciate. Um yeah no this is okay. So let me share that screen as well. So this is this is uh Mac Whisper, which is basically like just a UI for uh Whisper, which is a transcription software. Um and this can just transcribe anything that you put. Like I put in some uh some podcasts here that didn't have a transcript script just to to find like a word that uh that uh you know that that I was interested in. Um and you have different uh models here so these are your typical uh language models like your whisper is a very well known one. I don't I have to check that you can see that window. No you cannot Do this perfect. So here you have all your different different models that you can activate. Um, and you know there are Better models there, and you have different like benefits and weaknesses. So for example, what I'm currently using is Parakeet version 3 Which is a very good model, but it's especially it's very fast. So if you use Whisper here at the top. They are also very accurate, but they take like basically as long as your video is. And that is, you know, that can be like just a lot of time that just sitting here and waiting um and doing nothing and like the parakeet version three is super accurate. At least for me. Like this is also something like if you have a strong accent, um, which we probably most of us have because like we're all not native speakers, right? So um uh if you have a strong accent a mo some models might be better than others and you just have to try. Right. So um so yeah, I I'm generally uh pretty well off with these this parakeet one. Um and like the there's not like a lot of differences between them. Um you know if you're if you want to be very uh Save. I actually have not used those yet. Uh, you know, you can try out the like because this is on a Mac, uh, you can try the uh English ones. here from Apple and download those um and activate them. Uh I guess because I have the Engl English already here I can I can do that and then you can like load an audio file in. Let's just take this video that I just had. Go to my channel. Hello. In this hope. Shut up, Eric. Um, and then we can put that into Mac Whisper. there we are and then just click transcribe and I did now take the model the the Apple model And press on transcript and then we can see how long that takes. Now it of course it needs to download the video first or the audio of the video. I don't know how quick that goes. Apparently relatively quick, so that's nice. And you can see that it says WCAC S W I C A C, which like, yeah, that it's it totally sounds like that, but that that's not the the thing that one it also does does 2. 1. So yeah, uh it's it's not the worst, I think, but it's also not like 2. 1 expanded on Ritchie 2. 1 Oh I don't know. Yeah, it does obviously it struggles much more with my accent than other models. Uh but that's why we going to to try this with the parakeet model in a second. So this is this is good. Um Deoling so already did that pretty quick um And then we can do the Parakeet 3. The nice thing about some of these models is they are specifically multilanguage models. So if you speak German or or uh Swedish or whatever. In between it will transcribe that basically on the fly in in the correct language, which is pretty powerful. Um So putting this in again. Now we're using the parakeet model. And this is like this should be much quicker, like once it downloaded the video, which it does again. Um, I think. Should we here, here we go. Do it, do it. Oh of course it uh now takes l longer until something ex uh shows up because uh oh yeah there we go. Done. And uh and that goes relatively quick and it goes from context and so WCAG is correct. Well here it has a with with check, uh also fun. But in widjack are also not bad. Um So it it has a couple of those as well, but uh you you have seen like the the difference in speed of like how quickly it uh it did the actual transcription and it's it's fairly good. Like I would not publish this. I would be very embarrassed to publish this. Um but I think for a first version to then work work along and like say, oh, let's change this, let's change that, it's pretty good. Now you have in uh Mac Whisper you have these segments which uh tell you how long the individual segments are, uh which also likes a thing that I noticed. I'm always unhappy with like this is way too much text to put on one slide, like as as captions. Uh so I would usually like split that up then on YouTube or uh, you know, in any um uh web that I would use. And it all also tends to like if you do something like version 2. 0, it super often likes to like split that up there because it says like, oh there's a dot, it must be a new sentence. And I'm like, yeah, it's not not really like it works. Um so that's a little bit of a like dumb dumb way to do it. Like, oh that's a that's a Full stop, let's let's split it up. Uh but it's very nice. Uh you can you can make your edits, right? You can just type out what you want. Um and fix your transcript like that. You can play the audio on the bottom as well. Most transcription programs support that. Right, so you can just go through uh or read or here to like this section. Reactions with WICHAC 2. 1 And then you know what was that, and then you can make make your fixes pretty straightforward. Um So I like that. And what it this also does, it has speaker identification. Now it says it has found two speakers, but it doesn't show actually two speakers. So that's that's fun. Because it I don't I don't know. Um so you can delete that. But if you have a multi, like a podcast or something like that, speaker identification, these days works relatively okay. Um it's not perfect. Um and uh uh but I also haven't like tested it out with like having speakers with different audio channels So it would like make the transcription basically transcribe one audio channel and then transcribe the other audio channel and then it can make of course the distinction much better between the speakers. Um but yeah that's uh you don't always have that. Um yeah and then you can export it here at the top. And usually for subtitles you would uh use the SRT format, which is the uh sub something