WEBVTT - Talking Tech 2nd July 2024 0:00:21.198 --> 0:00:25.398 Hello everyone. Welcome to Talking Tech. This edition available from 0:00:25.398 --> 0:00:29.238 July the 2nd, 2024. I'm Stephen Jolly. Great to have 0:00:29.238 --> 0:00:32.868 you with us listening maybe through Vision Australia radio associated 0:00:32.868 --> 0:00:36.828 stations of RPA Australia or perhaps the Community Radio Network. 0:00:36.828 --> 0:00:39.558 There is also the podcast. To catch that, all you 0:00:39.558 --> 0:00:42.138 need to do is search for the two words talking 0:00:42.168 --> 0:00:44.598 tech and down. It can all come usually on a 0:00:44.598 --> 0:00:48.338 Tuesday afternoon just after it's been produced. Another option is 0:00:48.338 --> 0:00:51.098 to ask your Siri device or smart speaker to play 0:00:51.098 --> 0:00:56.228 Vision Australia Radio talking tech podcast Vision Australia Radio talking 0:00:56.258 --> 0:00:59.348 tech podcast with me, someone who can explain all this 0:00:59.348 --> 0:01:04.328 tech stuff really well. Vision Australia's national advisor on access technology, 0:01:04.328 --> 0:01:08.348 David Woodbridge. David, let's start with our Product Minute, a 0:01:08.348 --> 0:01:11.588 product from the vision store of Vision Australia. 0:01:11.608 --> 0:01:16.048 Indeed. So this is called the Archon Desk stand and 0:01:16.048 --> 0:01:22.858 Archons actually Archon. It's an extremely adjustable stand for a smartphone. 0:01:22.858 --> 0:01:25.858 So you've got the base which is made out of metal. 0:01:26.068 --> 0:01:29.398 Then you've got this telescopic pole that actually extends out, 0:01:29.398 --> 0:01:31.198 I think, up to a meter if you pull it 0:01:31.198 --> 0:01:36.168 right out. Um, normally it's about 30cm. And then the 0:01:36.168 --> 0:01:40.188 arm that comes out which holds the smartphone, it's got 0:01:40.188 --> 0:01:43.308 these little articulated little arms that you can adjust in 0:01:43.308 --> 0:01:46.668 all different directions, because each little arm of which is 0:01:46.668 --> 0:01:48.888 three has got little ball joints. So you can sort 0:01:48.888 --> 0:01:52.038 of rotate them around. So in effect you can have 0:01:52.038 --> 0:01:57.468 your smartphone in any position you like, whether it's portrait, landscape, 0:01:57.468 --> 0:02:01.518 upside down, whatever you like. So it's a really good 0:02:01.518 --> 0:02:04.218 stand for when you want to do things like video 0:02:04.218 --> 0:02:09.768 conferencing or optical character recognition and so on. My wife 0:02:09.768 --> 0:02:12.468 uses her one and her craft room on Facebook and 0:02:12.468 --> 0:02:15.228 so on. She uses it so that on her main 0:02:15.228 --> 0:02:18.318 laptop screen, on her Mac, she can see what she's 0:02:18.318 --> 0:02:21.768 doing on YouTube and Facebook, and then she just uses 0:02:21.768 --> 0:02:24.348 the camera on her smartphone to point down at what 0:02:24.348 --> 0:02:27.878 she's doing for crafting. So it does have multiple uses. 0:02:27.878 --> 0:02:29.858 It's a well known standard. It's been around for a 0:02:29.858 --> 0:02:33.908 long time, so if you're after a really good desktop, 0:02:33.908 --> 0:02:35.918 as in you put it on your table stand for 0:02:35.918 --> 0:02:39.188 a smartphone, it comes in really handy. So again, it's 0:02:39.188 --> 0:02:44.018 Arkon Arkon desktop smartphone stand. 