Fix Bad Audio Instantly? Inside Waves’ Voice Regeneration Tool

1
00:00:00,080 --> 00:00:00,440
Y'all ready?

2
00:00:00,440 --> 00:00:01,800
Be history. Get started.

3
00:00:01,800 --> 00:00:02,320
Welcome.

4
00:00:02,320 --> 00:00:03,680
Hi. Hi. Hi.

5
00:00:03,680 --> 00:00:06,600
Hello, everyone, to the pro audio suite.

6
00:00:06,600 --> 00:00:08,760
These guys are professional and motivated.

7
00:00:08,760 --> 00:00:09,680
Please take the video.

8
00:00:09,680 --> 00:00:12,960
Stars George Wisdom,
founder of Source Element Robert Marshall,

9
00:00:12,960 --> 00:00:16,960
international audio engineer Darren Robin
Roberts and global voice Andrew.

10
00:00:17,440 --> 00:00:21,240
Thanks to tribal Austrian audio lighting,
passionate elements

11
00:00:21,400 --> 00:00:25,360
George, the tech wisdom and Rob and APIs
international demos.

12
00:00:25,440 --> 00:00:26,520
Find out more about us.

13
00:00:26,520 --> 00:00:29,280
Check the Proteas sweetcorn line up.

14
00:00:29,280 --> 00:00:32,280
Ready? Here we go. I'm.

15
00:00:32,720 --> 00:00:34,480
And welcome to another pro audio suite.

16
00:00:34,480 --> 00:00:37,320
Thanks to try Booth. Don't
forget the code.

17
00:00:37,320 --> 00:00:41,240
Try Pip 200 to get $200 off yours.

18
00:00:41,240 --> 00:00:45,000
And by the time this goes to air,
I'll have mine set up somewhere

19
00:00:45,480 --> 00:00:48,160
remotely in Australia

20
00:00:48,160 --> 00:00:51,160
and Austrian audio making passion heard.

21
00:00:51,880 --> 00:00:54,880
We're hearing from our special guest
today, Michael Pearce and Adams.

22
00:00:54,960 --> 00:00:56,840
Or as we like to call him down here.

23
00:00:56,840 --> 00:01:00,200
Gomez from Waves
Audio has a brand new product,

24
00:01:00,200 --> 00:01:04,000
which I have to admit,
I think I looked at, but I can't remember.

25
00:01:04,360 --> 00:01:09,360
It's all to do with hormone treatment
that that is messing with my brain.

26
00:01:09,520 --> 00:01:12,520
I think the funniest thing
about that entire intro is

27
00:01:12,520 --> 00:01:16,440
it sounds like a testimonial on channel
ten at all. Yep.

28
00:01:16,480 --> 00:01:18,120
At 11:00 at night?

29
00:01:18,120 --> 00:01:18,480
Yeah.

30
00:01:18,480 --> 00:01:19,560
It's funny you should mention that.

31
00:01:19,560 --> 00:01:21,840
That's Andrew's pedigree. Yeah.

32
00:01:21,840 --> 00:01:24,400
I mean, a guy called Brad Turner

33
00:01:24,400 --> 00:01:28,560
used to film those in Melbourne years
and years at, like, 1990.

34
00:01:28,920 --> 00:01:29,360
It's like.

35
00:01:29,360 --> 00:01:33,960
And we've got Brad and Michael here from
double TFM to talk about a new product.

36
00:01:34,480 --> 00:01:36,400
Hang on a minute.

37
00:01:36,400 --> 00:01:37,920
Were you doing that? Yeah,

38
00:01:38,960 --> 00:01:39,960
I used to go down there.

39
00:01:39,960 --> 00:01:42,960
I was with Brad. It was a double T.

40
00:01:43,040 --> 00:01:44,840
Yeah. BT in me.

41
00:01:44,840 --> 00:01:45,480
It was fun.

42
00:01:45,480 --> 00:01:47,080
Oh my God, we must have cross paths.

43
00:01:47,080 --> 00:01:50,560
John Dramatis was also involved
at some point.

44
00:01:50,560 --> 00:01:52,440
We all used to make money like that.

45
00:01:52,440 --> 00:01:54,240
Those were the years indeed.

46
00:01:54,240 --> 00:01:56,480
Anyway, anyway,
thanks guys, for having me here.

47
00:01:56,480 --> 00:01:57,600
It's been a pleasure.

48
00:01:57,600 --> 00:01:59,800
I guess I'll see you next time, right?
Yeah. Thanks.