uh format. I always forget what the abbreviation is. And you can add speaker names to it if that's something that you need to do or you can only select the favorite favorite segments. You can allow for multiple lines, which I never understand how that works. And then you can just like export it on the bottom right and save it to a file and upload it to YouTube. Um and yeah, and that's I think it's it's very useful to like on the quick to have these transcripts ready, but um for like publication ready um stuff I think you really want to have a human reviewing it. Right. At least like do a spot review. Like common things and you get a feeling for it. Like I knew that this would be wick. w it it would terribly transcribe that. Because it all it it often does. So you so if you have your your um uh your commonly um like mistranscribed things. Uh you can even like put that into find and replace here and then you you can uh replace that during transcription. It will just do it. here um and that's that's fun and you also have the ignored segments which goes back to the previous question Things like silence, no audio, crunching, so if someone is eating during the things, crickets. The pretty fun. And I I also love that they have like copyright TV Gelderland twenty twenty and twenty twenty one in here. Because apparently the training of some of the models put that at the end of a lot of videos. So yeah. Uh like and subscribe this video if you think this is a great uh direction humanity goes in. Uh but yeah you can put in like um and um and stuff like that into here as well if those are like separate um sections or you can just find and replace them out. Yeah, so lot lots of things. Like this this started as a really small um program two or three years ago and I bought it like for five euros and now it's like 25 euros. Um and it has all these like different different features which you know are quite quite useful actually. So um so yeah. Yeah, this is uh has been been a good invest uh investment in there in that case. Um yeah, so that's Mac Whisper. I really like it. Like I think it's probably the best uh For Mac at least, uh do doing stuff like this. Alright, any other questions? No other questions. This was a lot of like caption talk Alright, I want to share uh or show a video because this is this is what teachers do when they don't have a lot of content and people don't ask questions. uh to share a video for f for us that uh that uh Apple has released uh just two weeks ago, so uh December third is always um International Day for People with Disabilities. Uh and uh Apple usually does like a video or like something that is like shows their um technology used in different circumstances and Apple is one of those very accessibility first companies so they always try to like you know feel good about themselves and I I think in these in these cases probably rightly so right um so uh I want to I want to show this video Which is not that one. It's this one. And maybe you have seen it. We will see it in two versions. So it's over two minutes long. The first one is just without uh with captions um and without uh audio description and then there is also a version with audio descriptions that I also gonna show um so yeah uh Have fun with this and I will mute myself and then I will press play
Speaker 4
26:27
Dear student, congratulations, you have been accepted into the bathroom. One, two, a lot, two, three, buy! Room to lecture hall.
Speaker 5
26:43
I'm not remarkable, I'm just finding my way.
Speaker 4
26:47
Turn left.
Speaker 5
26:47
It's not a walk in the park, but baby, I got the grades. Don't wanna be alone. I'm not your inspiration. You wanna be inspired, there's a library down the hall Have you no it's admiration? Sometimes smells a bit like peers I can be strong, I guess so wrong I don't always talk so pretty, I'm not remarkable.
Speaker 6
27:12
Not heroic garbre I've got no super powers I can't even do this stupid equation I'm not remarkable and neither are you We all have toast We all look funny
Speaker 5
27:28
I've got a face, I've got a place, I've got a life, I've got a name. And I need stuff, and you need stuff, and all our stuff is not the same. And some days I get so tired. She gets grumpy, it's bossy. Well I have a special need, I'll call it coffee.
Speaker 7
27:50
Do I have to be remarkable? I'm just having a tray. I'm not fragile or indestructible, and like everyone, I sometimes cry, and I'm not like the rest, but then the rest of you aren't like the rest, either. I'm not a scene, I'm not a scene, I'm not a loser. I'm not a winner.
Speaker 5
28:12
I'm not more and I'm not less. I'm just your average human mess. I'm not remarkable, and neither are you. I'm not remarkable and neither are you We've all got to and I don't We all look gorgeous in and take pictures I've got a face I've got a place I've got a life I've got a name and I need stars And you need stuff, and all our stuff is not the same. And we will strive and we will fail and we'll get hurt, but we'll prevail, and it'll be joyous, and it'll suck. And I'll be lonely and I'll be loved. And on the way, we will discover another. The same, but we're not the other. There's a life out there that I refuse to miss. I'm only remarkable because everybody is
Speaker 1
29:15
So good. I I I really like it. Um and yeah, and it's it's a it's also a nice thing because like um you know you you get these accurate like subtitling and these these like really good like w ways to to do it and that's why I I love to show them because there goes so much like craft into these. So so you know it's it's always nice to see how this all fits together. Yeah and now this is a YouTube functionality that is not open to everyone so I couldn't show you this with my account. uh because uh apparently I'm not as important as Apple, which I don't understand, um, but uh the this is how it goes. So under um audio track you can use English descriptive So we have English original and then English descriptive and that will give you the audio described version and I will put it back to the front And then we're gonna listen to it again and uh see what the changes are for audio description.