0:02:44.018 --> 0:02:45.188 How much does it cost? 0:02:45.368 --> 0:02:48.878 It's about $225. There is an attachment which we don't 0:02:48.878 --> 0:02:51.808 sell for an iPad, but you can always get them 0:02:51.808 --> 0:02:53.878 from a, you know, overseas or Amazon. 0:02:53.908 --> 0:02:58.948 The icon desktop stand from the vision store at Vision Australia. 0:02:59.938 --> 0:03:04.078 There was chatter last week about X. We used to 0:03:04.078 --> 0:03:08.458 call it Twitter. Its accessibility disappeared. It's come back. But 0:03:08.458 --> 0:03:09.568 tell us the story. 0:03:09.598 --> 0:03:13.288 This was actually quite bizarre. So early last week um, 0:03:13.288 --> 0:03:17.208 there was a shadow on Applevis saying, you know, you 0:03:17.208 --> 0:03:19.848 can't read tweets properly anymore. You only get the person 0:03:19.848 --> 0:03:22.998 who's tweeting his name or the handle, and you don't 0:03:22.998 --> 0:03:25.818 get anything else like the text of the actual tweet. 0:03:26.028 --> 0:03:29.478 So everybody started whinging about it and complaining about it 0:03:29.478 --> 0:03:33.258 on on Twitter, of course. And then everybody was thinking, 0:03:33.258 --> 0:03:34.818 I guess, or at least I wasn't like, oh God, 0:03:34.818 --> 0:03:36.348 here we go again. It's going to be like this 0:03:36.348 --> 0:03:37.878 all over again. We're going to have to actually fight 0:03:37.878 --> 0:03:42.698 our battle. But lo and behold, 48 hours later. So 0:03:42.698 --> 0:03:46.088 let's say from Wednesday to Friday last week, a new 0:03:46.088 --> 0:03:49.448 version popped up. So that was like a 0.01 release 0:03:49.448 --> 0:03:53.228 or whatever it was. And then the accessibility for VoiceOver 0:03:53.228 --> 0:03:55.958 came back. And again, I think it just said, you know, 0:03:55.958 --> 0:03:58.298 I release this for for bug fixes so you don't 0:03:58.298 --> 0:04:02.168 have to fix it, your typical sort of boilerplate statement. 0:04:02.248 --> 0:04:06.508 But it was fixed. So I don't know whether, you know, 0:04:06.508 --> 0:04:08.908 somebody saw it on Twitter and said, oh dear, here's 0:04:08.908 --> 0:04:13.378 a blindness community rightly getting annoyed again. Um, or there's 0:04:13.378 --> 0:04:17.758 still some software engineers at Twitter or X, because remember 0:04:17.758 --> 0:04:20.968 when good old Elon Musk took over, he got rid 0:04:20.998 --> 0:04:25.888 of lots of people, including the accessibility team for Twitter X. 0:04:25.888 --> 0:04:28.488 So everybody thought, well, here we go. This is going 0:04:28.488 --> 0:04:30.048 to be the end of end of days as we 0:04:30.048 --> 0:04:32.418 know it. Um, but yeah, as I just said, lo 0:04:32.418 --> 0:04:35.268 and behold, it works again. So just make sure that your, 0:04:35.268 --> 0:04:38.928 your app is completely updated and then you should be 0:04:38.928 --> 0:04:40.818 fine with voice over again. But I think it was 0:04:40.818 --> 0:04:44.388 just remarkable about how fast that bug was fixed. 0:04:44.388 --> 0:04:48.078 So for those who use ECS, things are back to 0:04:48.078 --> 0:04:53.598 normal or as they were smart AI apps with the 0:04:53.598 --> 0:04:58.008 smartphone camera. Let's have a talk about that today. Explain 0:04:58.008 --> 0:05:00.228 what that is. And then we can go through the 0:05:00.228 --> 0:05:02.478 various products that are available. 