49
00:01:59,800 --> 00:02:01,880
Yeah. And, Good idea. Brad,
if you're listening to.

50
00:02:01,880 --> 00:02:06,520
Yeah, tell us about this new product,
because for some reason,

51
00:02:06,520 --> 00:02:09,360
I thought I would be good to help you
with the beta testing,

52
00:02:09,360 --> 00:02:11,840
so I know all about it, but
you should probably feel our listeners.

53
00:02:11,840 --> 00:02:15,680
So the one of the first things that I need

54
00:02:15,680 --> 00:02:20,400
to mention about what we call
this region is it's

55
00:02:20,400 --> 00:02:24,600
not a plug in
and it will never be a plug in.

56
00:02:24,600 --> 00:02:26,920
And we can talk about that part
later down the line.

57
00:02:26,920 --> 00:02:31,840
But one of the largest and fastest
growing,

58
00:02:32,360 --> 00:02:35,520
audiences, or of users

59
00:02:35,520 --> 00:02:38,720
who create some kind of audio
in the world,

60
00:02:39,640 --> 00:02:44,920
not w users or video
editor users, the content creators.

61
00:02:44,920 --> 00:02:50,440
And, one of the things that I thought
really long and hard about was

62
00:02:51,960 --> 00:02:53,400
they might not

63
00:02:53,400 --> 00:02:56,840
understand how to fix, bad audio,

64
00:02:56,840 --> 00:03:01,320
but they do understand
what bad audio sounds like.

65
00:03:01,320 --> 00:03:03,680
So let's give them a solution.

66
00:03:03,680 --> 00:03:09,200
So we spent about three and a half years
working on building a very large language

67
00:03:09,200 --> 00:03:13,880
model, and training it on
what good audio sounds like.

68
00:03:14,400 --> 00:03:17,840
And we've come up with this product
called Voice Region,

69
00:03:17,840 --> 00:03:21,160
which can make even the worst

70
00:03:22,400 --> 00:03:25,680
meal or retrievable dialog sound

71
00:03:26,840 --> 00:03:28,040
nearly like a podcast.

72
00:03:28,040 --> 00:03:30,640
Mike and very, very usable.

73
00:03:30,640 --> 00:03:33,640
And on the website,
some of the demos I use,

74
00:03:33,960 --> 00:03:37,440
actually most of the demos
are very Australian because, you know,

75
00:03:37,440 --> 00:03:41,480
I have access to a lot of content
creators here, but

76
00:03:41,760 --> 00:03:45,200
literally one of the cases was a podcast,

77
00:03:45,280 --> 00:03:49,160
recording from Riverside, where on video

78
00:03:49,680 --> 00:03:53,040
the guest looked like they were recording
into a nice microphone.

79
00:03:53,040 --> 00:03:54,440
But the reality was

80
00:03:54,440 --> 00:03:55,120
they were actually

81
00:03:55,120 --> 00:03:58,320
recording on their MacBook
microphone, which was across the room.

82
00:03:59,440 --> 00:04:02,640
And in that came a classic mistake.

83
00:04:02,640 --> 00:04:03,120
Common.

84
00:04:03,120 --> 00:04:06,680
Well, it's it's a hugely common problem,
but it's also one

85
00:04:06,680 --> 00:04:08,280
that is emotionally very,

86
00:04:08,280 --> 00:04:12,040
very stressful for somebody who thinks
they've got great content and realizes

87
00:04:12,400 --> 00:04:16,440
it could be completely and utterly useless
because of that fact.

88
00:04:17,040 --> 00:04:21,440
And this is kind of why I built
voice Region, not for the people who

89
00:04:22,320 --> 00:04:25,960
mostly control the content,
but for all of those out of control

90
00:04:25,960 --> 00:04:29,040
scenarios where bad things happen

91
00:04:29,680 --> 00:04:32,440
and you have to work out,
how do I retrieve this?

92
00:04:32,440 --> 00:04:33,560
Do I pay somebody?

93
00:04:33,560 --> 00:04:38,360
Do I learn how to use, tool
that I've never seen before?

94
00:04:38,400 --> 00:04:40,720
Is this something
I can just drag and drop it onto

95
00:04:40,720 --> 00:04:43,200
and have it fix it for me,
which is what we built.

96
00:04:43,200 --> 00:04:46,520
It's quite seriously some of
those examples that you have up there,

97
00:04:47,840 --> 00:04:48,600
the normal

98
00:04:48,600 --> 00:04:51,720
plug in would probably be of no help
anyway.