Speaker 8
30:19
A blind student holds an iPhone.
Speaker 4
30:21
Dear student, congratulations. You have been accepted into the undergraduate.
Speaker 8
30:24
Students cheer. A student signs in American Sign Language. Subtitles. I'm going to college. I want to college. A student taps accessibility in their iPhone settings. On a college campus, the blind student holds an iPhone and
Speaker 4
30:34
Starting route to lecture hall.
Speaker 5
30:36
I'm not remarkable. I'm just finding my way.
Speaker 4
30:39
Turn left.
Speaker 5
30:40
It's not a walk in the park. But maybe I got the grades. Don't wanna be in love.
Speaker 8
30:46
A student types on back with her toes.
Speaker 5
30:49
If you wanna be inspired, there's a library down the hall. A student uses magnifier on Mac You know is admiration sometimes smells a bit like hair I can be strong, I guess no wrong I don't always talk so pretty I'm not remarkable
Speaker 8
31:14
A student draws an iPad. We all have to She holds Apple Pencil with her toes.
Speaker 5
31:21
I've got a place, I've got a life, I've got a name. And I need stuff, and you need stuff, and all our stuff is not the same. And some days I get so tired. She gets grumpy. Yeah, well I have a special need. I call it.
Speaker 8
31:38
Students hold coffee and bounce in their seats energetically. In class a student uses live captions on iPad. A blind student plays piano.
Speaker 7
31:43
Do I have to be remarkable? I'm just having a try. I'm not fragile or indestructible and like everyone I sometimes cry and I'm not like the rest But then the rest of you aren't like the rest is the student's party I'm not a sing I'm not a single I'm not a I'm not a winner.
Speaker 5
32:05
I'm not more and I'm not less. I'm just your average human.
Speaker 8
32:09
A student uses hearing aid on AirPods Pro. Confetti falls at a house party. Students with and without disabilities dance. They wake in a daze the next. morning. A row of students kicks out their feet.
Speaker 5
32:25
We all look gorgeous in pictures.
Speaker 8
32:31
Outside students dance. In cap and gown, sign the lyrics in ASL. A student with one leg leaps into the air.
Speaker 5
32:38
And we will fail, and we'll get hurt, but we'll prevail and it'll be joyous and it'll suck. And I'll be lonely and I'll be loved. And on the way we will discover we'll The same and all the other. There's a life out there that I refuse to miss.
Speaker 8
32:54
I'm only remarkable because everybody is The Dancing Students Pose They smile at each other, then disperse. Titles. Designed for every student. The Apple logo.
Speaker 1
33:08
Yeah, and that's uh that's how uh audio description works.
Speaker 3
33:13
Yes, Eric. Um so I thought about this when when I looked at your videos as well. I mean, I saw a lot more than was described. So I'm I'm guessing like where where do you know like how how do I know what to describe and what not?
Speaker 1
33:32
That's a good question. And like basically this is all experience based. The important thing is like what do you want to say with the video? And they and then the other thing is like where do you have space to like put audio description in Right. Sometimes they started by like the the first scene where they put the audio description before anything had even happened, you know, and you have this like little beat. Where in the non-audio described words it's just the anticipation, but in the audio-described version it's like you you're actually already getting information. Yeah, I I think this is an art more than like a science for for for a lot of things, especially like for artsy farzi videos like this, right? Um this is also like only two minutes forty, um so not a lot of space. And these Apple videos especially, they have this thing where they put so much visual and like content and stuff in. So you have to think about like okay what is clear from like the soundscape and from what is What what you can hear like for example the beer pong scene, um the the person throws the ball and it taps twice on the table and then falls off. And that in is probably enough information for someone who is blind and can't see the screen to say like, oh, there's like someone is playing like, you know, with a ping pong ball, something like that. Um, so they thought it doesn't need to be like more described than like, oh, this is a party scene, right? So um so for the sm for the shorter videos it's very hard um and uh you have to find the space and put that in there if you have longer things um Then it gets easier, like if you have movies, uh, for example, they usually have quiet scenes in in in between. Uh, so you have more time to uh describe more, but you also have to be like specific and try to not like over describe stuff, right? Because like you have y if if you only audibly see the video You probably can't remember everything that was said. So you really have to be specific in what you describe and not describe everyday stuff. you know, so yeah it's it's it's it's a it's it's really an art form. Um especially with like you know um um yeah, prosaic video or or like these music video types of things and stuff. Um it gets even more like an art form. If you have other things like you know technical instruction videos or something like that, I think it's easier than you just try to describe what's literally in front of you. I don't know, rotate the I don't know. I'm I'm bad with like these things. Like uh but but you know you you want to be like very exact and very detailed in your description in those cases. So yeah.
