0:21

Hello everyone. Welcome to Talking

0:23

Tech. This edition available from July

0:25

the 2nd, 2024. I'm

0:27

Stephen Jolly. Great to have you with us listening

0:30

maybe through Vision Australia radio

0:32

associated stations of RPA Australia

0:34

or perhaps the Community Radio Network.

0:36

There is also the podcast. To catch

0:39

that, all you need to do is search for the two

0:41

words talking tech and down.

0:43

It can all come usually on a Tuesday afternoon

0:45

just after it's been produced. Another

0:47

option is to ask your Siri device or

0:49

smart speaker to play Vision Australia

0:52

Radio talking tech podcast

0:54

Vision Australia Radio talking tech

0:56

podcast with me, someone who can

0:58

explain all this tech stuff really well.

1:00

Vision Australia's national advisor on

1:03

access technology, David Woodbridge.

1:05

David, let's start with our Product

1:07

Minute, a product from the vision store

1:10

of Vision Australia.

1:11

Indeed. So this is called the

1:13

Archon Desk stand

1:15

and Archons actually Archon.

1:19

It's an extremely adjustable

1:21

stand for a smartphone. So

1:23

you've got the base which is made out of

1:25

metal. Then you've got this telescopic

1:27

pole that actually extends out, I

1:29

think, up to a meter if you pull it right

1:31

out. Um, normally it's about 30cm.

1:34

And then the arm that

1:37

comes out which holds the smartphone,

1:39

it's got these little articulated little

1:41

arms that you can adjust in all different directions,

1:44

because each little arm of which is three

1:47

has got little ball joints. So you can sort of rotate

1:49

them around. So in effect

1:51

you can have your smartphone

1:53

in any position you like, whether it's portrait,

1:56

landscape, upside down, whatever

1:58

you like. So it's a really

2:00

good stand for when

2:02

you want to do things like video conferencing

2:05

or optical character recognition

2:07

and so on. My wife

2:09

uses her one and her craft room on

2:11

Facebook and so on. She uses it

2:13

so that on her main laptop screen,

2:16

on her Mac, she can see what she's doing

2:18

on YouTube and Facebook, and

2:20

then she just uses the camera on her

2:22

smartphone to point down at what she's doing for

2:24

crafting. So it does have multiple

2:27

uses. It's a well known standard. It's been

2:29

around for a long time, so if you're

2:31

after a really good desktop,

2:33

as in you put it on your table stand for

2:35

a smartphone, it comes in really handy.

2:38

So again, it's Arkon Arkon

2:41

desktop smartphone stand.

2:44

How much does it cost?

2:45

It's about $225.

2:47

There is an attachment which we don't sell for an

2:49

iPad, but you can always

2:51

get them from a, you know, overseas or Amazon.

2:53

The icon desktop stand from

2:56

the vision store at Vision Australia.

2:59

There was chatter last week about

3:02

X. We used to call it Twitter.

3:05

Its accessibility disappeared.

3:07

It's come back. But tell us the story.

3:09

This was actually quite bizarre. So

3:11

early last week um, there was

3:13

a shadow on Applevis saying,

3:16

you know, you can't read tweets

3:18

properly anymore. You only get the person who's tweeting

3:20

his name or the handle, and

3:22

you don't get anything else like the text of

3:24

the actual tweet. So everybody

3:27

started whinging about it and complaining about

3:29

it on on Twitter, of course. And

3:31

then everybody was thinking, I guess, or

3:33

at least I wasn't like, oh God, here we go again. It's going to

3:35

be like this all over again. We're going to have to actually fight

3:37

our battle. But lo and behold,

3:40

48 hours later. So

3:42

let's say from Wednesday to Friday last week,

3:45

a new version popped up. So that was like

3:47

a 0.01 release or whatever

3:49

it was. And then the accessibility

3:51

for VoiceOver came back.

3:54

And again, I think it just said, you know, I release

3:56

this for for bug fixes so you don't have to

3:58

fix it, your typical sort of boilerplate

4:01

statement. But it was fixed.

4:03

So I don't know whether,

4:05

you know, somebody saw it on Twitter

4:07

and said, oh dear, here's a blindness community

4:10

rightly getting annoyed again.

4:12

Um, or there's still some software engineers

4:14

at Twitter or X, because

4:17

remember when good old Elon

4:19

Musk took over, he got rid of

4:21

lots of people, including the accessibility

4:23

team for Twitter X.

4:25

So everybody thought, well, here we

4:27

go. This is going to be the end of end of days as

4:29

we know it. Um, but yeah, as I just said,

4:32

lo and behold, it works again. So just

4:34

make sure that your, your app is completely

4:36

updated and then you should be

4:38

fine with voice over again. But I think it was just

4:40

remarkable about how fast that

4:43

bug was fixed.

4:44

So for those who use ECS,

4:46

things are back to normal or

4:48

as they were smart

4:51

AI apps with the

4:53

smartphone camera. Let's have a talk

4:55

about that today. Explain

4:58

what that is. And then we can go through

5:00

the various products that are available.