99
00:04:51,720 --> 00:04:53,120
They're so far gone.

100
00:04:53,120 --> 00:04:54,120
Yeah, that you would

101
00:04:54,120 --> 00:04:57,200
only be detrimental to the audio
you were trying to keep anyway.

102
00:04:57,280 --> 00:04:59,400
And obviously,
if you think about the amount of time

103
00:04:59,400 --> 00:05:00,920
that it would, it would take.

104
00:05:00,920 --> 00:05:04,960
So for example, Robbo, if you've been paid
to take one of those pieces of audio,

105
00:05:05,240 --> 00:05:06,880
you'd spend a few hours on it.

106
00:05:06,880 --> 00:05:09,880
You'd still go back to it
and go, no, I'm still not happy with it.

107
00:05:11,360 --> 00:05:15,400
And one of the first things that we had
when we put the first demos out

108
00:05:15,400 --> 00:05:19,560
was people from, you know, our existing
audience, which are all plugin users,

109
00:05:19,560 --> 00:05:22,560
which we're very grateful for, saying,
make this a plugin,

110
00:05:22,840 --> 00:05:24,200
and we're constantly explaining.

111
00:05:24,200 --> 00:05:27,240
Now, the reason
this will never be a plugin

112
00:05:27,240 --> 00:05:32,640
is because it would break the
CPU and GPU on any computer.

113
00:05:33,640 --> 00:05:34,160
It doesn't

114
00:05:34,160 --> 00:05:37,600
matter if it's an M5
that came out last week or,

115
00:05:38,120 --> 00:05:41,520
you know, a really powerful PC,
it just would break it.

116
00:05:42,640 --> 00:05:46,760
It's not possible if this takes
too much energy and too much power.

117
00:05:47,360 --> 00:05:49,120
So it's in the cloud.

118
00:05:49,120 --> 00:05:51,160
Tell us, what are we doing to the audio?

119
00:05:51,160 --> 00:05:53,800
I mean it's called voice region.

120
00:05:53,800 --> 00:05:56,360
We're literally
well we're literally regenerating.

121
00:05:56,360 --> 00:05:58,600
So we're listening to what it is.

122
00:05:58,600 --> 00:06:00,680
We're taking the tonality.

123
00:06:00,680 --> 00:06:05,120
As much as we can find,
and we're literally rebuilding it

124
00:06:05,120 --> 00:06:09,160
so that it is
what the LM expects it to be.

125
00:06:09,600 --> 00:06:14,160
One of the examples
that displays that fairly clearly on

126
00:06:14,160 --> 00:06:17,520
the website is a Creative Commons
public recording

127
00:06:17,520 --> 00:06:21,320
from one of the 60s trips to the moon,

128
00:06:21,600 --> 00:06:24,880
where we take the NASA

129
00:06:24,880 --> 00:06:28,640
radio from the moon and make it sound
like he's talking into a podcast mic.

130
00:06:30,400 --> 00:06:32,000
And you can tell it's the same person.

131
00:06:32,000 --> 00:06:35,520
It's just it just one minute he's
on the moon, in the next minute he's not.

132
00:06:35,960 --> 00:06:38,760
And so this.

133
00:06:38,760 --> 00:06:40,440
But was he that was he on the moon?

134
00:06:40,440 --> 00:06:43,800
I wasn't going to go there,
but I didn't want to, you know. So.

135
00:06:44,360 --> 00:06:48,120
And my point would be that, Robert,
there's hope for our podcast.

136
00:06:48,120 --> 00:06:49,040
Yes, man.

137
00:06:49,040 --> 00:06:50,040
Yes. That's right.

138
00:06:50,040 --> 00:06:52,880
Well, I mean,
I think I'll just stop recording right now

139
00:06:52,880 --> 00:06:54,920
and you can just capture
this directly off the phone.

140
00:06:54,920 --> 00:06:56,160
Right.

141
00:06:56,160 --> 00:06:56,960
Good enough.

142
00:06:56,960 --> 00:07:00,240
No, no you that Robert it it please,
please please don't.

143
00:07:00,280 --> 00:07:04,600
But take into account Robert that
I actually can do that with voice region.

144
00:07:04,760 --> 00:07:05,680
Right. That's what I'm saying.

145
00:07:05,680 --> 00:07:07,800
I mean like live by the sword, die
by the sword.

146
00:07:07,800 --> 00:07:08,720
Let's let's do it, Rob.