Speaker 3
36:52
And what about, because I've seen on some streaming services in America, from America, where they use the CC. And if you select that as the subtitles, you have this, I think it's called Asterix, I don't know, the star, and it says like melancholic music playing. What is that? Is that an alternative to hearing what they're saying?
Speaker 1
37:14
Or so so basically when you have your captions you want to include um all music that is important to like understanding what's what's going on. So if you only see the screen and read the titles, you know what's going on. It has started, and this is for for maybe ten years now, that things like melancholy music or like upbeat music, like giving you more of a understanding like where this scene is leading to because like if you have like an uh a scene where there's tense music in the background But you can't, you know, because like I don't know, the shark uh is of of uh uh of Jaws is like not is just outside of the frame, but you can't hear the dun dun dun dun you probably want to know like oh this is you know the the the shark is coming and you you you would say like um you know dangerous music or like music uh t t t tense music I guess. uh is playing so you you know okay this is this is happening. Um so yeah this I I think it's good information. It's more important if you have things like Oh, there's a gunshot outside, which you can't see obviously because you're inside. Um, then it's like it's really, really necessary to to mention it. um, you know, or the doorbell is ringing, something like this. Um but if you're if yeah, but now we have more of these ambient things. Like it was also in the video uh it said I think uh students cheer, you know, which like I think was visually very clear that that was happening, but they also have it as the the uh caption Um because like they could cry or they could like you know do do do other things, it might not be as clear. So uh so yeah, I think uh that has uh Been very common to say like, oh, this is this is the emotional thing that we're doing here right now. Okay, thank you. Yeah. Any other questions?
Speaker 9
39:30
Um go ahead. So uh you have to do uh if I can't do the proper subtitles for some reason, um I have to do a transcribe with uh like timing stuff also or I can't just publish a bunch of text and be like this is what's sent in the video
Speaker 1
39:55
Uh I mean t technically I think you could just do a text and say this is my textual alternative for this video. Um it's probably won't be a good a good experience for anyone. Um Uh it it also depends a little bit on like is the timing important for the content? Like, you know, some in some circumstances you probably would want to say like, oh, this is at this minute. of the video to someone else like if you're if you're working let's say uh in a newspaper uh You know, and uh and you say like, oh, I've uh I've listened to the transcript and at minute twenty-five there is a this really important sentence that we need to like Should we put on like the the the news because now you're you're working in a TV studio? I'm very flexible with that Um it's like it's like, oh y you have that, you know, you you can you can look it up there. So it might be more needed to have timings in there in these situations than when you um when you have other types of videos. So you know that that depends. In general, you know, captions are important because like some people just can't hear what's on the screen. I always encourage people like putting accessibility in in the planning of the video. And I think that was also in the in the videos that I recorded for you. Because like when you have it in the back of your head i in the back of the head that people might not be able to see what you're doing. You're just describing it automatically. You're just doing like more. And then you The more you do in your main thing and in the captions and transcripts, the less you need to do like extra audio description, right? Um and I think that's That's important. And sometimes like this music video wouldn't work without audio description, right? You don't want everything that is in the audio description also in the quote unquote normal version of the video. I don't think that makes a lot of sense. And I think even like for people who are do w who want to have audio description, the audio described version is better than a version that is for all people. And includes the audio description within, because that would be more clumsy and like more difficult, right? So yeah, it's um yeah, it's uh the these things are difficult and like It's uh there there is a reason why creatives are always involved into the in in in these videos. Awesome. Any other questions?
Speaker 10
42:44
Yes, I have a question in something that you say in regards to post, stop, and height. So you mentioned that uh you can do this in parallel with other content and being able to post up and hide it. But then you mentioned that unless that is part of an activity that is essential But then I didn't understand like what is an activity that is essential.
Speaker 1
43:09
Okay, good, good question. So this is one of these wicacisms. that that we can that we can explore. Yeah, no, really good question. So if we look at post stop height, um and I'm just skipping directly to the understanding. So uh The important thing is and and here is the essential part, right? Um the important thing is That it could be a distraction. Now, when you have these essential exceptions in WCHA, they are always four cases where you are like in this case where you want the distraction. So let's say it's a cognitive test like that tests like how good are you concentrating. And the idea is, oh, there is a rotating thing like in the top right, and how much does this distract you? Then that would be an essential use of this success criterion. So then it's allowed. Um so that that that's basically whenever it says it's essential, when you remove it, it would not do the same thing because it wouldn't test you for being distracted, right? So that's the idea be behind the essential uh exceptions. Does that make sense? Uh for moving, blinking, and scrolling. And for auto updating, it's basically uh if you have Oh, I don't have a good example for that. Do they have one? Because I think auto updating as difficult to to justify as an essential thing. But it might it might also, oh yeah, auto-updating would be like, oh you have um a series of numbers just appearing. And you want to and and your um task is to put that number into a text field. So they basically test your reaction. So in that case, like this would al always auto-update, but it would be essential to do it because otherwise you can't test that, you can't make that reaction test. Something like that. Does that help?
