Latest / Sportstalk with D'Arcy Waldegrave / Sir Ian Taylor: former broadcaster on the technical issues impacting the Ashes
Transcript
- 0:06Speaker 1: You're listening to the Sports Talk podcast with Dancy Wildergrave
- 0:10Speaker 1: from News Talk Zed.
- 0:11Speaker 2: Be right, let's talk about the ashes from yesterday. A
- 0:15Speaker 2: bit of drama, a technical issue. Alix Carey scored a
- 0:18Speaker 2: century apparently should have been out when it's in seventy two.
- 0:21Speaker 3: They used to review the.
- 0:22Speaker 2: English on a court behind. They thought they heard something
- 0:25Speaker 2: microphones on the stumps, clear noise, but the replay reckon
- 0:29Speaker 2: it all happened before past the back, resulting in a
- 0:31Speaker 2: not out call. So Kerry comes out, admitted that he
- 0:34Speaker 2: hit the ball, and the company that owns Snacko concluded
- 0:37Speaker 2: that the operator at the time must have chosen the incorrect.
- 0:39Speaker 3: Stump MIC for audio. How do you do that?
- 0:42Speaker 2: There must be a process, not this one, this one.
- 0:45Speaker 2: Sirien Taylor joins us now he's a synonymous with the
- 0:47Speaker 2: on screen graphics for televised sporty runs Virtuly and duaneed
- 0:51Speaker 2: and Ian thanks very much for joining us. Of course
- 0:54Speaker 2: your name sprang to mind straight away after the Snacko
- 0:57Speaker 2: issues and the ashes yesterday, and you're at the forefront
- 1:00Speaker 2: of the way this technology works. So where are we
- 1:03Speaker 2: at at the moment that tech and sport identifying things
- 1:06Speaker 2: like next because this goes through all sport, right.
- 1:10Speaker 3: Yes it does.
- 1:11Speaker 4: And I mean the interesting thing is it's the technology
- 1:16Speaker 4: has been getting better. So we've been doing it for
- 1:19Speaker 4: I don't know many, twenty twenty five years, I guess,
- 1:22Speaker 4: and you know, just probably in the last three to
- 1:26Speaker 4: four years, we've noticed a massive move in the technology
- 1:31Speaker 4: that helps. And of course everyone talks about AI, but
- 1:34Speaker 4: AI and machine learning is another thing that is really
- 1:38Speaker 4: make a big difference. I think where they got caught
- 1:41Speaker 4: out yesterday. We don't do the edge detection.
- 1:45Speaker 2: We do the.
- 1:46Speaker 4: Runouts and the LBWs, but I think they got caught
- 1:49Speaker 4: out yesterday, and the edge detection is it's really important.
- 1:53Speaker 3: You've got two elements.
- 1:54Speaker 4: You've got what the camera is seeing and what the
- 1:57Speaker 4: microphone's hearing, and those have to be absolutely aligned with
- 2:01Speaker 4: each other, or you get the thing that appears to
- 2:05Speaker 4: have happened yesterday. And from the reports I'm seeing, those
- 2:08Speaker 4: two were out of sync. And so that puts the
- 2:11Speaker 4: umpires in a bit of a tough position because they
- 2:13Speaker 4: can see the ball is past the bat when they
- 2:17Speaker 4: see the sound, So what do they do? You know,
- 2:21Speaker 4: They've got to go with what they can see and
- 2:25Speaker 4: I think that was the issue that happened yesterday.
- 2:27Speaker 2: You used to have hot spot, whatever happened to that.
- 2:29Speaker 2: That was a very popular piece of technology. What did
- 2:33Speaker 2: that get phase here?
- 2:34Speaker 4: Yeah, so I thought hot thought was really really good.
- 2:38Speaker 4: It was only ever used in Australia. It didn't get
- 2:41Speaker 4: it picked up by the ICC.
- 2:43Speaker 3: Well.
- 2:44Speaker 4: One of the really important run of the ridicol things
- 2:46Speaker 4: about hotspot was that if you saw the edge of
- 2:49Speaker 4: the bat light up, it meant the ball hit the bat.