147
00:07:08,720 --> 00:07:12,640
Oh, come on, I'm already doing it
with Gomezs audio actually going to be.

148
00:07:12,880 --> 00:07:15,480
And there's and there's
a couple of different ways to use this.

149
00:07:15,480 --> 00:07:17,440
And there's a couple of different things

150
00:07:17,440 --> 00:07:20,600
that I've built in
and that we're about to kind of add.

151
00:07:20,920 --> 00:07:26,080
One of them is I've given, people
the ability to record directly on to the,

152
00:07:26,120 --> 00:07:29,120
in the app on the either on their phone

153
00:07:29,600 --> 00:07:31,880
or on a computer.

154
00:07:31,880 --> 00:07:33,600
And I've given them a place
to put this script.

155
00:07:33,600 --> 00:07:34,680
And you can either use it

156
00:07:34,680 --> 00:07:37,960
as a teleprompter and have it scroll,
or you can just kind of stick

157
00:07:38,000 --> 00:07:39,600
speed to zero in the script,

158
00:07:39,600 --> 00:07:42,040
will stay there in front of you
so that you can actually

159
00:07:42,040 --> 00:07:43,800
look at the script
while you're talking to a mic,

160
00:07:43,800 --> 00:07:47,720
the same way a voiceover does,
and then for for a.

161
00:07:47,720 --> 00:07:49,320
And one of the reasons we did

162
00:07:49,320 --> 00:07:53,280
this is because a lot of the time,
if talent is remote,

163
00:07:53,960 --> 00:07:58,800
whether it's a podcast guest or somebody
else, you can still make them sound

164
00:07:59,320 --> 00:08:01,960
like they're,
in front of a better microphone

165
00:08:01,960 --> 00:08:05,760
and the room it
with just drag and drop and process.

166
00:08:06,320 --> 00:08:10,680
We're about to add in two weeks,
a pop out teleprompter

167
00:08:11,280 --> 00:08:12,560
so that you can actually put

168
00:08:12,560 --> 00:08:16,400
the teleprompter, on a script,
a separate screen, make it really large

169
00:08:16,680 --> 00:08:20,040
just underneath your camera
so that you can use a teleprompter.

170
00:08:20,040 --> 00:08:23,400
And we're also adding record video
onto the recording page

171
00:08:23,760 --> 00:08:26,760
so you can select a video camera,
select your microphone.

172
00:08:27,000 --> 00:08:29,800
And even if it's the MacBook microphone
and you're in the kitchen,

173
00:08:29,800 --> 00:08:32,640
we'll make it sound like you've got
you're wearing a lapel.

174
00:08:32,640 --> 00:08:34,920
So this all sets up on a server,
is that correct?

175
00:08:34,920 --> 00:08:35,640
Yeah.

176
00:08:35,640 --> 00:08:39,600
And is this part of the,
what's the bundle?

177
00:08:39,600 --> 00:08:42,160
You know, the the subscription.

178
00:08:42,160 --> 00:08:43,200
Do you get that? No, no.

179
00:08:43,200 --> 00:08:46,400
If you're if you're if you're
if you're paying for creative access

180
00:08:46,400 --> 00:08:49,760
plug ins, then you're not getting region.

181
00:08:49,760 --> 00:08:52,760
If you want region
then just go and subscribe.

182
00:08:53,120 --> 00:08:54,920
I mean, I just made it super cheap.

183
00:08:54,920 --> 00:08:58,120
I mean, to get five hours of processing
per month,

184
00:08:58,120 --> 00:09:01,440
I'm charging $4.99 for us right now.

185
00:09:02,200 --> 00:09:02,640
That's cheap.

186
00:09:02,640 --> 00:09:06,600
But I also give you,
everybody on a free account

187
00:09:06,600 --> 00:09:10,520
which needs no billing address,
no credit card or anything.

188
00:09:10,880 --> 00:09:12,600
You get five minutes free per day.

189
00:09:12,600 --> 00:09:16,160
So if you think about all the short
form content and then people could

190
00:09:16,160 --> 00:09:21,040
literally record on their phone,
with the video, not have to worry.

191
00:09:21,040 --> 00:09:26,000
Lapel put the video straight into voice
region and then post it on TikTok, etc..

192
00:09:27,480 --> 00:09:28,440
And by the way, we're

193
00:09:28,440 --> 00:09:32,680
finding the video
files are about 50% of the files

194
00:09:32,680 --> 00:09:35,680
being uploaded to voice region
right now. So

195
00:09:36,080 --> 00:09:37,680
so it takes in video.