Speaker 10
45:25
Got it. Yes. Thank you.
Speaker 1
45:27
Yeah. You ca you cannot say like, oh, we are a news uh agency and we want our videos to to play next to the test. and be unstoppable and uh and we think it's essential that's not what is meant. Um and I only say that because I have heard clients claim that and I was like no no no no that's not that that's not how this goes
Speaker 10
45:50
And in the same topic, like for example, if you have not in the essential, but in moving, uh sorry, posing, hiding. If you for example have a campaign banner, let's say you have Black Friday and then you have this vertical scrolling of sentences. It says something that if you have uh more than five seconds then you need to have kind of a control to stop it, but any of these ones use it. So when you say like it's five seconds in between the whole scrolling. in between each sentence of each sentence waiting for five seconds until the next comes.
Speaker 1
46:32
That would be longer than five seconds. It's not it's not about the individual like length of the frame, it's about the whole animation for the whole thing. Uh so I would even even if it would um it it says if it's longer than five seconds. It lasts more than five seconds. So it would last five seconds and then the next sentence comes in and that's the neck another five seconds. And that means it's t it's lasting already 10 seconds. So you can't wiggle yourself out. You need to have a pause button there. It it doesn't say that there's no nothing that says like, oh, if it stays still for five seconds, it's um uh it's it's uh not um it doesn't apply. It just says the animation needs to needs to last longer than five seconds. So you can do five seconds, you can you can scroll for five seconds and then stop. That's okay. If it stays stopped. But if as as soon as you start making an animation, all the time in between also counts. as the uh as the animation. I could I could make an exception if it's like right on the top of the browser and and users usually scroll down. Um maybe then I would not find this like too offensive, but um apart from that, I think it's it's difficult to argue.
Speaker 10
47:57
So you say like if it everything lasts five seconds or less, stops and restarts again, that's okay. If people are able to read it, read everything
Speaker 1
48:10
No, if it starts from the beginning then it ta it it's like the uh it this does not say like your your animation needs to last more than five seconds. It says there is moving, splinking, or scrolling information. that lasts more than five seconds. That can be the same animation or the same graphic or whatever in in repeat for for more than five seconds. So it's not it doesn't depend on like how long your file is or how long your animation is or like if it repeats that all does not matter. The result what's on the screen matters. Like if that's longer than five seconds, um then it has to have a pause button
Speaker 10
48:54
Okay, it's funny because I was looking a lot of you know in Black Friday, a lot of these banners. Many people have like is this Actually a requirement or is uh everyone just is not aware of that?
Speaker 1
49:13
It's actually a requirement and I think uh people are like playing with fire, especially with like the European Accessibility Act now breathing in their neck. So I think I wouldn't do it if I was them. But like, yeah, and we and we have that discussion with some clients. They say like, oh, but we have 20 like animations on our page. So have we so we put a pause button on every single animation? And I'm like, this is not useful, nobody will do that. Right? You need to provide one pause button at the top. But that's that's uh only a recommendation, that's not a must. But it's like uh The things that are happening on the web, it's like it's really bad. So, you know, it's like uh yeah, no only because a lot of people are doing it doesn't mean it's uh it's valid, unfortunately.
Speaker 10
50:08
Of course. Thank you.
Speaker 1
50:10
No problem. Yeah. And this is this is one that is also for non-interference. So if you fail this. uh you also your whole um page fails for all WCAG success criteria technically um because uh This leads often to like completely unusable um things uh because people are distracted. So uh yeah. It's it's I I think it's a pretty big deal and I would I would like it to be like much more punished and like people not doing it, but like a lot of people like their animations on their websites and I don't understand why. I'm very old school like that.
Speaker 3
51:00
I have a question. So it's not about audio, but kind of. If you if there are non-A criterions, do you just go after the single A's? Is that what it means? And AAA is only if you're going to go uh above and beyond, right?