5:02

Yep. So you might remember last week we spoke about

5:04

three main wearables. We spoke

5:06

about AR, vision, Orcam

5:09

and envision AI. Now

5:11

there's different sort of versions of that

5:14

for your smartphone. So rather than

5:16

wearing your,

5:18

you know, your smart eye stuff, this

5:20

stuff, as far as apps are concerned,

5:22

are actually built into the smartphone, i.e. both

5:25

Android and iPhone. So what we're

5:27

really talking about is similar

5:29

wearable hardware

5:31

that is really apps running on your smartphone.

5:34

Let's start then, with what's available

5:37

through both iOS and Android.

5:40

So the first one that I always love

5:42

to mention, of course, is the be my eyes

5:44

API function. And

5:46

that's where you can take a photo. Um,

5:49

and for me it's a lot more extensive

5:51

than these sort of automatic scene detection

5:54

apps. This one really

5:56

goes through extensively what it

5:58

sees in the photo. And then

6:00

of course, you can also interrogate and ask questions

6:02

about the photo. Um, it actually said

6:04

my desk had coffee stains on it, which I was

6:06

horrified by. So

6:08

I went and quickly wiped it up. And then I ran

6:10

it again and said, nope, nope. I didn't detect any coffee

6:13

stains, so that was really good. But no,

6:15

it actually does work really nicely. So,

6:18

um, and as I said, for be my Eyes, I

6:21

um, for both Android and iPhone, it works

6:23

really nice. And also it's very,

6:25

very quick.

6:26

It's my go to place for getting

6:28

descriptions of photos.

6:30

Yeah, absolutely. Same same here.

6:32

The next one envision I.

6:35

Yeah. So this has the the similar functionality

6:38

to what we talked about last week with

6:40

the envision smart glasses.

6:42

Uh, so you've got things like, you know, text

6:45

detection, scene detection, currency

6:47

detection, all that sort of normal stuff.

6:50

Um, it's a very robust

6:53

app. Now, I also happen to see

6:55

in a beta testing last week

6:57

to there's a new one coming

6:59

out besides the envision AI app

7:02

called Envision Assistant

7:04

and what it is, it's a similar

7:06

thing to what the Be My Eyes AI does.

7:08

You take a photo and then you can use

7:10

AI to interrogate the photo. And

7:12

I've been having a look at the beta version of it, and

7:15

it does work very nicely. And I think again,

7:17

it's based on ChatGPT. So

7:19

that's something else to look forward to,

7:22

um, coming up in the next couple of months,

7:24

or maybe even less.

7:25

Tell us about seeing AI, which has been around

7:27

a while.

7:28

Yes. So this is a really famous one from Microsoft.

7:31

Um, and you will remember that it does

7:33

things like short text. So instant text,

7:35

document reading, barcode

7:38

scanning, scene detection, face

7:40

detection, Handwriting detection.

7:42

Color detection. Currency detection

7:45

and also handwriting. So it does

7:47

everything in the toolbox and

7:49

actually works really well. So it does expand

7:52

the functionality of both

7:54

Android and iOS very

7:56

dramatically. So that's almost

7:58

my other one of my my favorite toolkit

8:00

apps. When I'm talking about AI all the time.

8:02

What about scene detection across

8:04

both platforms?

8:06

It really does get a bit wonky because I think

8:08

last week when we talked about the hardware,

8:10

I did leave this to the end of the discussion

8:13

when I was talking about all the features, because I did say

8:15

scene detection is not robust

8:17

enough. It's based on what

8:20

the camera is seeing, it's based on

8:22

light and everything else.

8:24

And then it has to then, you know,

8:26

put that image against other

8:29

images that it's previously had a look at and then come out

8:31

of result. And I always find it's

8:33

about, oh, 85,

8:35

90% accurate. But I noticed on

8:37

an app that I haven't mentioned the program, it's not

8:39

very stable. There was a

8:41

warning on it. Please do not take

8:43

the information that this app will give

8:45

you as being true,

8:48

because images are affected by

8:50

lots of different processing issues.

8:53

So I thought that was a really, really good warning

8:55

and a good thing to keep in mind in general.

8:57

And is that on both Android and iOS?

9:00

Yes. Mhm. Tell us about

9:02

an Android One called lookout.

9:05

So this is from Google itself. And

9:07

it is almost like

9:10

Google's versions of Microsoft

9:12

seeing AI app. So it basically

9:14

does everything that they're

9:16

seeing AI app does. Now

9:18

there's one little thing that's a bit of an unknown

9:21

feature in Seeing Eye that people might have forgotten about. You

9:24

can say, I want to find

9:26

something, and then you determine what you

9:28

want to find, such as,

9:30

you know, your keys and someone. So you're literally take

9:32

a photo of your keys and then you say, these

9:34

are my keys. And next time it's looking for your keys,

9:37

it knows what your keys look like. Well,

9:39

there's sort of a similar version on

9:41

the lookout app called find,

9:44

except it's a bit more generic. So

9:46

when you say I want to find something, it brings up

9:48

a list of things like door, chair,

9:50

table, cup and

9:53

so on. And then you choose one of those particular

9:55

items so it knows it's got to

9:57

look for a cup, and then

9:59

it will tell you how far it's away when it sees

10:01

that particular item, in this case a cup.