- 2:52Speaker 4: The issue two issues was that because well the same issue,
- 2:59Speaker 4: because the cameras were really really expensive. They're very specialist
- 3:02Speaker 4: cameras to be able to see that hotspot come up.
- 3:07Speaker 4: They couldn't afford to have as many as you'd need
- 3:10Speaker 4: to be able to see all the edges of the bat,
- 3:13Speaker 4: and so it just kind of priced itself out because
- 3:18Speaker 4: to be really effective you needed it on both sides
- 3:21Speaker 4: of the wicket, So that was at least four of
- 3:24Speaker 4: them that you needed, and probably a couple of front
- 3:26Speaker 4: on and it just got too expensive to use. But
- 3:30Speaker 4: you know what we're seeing with some of the new
- 3:32Speaker 4: technologies which will probably start you know, you'll start seeing
- 3:36Speaker 4: unravel in the next year or so, is that AI
- 3:39Speaker 4: and machine learning is helping to give way more certainty.
- 3:43Speaker 4: It's made a big difference with the lbw's not many
- 3:46Speaker 4: people question those anymore because the technology has advanced so
- 3:51Speaker 4: far with the high speed cameras well.
- 3:53Speaker 2: It makes sense right that we do utilize AI and
- 3:56Speaker 2: its best form as opposed to the scary, more sinister
- 4:00Speaker 2: sides of it. So this is good, But does that
- 4:01Speaker 2: make the technology after it's being developed? Does that make
- 4:05Speaker 2: it cheaper or less expensive? I will it be accessible
- 4:10Speaker 2: right around the globe? Because in order for it to
- 4:12Speaker 2: be I suppose given the tick by the ICC, it
- 4:15Speaker 2: needs to be applied everywhere, doesn't it. It can't just
- 4:18Speaker 2: be with the Aussies or you know, with a couple
- 4:20Speaker 2: of high rolling teams. So is it going to be
- 4:22Speaker 2: many cheaper really less expensive.
- 4:25Speaker 3: No, it's it's getting less expensive.
- 4:28Speaker 4: And the thing is that you know, there's there's two
- 4:33Speaker 4: edge detection applications now, so the one they are using
- 4:37Speaker 4: in Australia which is Snicko, and the other one is
- 4:41Speaker 4: called Ultra Edge. So you have to have ball tracking,
- 4:44Speaker 4: LBW and edge detection in every game now everywhere, and
- 4:50Speaker 4: so that is going all over the world, and I
- 4:52Speaker 4: think any changes that technology bring will be applied to everything.
- 4:57Speaker 4: One of the things we're seeing which is really interesting
- 5:01Speaker 4: is like I'm on my iPhone seventeen whatever it's called,
- 5:05Speaker 4: you know, whatever numbers they're up to at the moment,
- 5:08Speaker 4: and these things are getting incredibly sophisticated as well now,
- 5:13Speaker 4: so you know, they have lighter in them, they have
- 5:15Speaker 4: all sorts of things. So we're seeing that technology change
- 5:19Speaker 4: really fast, and it wouldn't surprise me if we start
- 5:22Speaker 4: to see iPhones being used in those lesser matches, maybe
- 5:28Speaker 4: even club matches in the future, to give people ball
- 5:32Speaker 4: tracking and maybe even you know, edge detection.
- 5:35Speaker 2: The acceleration of this technology, you see, you've been around
- 5:38Speaker 2: twenty five years or so doing this exponential it's just
- 5:41Speaker 2: going through the roof. It's got how much more can
- 5:44Speaker 2: we do? You mean, you mentioned the seventeen, the iPhone seventeen,
- 5:46Speaker 2: But where's the limit? Is there a limit from what
- 5:49Speaker 2: you can see?