0:05:02.478 --> 0:05:04.518 Yep. So you might remember last week we spoke about 0:05:04.518 --> 0:05:09.588 three main wearables. We spoke about AR, vision, Orcam and 0:05:09.708 --> 0:05:14.118 envision AI. Now there's different sort of versions of that 0:05:14.118 --> 0:05:18.758 for your smartphone. So rather than wearing your, you know, 0:05:18.758 --> 0:05:22.388 your smart eye stuff, this stuff, as far as apps 0:05:22.388 --> 0:05:25.118 are concerned, are actually built into the smartphone, i.e. both 0:05:25.118 --> 0:05:28.838 Android and iPhone. So what we're really talking about is 0:05:28.838 --> 0:05:34.718 similar wearable hardware that is really apps running on your smartphone. 0:05:34.748 --> 0:05:39.818 Let's start then, with what's available through both iOS and Android. 0:05:40.468 --> 0:05:42.988 So the first one that I always love to mention, 0:05:42.988 --> 0:05:46.978 of course, is the be my eyes API function. And 0:05:46.978 --> 0:05:50.278 that's where you can take a photo. Um, and for 0:05:50.278 --> 0:05:52.678 me it's a lot more extensive than these sort of 0:05:52.678 --> 0:05:57.808 automatic scene detection apps. This one really goes through extensively 0:05:57.808 --> 0:06:00.658 what it sees in the photo. And then of course, 0:06:00.658 --> 0:06:03.718 you can also interrogate and ask questions about the photo. Um, 0:06:03.718 --> 0:06:06.048 it actually said my desk had coffee stains on it, 0:06:06.048 --> 0:06:09.648 which I was horrified by. So I went and quickly 0:06:09.648 --> 0:06:11.898 wiped it up. And then I ran it again and said, nope, nope. 0:06:11.898 --> 0:06:14.778 I didn't detect any coffee stains, so that was really good. 0:06:14.838 --> 0:06:18.108 But no, it actually does work really nicely. So, um, 0:06:18.108 --> 0:06:21.228 and as I said, for be my Eyes, I um, 0:06:21.228 --> 0:06:24.408 for both Android and iPhone, it works really nice. And 0:06:24.408 --> 0:06:26.358 also it's very, very quick. 0:06:26.358 --> 0:06:29.888 It's my go to place for getting descriptions of photos. 0:06:30.098 --> 0:06:32.228 Yeah, absolutely. Same same here. 0:06:32.528 --> 0:06:34.718 The next one envision I. 0:06:35.438 --> 0:06:38.678 Yeah. So this has the the similar functionality to what 0:06:38.678 --> 0:06:42.998 we talked about last week with the envision smart glasses. Uh, 0:06:42.998 --> 0:06:47.048 so you've got things like, you know, text detection, scene detection, 0:06:47.048 --> 0:06:52.058 currency detection, all that sort of normal stuff. Um, it's 0:06:52.058 --> 0:06:55.348 a very robust app. Now, I also happen to see 0:06:55.348 --> 0:06:59.098 in a beta testing last week to there's a new 0:06:59.098 --> 0:07:03.388 one coming out besides the envision AI app called Envision 0:07:03.388 --> 0:07:06.688 Assistant and what it is, it's a similar thing to 0:07:06.688 --> 0:07:09.028 what the Be My Eyes AI does. You take a 0:07:09.028 --> 0:07:12.778 photo and then you can use AI to interrogate the photo. 0:07:12.778 --> 0:07:14.458 And I've been having a look at the beta version 0:07:14.458 --> 0:07:16.798 of it, and it does work very nicely. And I 0:07:16.798 --> 0:07:20.718 think again, it's based on ChatGPT. So that's something else 0:07:20.718 --> 0:07:23.568 to look forward to, um, coming up in the next 0:07:23.568 --> 0:07:25.458 couple of months, or maybe even less. 0:07:25.458 --> 0:07:28.338 Tell us about seeing AI, which has been around a while. 