Speaker 1
51:24
Yes, that's in general how how it's handled. So so to be um A compliant, you need to do all the A criteria. If you want to be double A compliant, which is what like most of the laws say you must be, you have to do A and A. So it's basically they they stack. You cannot do just double A uh criteria if you want to to have uh compliance with WCAC. Okay, thank you Yeah, it's a it's a little bit of a weird thing and I I have argued a long time that they should get rid of the triple A criteria or should them integrate into the AA criteria because like They are pe people sometimes look at things uh in triple A like uh what what is a good example? Um Here like abbreviations like triple aces, there must be a mechanism that every abbreviation is like you you can know what it is. So when you use things like triple A, you basically need to have like a little thing on top and say like oh this is triple A means like the highest level of of things And like it gets really, really annoying really quickly. And like we're also using a lot of abbreviations, like very casually, like you know, I mean I use Wicca casually, but like you might use UN for United Nations casually, uh you know, or like USA, technically you would need to make a expanded form available with AAA. And I think that can be distracting for a lot of people who would say like, oh, this is like this is too much or like why why do we need to do this? This has no benefit, you know. So it's like Yeah, um the simple best way is to have the abbreviation at the start, explain it once, and then just use it that like meets most people's needs Um and then you use this only when you want to go above and beyond. But that's why we have these like three levels and nobody is looking at to play because like people think it's too much work And I think that's okay, like you know, if you do double A really well, I think you have more benefits than like shooting for individual triple A criteria Because you need and and not having double A, of course. If if you have double A, do what you want. But like having that foundation in is really important. All right, other questions? Nope, okay. I have one other thing that I wanted to show because we talked about subtitles already a little bit. And there are these subtitle guidelines from the BBC. These links will all be in the uh in the uh canvas um as well And I think the most important thing to say is like subtitling guidelines generally differ from country to country. And even inside of a language they can be different in different uh countries. So the Germans do subtitling a little bit different than the people in Austria and the people in Switzerland. And so the there are a lot of like regionalisms in there as well. But I like to show this because it is very detailed and it like uh covers also a couple of things you wouldn't think of like if you're just working for web uh and this might be interesting for people who do like cross media subtitling because uh you know on the web we can do almost anything with subtitles like we have a lot of control we can color them we can like you know change their appearance we can do stuff but if you have it on TV or if you have it on mobile or something like that it gets like a lot you know more difficult if you have it in an app or whatever Um so uh so the BBC tries to cover for all of this. Um and I think it's it's just good to know. Um that this resource exists. Like I will not force you to like learn this or know this or anything like that. That's too may way too much detail and even I don't I don't know. Um, but like for me like one of the big things that we do in these um presentations and like what we give you is that you know where stuff is and you you know how to get to it and how to understand it. Um so let's go to the presentation Good subtitling is an art that requires negotiating conflicting requirements. And that you know, that's true for like captions and uh They use subtitles for captions. It's the oh like the the thing that we in Europe use more subtitles for subtitles. Um And in the US they use captions for when the language of the video is the same as the language of the captions, but subtitles they use only if it's a different language We use subtitles for everything and we then say subtitles for people with uh who are hard of hearing or something like that, um, which is also a little bit weird, so you know it's like wordings. Why do we have that? So and they say prefer verbatim transcriptions, uh like uh but so or two, you should leave that in. um because they are often essential for expressing meaning. So uh so so yeah you want to to be as verbatim as possible uh in those caption uh things. But if you have a very busy scene full of action and disconnect conversations, it might be confusing if you subtitle fragments of speech here and there, and then it gets like all mumbo mumbo jumbo. So Yeah, this is as I said, it's more an art than in a science. You don't want to simplify, you want to retain the speaker's first and last words. So if uh if you have to contract or like change the language a little bit, you at least want to keep those for libraries and stuff like that So there's a lot of of stuff like that. Preserve the style, consider the previous subtitle, keep the form of the verb. It's like super, super good stuff. Strong language is a fun one because like a lot of people struggle with that. Uh and if you upload content to YouTube, you will hear strong language, you so swearing and stuff like that. You will hear it, but the YouTube captions will automatically like make a space uh thing in there, which this is not an equal way to experience the same content if you know if a creator wants to swear uh and YouTube allows content on their platform, that's where they should like also pro uh you know preserve that for people who cannot hear but can only read the the captions. I think that's That's very important because you're basically saying, oh, deaf people are not allowed to swear, which is like uh if you ever heard like deaf people or seen deaf people swearing Yeah, that's it's a real thing. Uh so you know, it's like it's like that's not a thing. Um yeah, yeah, and you can of course use like bleep or something like that. Um or you can use uh uh muted words. Um so there are a couple of things that you can do, but in general like You know, only if you if you also do the pleeping in the audio, then you also want to like of course pleep the subtitle. Yeah, and then there are a lot of things like with line breaks, uh so you get like these fun things like on broadcast you have 37 characters. uh per per line. And online they recommend 68% of the width of the landscape, 16 to 9 video, 90% of 4 to 3 video, 90% of a 1 to 1 video, and 90% of a