10:03

So it's not as

10:06

flexible as the CI1

10:08

where you can either choose generic

10:10

ones or customize it, but it still works

10:12

really well. So look out to me

10:15

is, uh,

10:17

let's say 95% of what the

10:19

AI app can do.

10:20

Let's look forward to iOS 18,

10:23

which comes to the public late

10:26

September, probably. What

10:28

have you been finding there with your

10:30

beta exploration?

10:32

This stuff is actually pretty

10:34

amazing what Apple is doing. So

10:36

all the stuff I'm about to talk about was originally

10:38

in the magnification app. It's now part of VoiceOver

10:40

rotor, and the whole

10:42

sort of menu on your radar is called Live Detection.

10:45

And underneath live detection you've got

10:47

things like door detection,

10:50

scene detection, the point and

10:52

speak function for using touch screens,

10:55

people detection, and of course

10:57

text detection. So what happens

10:59

is you use the rotor to

11:01

rotate around to life detection. You

11:04

flick down to the one you want, such as

11:06

door detection. You double tap it, it

11:08

then turns it on. But then what you've got is

11:10

a four finger triple tap.

11:12

That ten turns that voice

11:15

detection on or off.

11:17

So depending on what function you've

11:19

got running in this case, you know,

11:21

door or whatever else, it then only

11:23

detects that particular function.

11:25

The only one that's missing for me.

11:27

And this is why I put it in the show notes

11:29

for a list is object detection,

11:32

where it actually uses the camera

11:34

to tell you how far an object is away,

11:36

whether it's in within, you know, arm's

11:38

reach within about 1 or

11:40

2m away or further away.

11:43

So I think if

11:45

Apple added that to the live detection

11:47

rater, they'd practically have everything

11:49

covered. And what I do find with

11:51

the live detection very quickly is

11:54

the scene detection on all these ones

11:56

tends to babble a lot. So it keeps saying

11:59

window blind, window blind, window blind,

12:01

but because you can quickly toggle it on and off,

12:04

you just turn it on when you want. As soon as you

12:06

got the information you want, you turn it off again

12:08

with that four finger triple

12:10

tap and off you go. So it's

12:12

getting very, very flexible in

12:14

what VoiceOver can do in iOS 18

12:17

beta.

12:17

This is very nice, all this stuff,

12:19

but ergonomically it can be a bit challenging.

12:22

A lot of us walk along with only one hand

12:25

available, because the other hand

12:27

might be holding a cane or

12:29

the harness of a dog. How do we get around

12:31

this? Because you need to be able to manage

12:33

the phone.

12:34

The first thing is you've got to have a thing to put the phone

12:36

in. So normally it's a chest harness. My

12:39

chest harness that I've got for my smartphone does

12:41

portrait. So up and down all landscape.

12:43

So that's number one. Number two

12:45

which I haven't found to be that dramatically

12:48

fantastic, is using voice control

12:51

to use voice over to then navigate

12:53

the smartphone screen and for iOS.

12:55

So things like, you know, flick left flick

12:57

right, magic double tap. That

13:00

gets very laborious when you're trying to use,

13:03

um, an object recognition type app. The

13:06

third one, which is what I always use

13:08

on my one, is a little keyboard

13:10

that I've had for a long time called the Revo

13:12

Keyboard Revo. It's

13:14

about the size of a credit card. Um,

13:17

of course a little bit thicker because it's got keys on it

13:19

and because I can use it one handed.

13:21

All I really need to do with that keyboard is

13:23

press four and six to go left and right.

13:25

I press five to select

13:28

and a few other little keys. So basically

13:30

I can use about 5 or 6 keys

13:32

to completely navigate voice over.

13:34

Of course, the only thing it hasn't got for iOS

13:36

beta yet is to turn that

13:38

voice live recognition on and off toggle.

13:41

But besides that, I can completely

13:43

control my iPhone from this little keyboard.

13:46

Well, there's a lot more with AI

13:48

and the camera we'll talk about in the future,

13:50

and I look forward to that. It's going to be interesting.

13:53

Yep. Before we go, a reminder

13:55

of where there are details of what we've been talking

13:57

about in this and previous editions of the program.

14:00

Indeed. So as always, you can check out my blog site

14:02

which is David Woodburn dot Podbean

14:04

pad Bayern Comm David.

14:06

Wood beard Podbean podbean.com

14:11

to right to the program, you can.

14:12

Write to me at Vision Australia or Work, which is David

14:14

Woodbridge how it sounds

14:17

at Vision australia.org David.

14:19

Dot Woodbridge at Vision Australia.

14:22

Org this has been talking tech with

14:24

me has been Vision Australia's national advisor

14:26

on access technology David Woodbridge

14:28

I'm Stephen Jolly take care. We'll talk

14:31

more tech next week. See you.