0:07:28.398 --> 0:07:31.278 Yes. So this is a really famous one from Microsoft. Um, 0:07:31.278 --> 0:07:34.518 and you will remember that it does things like short text. 0:07:34.518 --> 0:07:41.508 So instant text, document reading, barcode scanning, scene detection, face detection, 0:07:41.508 --> 0:07:47.208 Handwriting detection. Color detection. Currency detection and also handwriting. So 0:07:47.208 --> 0:07:50.928 it does everything in the toolbox and actually works really well. 0:07:50.928 --> 0:07:55.098 So it does expand the functionality of both Android and 0:07:55.098 --> 0:07:59.058 iOS very dramatically. So that's almost my other one of 0:07:59.058 --> 0:08:01.878 my my favorite toolkit apps. When I'm talking about AI 0:08:01.878 --> 0:08:02.598 all the time. 0:08:02.598 --> 0:08:06.278 What about scene detection across both platforms? 0:08:06.488 --> 0:08:08.498 It really does get a bit wonky because I think 0:08:08.498 --> 0:08:11.858 last week when we talked about the hardware, I did 0:08:11.858 --> 0:08:13.598 leave this to the end of the discussion when I 0:08:13.598 --> 0:08:15.548 was talking about all the features, because I did say 0:08:15.548 --> 0:08:20.048 scene detection is not robust enough. It's based on what 0:08:20.048 --> 0:08:24.398 the camera is seeing, it's based on light and everything else. 0:08:24.398 --> 0:08:27.908 And then it has to then, you know, put that 0:08:27.908 --> 0:08:30.518 image against other images that it's previously had a look 0:08:30.518 --> 0:08:32.368 at and then come out of result. And I always 0:08:32.368 --> 0:08:37.678 find it's about, oh, 85, 90% accurate. But I noticed 0:08:37.678 --> 0:08:39.688 on an app that I haven't mentioned the program, it's 0:08:39.688 --> 0:08:43.198 not very stable. There was a warning on it. Please 0:08:43.198 --> 0:08:45.928 do not take the information that this app will give 0:08:45.928 --> 0:08:50.968 you as being true, because images are affected by lots 0:08:50.968 --> 0:08:54.538 of different processing issues. So I thought that was a really, 0:08:54.538 --> 0:08:56.418 really good warning and a good thing to keep in 0:08:56.418 --> 0:08:57.258 mind in general. 0:08:57.258 --> 0:09:01.308 And is that on both Android and iOS? Yes. Mhm. 0:09:01.338 --> 0:09:04.848 Tell us about an Android One called lookout. 0:09:05.478 --> 0:09:09.738 So this is from Google itself. And it is almost 0:09:09.738 --> 0:09:14.058 like Google's versions of Microsoft seeing AI app. So it 0:09:14.058 --> 0:09:18.888 basically does everything that they're seeing AI app does. Now 0:09:18.888 --> 0:09:21.488 there's one little thing that's a bit of an unknown 0:09:21.488 --> 0:09:23.438 feature in Seeing Eye that people might have forgotten about. 0:09:23.678 --> 0:09:27.248 You can say, I want to find something, and then 0:09:27.248 --> 0:09:31.088 you determine what you want to find, such as, you know, 0:09:31.088 --> 0:09:33.098 your keys and someone. So you're literally take a photo 0:09:33.098 --> 0:09:35.528 of your keys and then you say, these are my keys. 0:09:35.528 --> 0:09:37.568 And next time it's looking for your keys, it knows 0:09:37.568 --> 0:09:40.208 what your keys look like. Well, there's sort of a 0:09:40.208 --> 0:09:45.088 similar version on the lookout app called find, except it's 0:09:45.088 --> 0:09:47.428 a bit more generic. So when you say I want 0:09:47.428 --> 0:09:49.288 to find something, it brings up a list of things 0:09:49.288 --> 0:09:53.848 like door, chair, table, cup and so