vertical video. So that you have enough space and also the line length helps. And so there's uh a lot of of considerations about that. You usually don't need to really think about that. I don't. But it's like, you know, when I showed the uh YouTube channel stuff Um yeah, you want to make sure that like everything aligns really nicely in the center. Single sentences, number of lines So they say three lines is the maximum for for the UK and stuff like that. Break at natural points. So yeah, it's like it's really like a lot of things that you would think is common knowledge or like not not that like you you wouldn't need to like write that out um but in reality you probably should um So uh yeah, consider the image. So if you have like uh stuff on the bottom, you you cannot well you can do that online, but most people don't. Like if you have text at the bottom, usually can do the captions at the top and on TV or like movies or something like that you often see that. Uh but I don't think we see that quite enough. um in the um in practice in in most videos because it's actually tough to do like you know you don't need to know like how to put that into your uh caption file and stuff like that. So there's stuff to do. Here how you split up um long sentences, the timing, stuff like that, synchronization, matching shots, identifying speakers, colours, typography. Super fun things like you can dive into. And I don't have the Uh uh I I don't I don't want to, but like it's it's really fun and interesting uh to do that. Like how do you how do you uh show sarcasm? How do you stress stuff If you especially if you can't use like color, you can use like uppercase. Should not use italics ever How do you do questions? You know, using uh question exclamation mark if it's like a question that is like, you know, you want to highlight. Accents, um incorrect grammar, use label, difficult speech, so if you can't like uh really understand it um like slurred speech and stuff like that so there are all all these things that you can do to just say like oh this is how how this this works like uh Yeah, and I I really think this is uh is a good resource like if you really want to dive deep into subtitling and writing captions. this subtitle guide is uh uh is is great. Um and I I really like it. Like it also shows like when you're using um uh these uh hash symbols or uh yeah hash symbols this is also song related like we saw in the YouTube version they had the little notes uh at the front and the back to indicate singing. You can also use the hashes, so that's just convention. Yeah, and that's you know it's it's very long, but it's a really good thing. So I like uh I like this, and then it goes on, talks about subtitles and file formats, and then as an appendix like all the characters you can use and stuff like that and some sample files if you want to try that out. So it's a really good resource. I like it because it's so comprehensive And it was just recently updated as well, so they keep on like putting really good information in here. Um and I I really like that So what else do I have here? The subtitle guidelines. I showed the transcript. Uh yeah, I th I I think uh I did talk about the um audio and media uh making audio and uh video media accessible resource. This has all the infos that I presented in text and with examples. So I think it's it's really good. But if you have any questions about those, you know, feel free to reach out. And the most questions I always get is about audio description or describe video. And uh this also outlines like what the best ways to do stuff. So you have integrated or alternative video or you just try to put it into the in-between. time uh between the uh b between the actual like main audio content Alright, any other questions so far? And if not, I have one example that I want to show that I think is pretty funny, and then uh uh I would close it out.
Speaker 2
01:05:53
Well I have a tips. I found some guidelines for Swedish captions. Oh good. I will put it in the chat.
Speaker 1
01:06:03
Awesome. I will save those uh and uh We'll share them to you and for the next group that does this. This is this is the behind the scenes, you know. We did this this is the fifth time I think that we're doing this. And I'm just collecting from from all the people that are bringing stuff in. No, it's really really good. Thank you very much.
Speaker 9
01:06:32
I have one more question about bleak words. Because like one of the funniest scenes ever made on television is in the series. Arrest the development when Buster is talking and they have to bleep everything because it's so bad, so you just have to imagine what he's saying.
Speaker 1
01:06:52
Right.
Speaker 9
01:06:52
In the subtitle Um yeah, how would you translate that?
Speaker 1
01:06:59
Well if it's if it's uh beeped in the audio you would also you would would either put beep Yeah. Uh in i just in the uh in in the subtitles. So it could be beep, beep, beep, beep, beep, beep, beep, beep, beep, beep, beep, beep, beep, be You know, so that's Totally totally okay. And you could use uppercase and lowercase to like emphasize stuff. Um but I would because it's an American series, right? I would Try to get onto a streaming service and see how they did it because I Yeah I just saw the episode the television so I can see. But that th that's it that's what you're supposed to be doing. Like you know, everything that you the captions are uh mirroring what you would be hearing otherwise. So if you hear a beep, you would put beep in it or you would say like You know, in these cases you could say like um he gets increasingly angry and swears. That would also be like a way to convey it, you know, uh and swears a lot But I think it would be more fun if it if they put beep beep beep beep beep into the subtitles. And this this is like where it's also get like fun to do these types of things where you go like oh how can I make this visually interesting and this is actually where I do have my last um uh example I want to show so uh content warning. There are some mouth noises in here. Um, and also these people have tried hot sauce. And uh Tom Scott, who is on the right. Um who uh is a British uh YouTube person, uh or was a British YouTube person. Um he has just taken like a big scoop of hot sauce and uh he will soon feel what that hot sauce does to him. If you don't want to see this, you can just switch off right now, which is totally fine. But it has a really fun subtitle treatment and that's why I want to show it Um all right, just making sure you can see it too, yes. And uh I will play it now. Um Because a lot of people don't know that stuff like this is possible with captions even on YouTube. And actually I don't I didn't know as well. Okay, let's play this.