- 5:50Speaker 4: Well, you know, it's funny you should say that because
- 5:54Speaker 4: you know, we started we started the company in nineteen
- 5:57Speaker 4: ninety and that was the same year that the World
- 6:00Speaker 4: Wide Web was invented, and it's sort of like, what's
- 6:03Speaker 4: this thing, the Internet? What are we going to use
- 6:05Speaker 4: that for? And the acceleration. In nineteen ninety two, the
- 6:11Speaker 4: computer we use to do the first America's Cup cost
- 6:14Speaker 4: half a million dollars and it was the size of
- 6:16Speaker 4: a fridge. And now you know, we deliver all this
- 6:19Speaker 4: stuff on the phone you carry in your pocket, so
- 6:22Speaker 4: that the escalation has been I guess over thirty five
- 6:26Speaker 4: years gradual. But the biggest change we've seen really is
- 6:32Speaker 4: AI and that has just gone to another level where
- 6:36Speaker 4: we're seeing the changes almost every week.
- 6:39Speaker 2: As far as the work behind it, I'm presuming you've
- 6:42Speaker 2: got to input piles and piles of data, so much
- 6:47Speaker 2: data to I suppose train these machines up, it takes
- 6:51Speaker 2: the longest Doore.
- 6:52Speaker 4: Yeah, that's a really good point because the thing about
- 6:57Speaker 4: about AI is it's only as good as you know
- 7:01Speaker 4: what it's been taught and how much it's been taught.
- 7:04Speaker 3: So we have for the last year we have we.
- 7:08Speaker 4: Built an AI engine for cricket, for scoring, for identifying,
- 7:13Speaker 4: you know, what are all the cool things happening in
- 7:14Speaker 4: a match. And over the last year we've put ten
- 7:19Speaker 4: thousand hours off cricket through it, and we've had students
- 7:23Speaker 4: from the Polytech and we've they come in and as
- 7:27Speaker 4: part of a training course, they sit in there and
- 7:29Speaker 4: they identify. So give you one example, and there's lots
- 7:33Speaker 4: of examples. Let's say the noball. So nobles get missed
- 7:37Speaker 4: quite often. You know, they look at them really closely
- 7:41Speaker 4: when they're when there's an appeal, but actually they get
- 7:44Speaker 4: missed a lot. So over the year there's been ten
- 7:48Speaker 4: thousand hours of cricket gone through our AI engine and
- 7:51Speaker 4: somebody has sat there and taught the AI exactly what
- 7:57Speaker 4: a noball is, and they we've got doing it now.
- 8:01Speaker 4: And the people who worked on it were laughing saying,
- 8:03Speaker 4: you know, it's taken us a year to teach AI
- 8:06Speaker 4: what a noble looks like, and all we have to
- 8:09Speaker 4: do is look at it and we know straight away.
- 8:12Speaker 3: That's one side of it.
- 8:13Speaker 4: But the other side of it is, now that the
- 8:15Speaker 4: AI knows what a no ball is, it can do
- 8:18Speaker 4: a million games at once, a million balls at the
- 8:20Speaker 4: same time anywhere in the world. And that's a huge
- 8:24Speaker 4: change that's happening now. So things get easier, they get cheaper,
- 8:28Speaker 4: they'll get done in the cloud, but it means you
- 8:30Speaker 4: can do it way more effectively because you've finally taught
- 8:35Speaker 4: it how.
- 8:36Speaker 3: To do it cold come, but a lot of work.
- 8:38Speaker 4: Yeah, oh, I mean it was probably the worst worst
- 8:43Speaker 4: person that could happen to, reallyy.
- 8:46Speaker 2: But on that look, ten thousand hours. That's so significant,
- 8:50Speaker 2: Sirian Tartar, and thanks very much for joining us, because
- 8:52Speaker 2: you know, to become an expert in something, you've got
- 8:54Speaker 2: to spend ten thousand hours, right, that's our humans operator
- 8:56Speaker 2: as well.
- 8:57Speaker 3: Mate, So well done it is, isn't it. Well you
- 8:59Speaker 3: have a great Christmas.
- 9:00Speaker 2: Doting and you as well.
- 9:01Speaker 3: Thanks for your time, yere goy dear see it.
- 9:04Speaker 1: For more from sports Talk, listen live toin News Talks.
- 9:07Speaker 1: It'd be from seven pm weekdays, or follow the podcast
- 9:10Speaker 1: on iHeartRadio.