on. And then 0:09:53.848 --> 0:09:56.668 you choose one of those particular items so it knows 0:09:56.668 --> 0:09:59.548 it's got to look for a cup, and then it 0:09:59.548 --> 0:10:01.768 will tell you how far it's away when it sees 0:10:01.768 --> 0:10:05.128 that particular item, in this case a cup. So it's 0:10:05.128 --> 0:10:10.098 not as flexible as the CI1 where you can either 0:10:10.098 --> 0:10:12.918 choose generic ones or customize it, but it still works 0:10:12.918 --> 0:10:17.508 really well. So look out to me is, uh, let's 0:10:17.508 --> 0:10:20.688 say 95% of what the AI app can do. 0:10:20.838 --> 0:10:24.948 Let's look forward to iOS 18, which comes to the 0:10:24.948 --> 0:10:29.958 public late September, probably. What have you been finding there 0:10:29.958 --> 0:10:32.418 with your beta exploration? 0:10:32.418 --> 0:10:35.808 This stuff is actually pretty amazing what Apple is doing. 0:10:35.808 --> 0:10:38.178 So all the stuff I'm about to talk about was 0:10:38.178 --> 0:10:41.838 originally in the magnification app. It's now part of VoiceOver rotor, 0:10:41.838 --> 0:10:44.238 and the whole sort of menu on your radar is 0:10:44.238 --> 0:10:48.258 called Live Detection. And underneath live detection you've got things 0:10:48.258 --> 0:10:53.088 like door detection, scene detection, the point and speak function 0:10:53.088 --> 0:10:58.628 for using touch screens, people detection, and of course text detection. 0:10:58.628 --> 0:11:01.988 So what happens is you use the rotor to rotate 0:11:01.988 --> 0:11:05.048 around to life detection. You flick down to the one 0:11:05.048 --> 0:11:07.898 you want, such as door detection. You double tap it, 0:11:08.198 --> 0:11:10.238 it then turns it on. But then what you've got 0:11:10.238 --> 0:11:14.498 is a four finger triple tap. That ten turns that 0:11:14.498 --> 0:11:19.028 voice detection on or off. So depending on what function 0:11:19.028 --> 0:11:21.838 you've got running in this case, you know, door or 0:11:22.048 --> 0:11:26.038 whatever else, it then only detects that particular function. The 0:11:26.038 --> 0:11:28.738 only one that's missing for me. And this is why 0:11:28.738 --> 0:11:30.388 I put it in the show notes for a list 0:11:30.388 --> 0:11:34.738 is object detection, where it actually uses the camera to 0:11:34.738 --> 0:11:37.018 tell you how far an object is away, whether it's 0:11:37.018 --> 0:11:40.998 in within, you know, arm's reach within about 1 or 0:11:40.998 --> 0:11:45.828 2m away or further away. So I think if Apple 0:11:45.828 --> 0:11:49.338 added that to the live detection rater, they'd practically have 0:11:49.338 --> 0:11:52.548 everything covered. And what I do find with the live 0:11:52.548 --> 0:11:56.268 detection very quickly is the scene detection on all these 0:11:56.268 --> 0:11:59.118 ones tends to babble a lot. So it keeps saying 0:11:59.118 --> 0:12:02.688 window blind, window blind, window blind, but because you can 0:12:02.688 --> 0:12:05.028 quickly toggle it on and off, you just turn it 0:12:05.028 --> 0:12:06.408 on when you want. As soon as you got the 0:12:06.408 --> 0:12:08.928 information you want, you turn it off again with that 0:12:08.928 --> 0:12:12.648 four finger triple tap and off you go. So it's 0:12:12.648 --> 0:12:16.248 getting very, very flexible in what VoiceOver can do in 0:12:16.248 --> 0:12:17.598 iOS 18 beta. 0:12:17.598 --> 0:12:21.288 This is very nice, all this stuff, but ergonomically it 0:12:21.288 --> 0:12:23.838 can be a bit challenging. A lot of us walk 0:12:23.838 --> 0:12:27.138 along with only one hand available, because the other hand 0:12:27.138 --> 0:12:30.588 might be holding a cane or the harness of a dog. 0:12:30.588 --> 0:12:33.108 How do we get around this? Because you need to 0:12:33.108 --> 0:12:34.708 be able to manage the phone. 