Speaker 11
01:09:22
Oh. Seems pleasant maybe? I see why he said it's just wet. There's nothing. Oh, seriously? I was almost disappointed. Should I go and have a walk by the paint? That takes ten seconds and then That was interesting. Your reaction at the beginning was
Speaker 1
01:09:58
So I've if I I I I mean I don't like it when p other people have pain. But I but I like eating hot sauce, so um So the the this was fun. Um and uh and and I think the idea that they were like doing the oh As these this um this clear like subtitle thing that basically lets you feel even if you don't hear like what he's going through, you know, you have this basically explosion in the terms of like, oh Like I think I think it's pretty good. Like this this feels like oh someone um really put their heart into it and trying to caption this video. So uh that's that's why I like to show it. And uh yeah, and and I think that's really well done. Also like the little uh thing that they did in this um uh series is that they always put the guest on y like yellow text and the and Tom on on white text. So you always know what's going on and who is speaking. Uh because in this case they're just sitting on the table, but there have been other videos where like it's a little bit more dynamic. So it's it's it's really nice to have that uh kind of um Yeah, uh c color difference uh to see it. Uh yeah, but I I really like that and uh and I think uh we we not doing that too often. And you wouldn't want to do it like all the time for every sentence. You know, because then it like it's n it's g it is harder to read and it only works because it's this like He is doing that sound and that sound then gets verbalized and like centered. Um Oh love it. Alright, there's another question. Go ahead.
Speaker 3
01:12:02
Yeah, so let's say we have a video that's not for entertainment purposes. So it's more like um instruction video. And As we've heard today, it's up to you as the creator to make it as accessible as possible. But let's say it's something important, as I said, it's an instruction video. Is that something you could test or how how would you know that you've conveyed the message in the correct way. Because it would take a lot of resources to kind of reshoot the video. I mean
Speaker 1
01:12:43
Yeah, I mean you can always have like a textual alternative to the video and say basically, oh, the video is another way to experience this like description. So that's usually how I would do that. Um if you haven't planned from the beginning, then it gets really hard to make stuff super accessible. Um and then you always have to negotiate like um w what is really needed and what you know what is essential to convey and what is like stuff that we maybe not need to convey. And like if you have instructional videos Um you just need to make sure that everything that you show and you say is also available. Like either like if you say it, it needs to be in the this in the in the caption. If you show it, it needs to be somehow audio described that can be like the person who does it or it needs to be in the video. Um in some rare cases Um for audio descriptions you can do what we call extended audio descriptions. So you pause the video And you just put audio over the still image of the video and you say like, okay, this is happening right now. Um and describe what what's in the scene uh instead of like just l letting the video move on. That might be necessary in some circumstances Yeah, but yeah, it's it it's difficult if you don't think about it from the start for sure.
Speaker 3
01:14:15
Thank you.
Speaker 1
01:14:18
All right, any final questions? This is your chance And if not, um oh yeah, I wanted to say all the I think three sessions like week ten For sure, week 11 as well, and I think week 12 too. Yes. Um no, it doesn't actually show this. I think we did week 12. Let me quickly check here. Oh, you can't see this. Huh. Would be easier if you can like see it if you want to. Um So because because now we have a very long Christmas break, right? Like three weeks. Um the idea is that um We give you uh all that content so you can do that in your own time because some of you will have time before Christmas to do a lot of stuff. Some people will have time in the new year to do a lot of stuff. And like, why should we stagger this when you can organize yourself, right? You're all grown-up people. Uh so uh yeah for ten, eleven and twelve uh I published everything. Uh and even thirteen? Maybe even fur 14? I don't know if uh if uh Humphus did did publish his uh as well, but it looks like it. So you have all the the stuff for like in the new year. Uh we have a break now and meet back in um the second week of January, so that's January 13. Um nothing scary about that. Uh and uh Uh yeah, and you know, so so we have basically a month until we meet again uh so you can uh like really get into these um and then the next Course that we do. The next regular thing is forms and interaction. You can only do that if you want to And then come back and do your weekly regular thing. Or if you have like time over uh the holidays, you can just like you know do one or two or three of the courses and uh uh and watch those video if you fancy it. Uh and if you don't, you know, there's we we don't judge. We don't care. No, we do care. But like you you you organize yourself uh with that uh because that's I think the easiest. uh to do it and with that uh I think I want to wish you all uh a good uh holiday season and like you know uh I hope uh some calmness and like relaxation is happening. Uh I wish that for myself and I wish that for you. And I hope to see everyone refreshed in the new year. Thank you very much. Thank you.
Speaker 3
01:17:25
Thank you all. Take care. Bye-bye. Thank you. Thank you. Bye.