0:12:34.738 --> 0:12:36.268 The first thing is you've got to have a thing 0:12:36.298 --> 0:12:38.758 to put the phone in. So normally it's a chest harness. 0:12:38.758 --> 0:12:41.938 My chest harness that I've got for my smartphone does portrait. 0:12:41.938 --> 0:12:44.458 So up and down all landscape. So that's number one. 0:12:44.728 --> 0:12:49.498 Number two which I haven't found to be that dramatically fantastic, 0:12:49.498 --> 0:12:53.008 is using voice control to use voice over to then 0:12:53.008 --> 0:12:56.728 navigate the smartphone screen and for iOS. So things like, 0:12:56.728 --> 0:13:00.408 you know, flick left flick right, magic double tap. That 0:13:00.408 --> 0:13:03.738 gets very laborious when you're trying to use, um, an 0:13:03.738 --> 0:13:07.548 object recognition type app. The third one, which is what 0:13:07.548 --> 0:13:10.728 I always use on my one, is a little keyboard 0:13:10.728 --> 0:13:12.798 that I've had for a long time called the Revo 0:13:12.798 --> 0:13:17.118 Keyboard Revo. It's about the size of a credit card. Um, 0:13:17.118 --> 0:13:18.798 of course a little bit thicker because it's got keys 0:13:18.798 --> 0:13:21.438 on it and because I can use it one handed. 0:13:21.438 --> 0:13:23.508 All I really need to do with that keyboard is 0:13:23.508 --> 0:13:26.478 press four and six to go left and right. I 0:13:26.478 --> 0:13:29.868 press five to select and a few other little keys. 0:13:29.868 --> 0:13:32.568 So basically I can use about 5 or 6 keys 0:13:32.568 --> 0:13:35.598 to completely navigate voice over. Of course, the only thing 0:13:35.598 --> 0:13:38.478 it hasn't got for iOS beta yet is to turn 0:13:38.478 --> 0:13:42.618 that voice live recognition on and off toggle. But besides that, 0:13:42.618 --> 0:13:45.828 I can completely control my iPhone from this little keyboard. 0:13:46.038 --> 0:13:49.448 Well, there's a lot more with AI and the camera 0:13:49.448 --> 0:13:51.818 we'll talk about in the future, and I look forward 0:13:51.818 --> 0:13:54.878 to that. It's going to be interesting. Yep. Before we go, 0:13:54.878 --> 0:13:57.008 a reminder of where there are details of what we've 0:13:57.008 --> 0:14:00.008 been talking about in this and previous editions of the program. 0:14:00.008 --> 0:14:01.928 Indeed. So as always, you can check out my blog 0:14:01.928 --> 0:14:06.938 site which is David Woodburn dot Podbean pad Bayern Comm David. 0:14:06.938 --> 0:14:12.628 Wood beard Podbean podbean.com to right to the program, you can. 0:14:12.628 --> 0:14:14.458 Write to me at Vision Australia or Work, which is 0:14:14.458 --> 0:14:19.438 David Woodbridge how it sounds at Vision australia.org David. 0:14:19.438 --> 0:14:23.608 Dot Woodbridge at Vision Australia. Org this has been talking 0:14:23.638 --> 0:14:26.848 tech with me has been Vision Australia's national advisor on 0:14:26.848 --> 0:14:30.868 access technology David Woodbridge I'm Stephen Jolly take care. We'll 0:14:30.868 --> 0:14:32.398 talk more tech next week. See you.