196
00:09:37,680 --> 00:09:40,480
Dymocks is it does the audio processing
remarks is a video.

197
00:09:40,480 --> 00:09:41,280
Absolutely.

198
00:09:41,280 --> 00:09:43,680
We give you back in MP4 with clean
audience.

199
00:09:43,680 --> 00:09:44,680
Here's a question then.

200
00:09:44,680 --> 00:09:48,480
And look I know it probably doesn't
need to be answered by you guys, but

201
00:09:48,480 --> 00:09:50,240
I'm sure plenty of people have asking it.

202
00:09:50,240 --> 00:09:54,440
Given the AI temperature at the moment,
shall we say,

203
00:09:54,440 --> 00:10:00,000
what happens to all these AI samples
once they've been processed?

204
00:10:00,000 --> 00:10:04,120
Your audio stays on your dashboard
until 14 days,

205
00:10:04,120 --> 00:10:06,680
and then at that point
it's automatically deleted.

206
00:10:06,680 --> 00:10:08,680
And if you look at the user dashboard, it
says,

207
00:10:08,680 --> 00:10:10,680
I think I've got it in like two places
right now.

208
00:10:10,680 --> 00:10:15,240
It says anything you have in
your files is deleted after 14 days.

209
00:10:15,800 --> 00:10:18,560
One of the things that we do not do

210
00:10:18,560 --> 00:10:21,240
is use any user's files

211
00:10:21,240 --> 00:10:24,040
to try and improve their model,
and model is our model.

212
00:10:24,040 --> 00:10:27,360
We improve it by paying for content

213
00:10:27,360 --> 00:10:31,120
from specific companies that do that
kind of thing for two reasons.

214
00:10:31,120 --> 00:10:32,920
Number one, it's because it's, you know,

215
00:10:32,920 --> 00:10:35,920
we want to stay 100% legal
with where our content comes from.

216
00:10:35,960 --> 00:10:39,400
Number two,
scraping is a really, really bad idea.

217
00:10:39,400 --> 00:10:42,400
It's unethical, it's not moral,
and it's lying.

218
00:10:42,760 --> 00:10:45,600
So we do not use any user's

219
00:10:45,600 --> 00:10:48,600
audio to help educate our model at all.

220
00:10:48,800 --> 00:10:51,800
And by the way,
there's another reason for that.

221
00:10:52,200 --> 00:10:55,920
When you've got a large language
model like this, you don't educate it

222
00:10:55,920 --> 00:10:57,520
by giving it terrible audio.

223
00:10:57,520 --> 00:11:01,040
You you educate it
by giving it very high quality audio.

224
00:11:01,880 --> 00:11:05,240
It's not like
we can educate this by giving it

225
00:11:05,840 --> 00:11:09,000
region processed audio,
because that's a bit like,

226
00:11:09,680 --> 00:11:13,400
you know, open AI chat about scraping
and finding more information

227
00:11:13,400 --> 00:11:15,040
that it created that was wrong.

228
00:11:15,040 --> 00:11:18,720
And it's it's, you know, eating it's
the beast is eating its tail.

229
00:11:18,720 --> 00:11:19,920
Yeah. It's exactly.

230
00:11:19,920 --> 00:11:21,840
Yeah. It's like it's not the way to do it.

231
00:11:21,840 --> 00:11:24,720
But you know, it's
nobody's information gets used.

232
00:11:24,720 --> 00:11:29,520
When does the 499 limited time
because it's normally 999 right.

233
00:11:29,520 --> 00:11:31,000
Well for the creator.

234
00:11:31,000 --> 00:11:31,360
Fine.

235
00:11:31,360 --> 00:11:35,880
Normally it's like ultimately
as far as I'm just, unconcerned,

236
00:11:36,320 --> 00:11:42,240
I will keep going on the full 99
until I feel like I've reached,

237
00:11:42,320 --> 00:11:47,160
a point where, we're established.

238
00:11:47,200 --> 00:11:51,240
We released this on January 26th,
and in a very quiet launch

239
00:11:51,840 --> 00:11:55,920
so far, we we have about 35,000 members,
which, you know, for.

240
00:11:56,120 --> 00:11:58,480
Wow, it's a month and a half
I'm pretty proud of.

241
00:11:58,480 --> 00:11:59,920
Wow. Yes.

242
00:11:59,920 --> 00:12:03,400
We've processed
a ridiculous amount of audio

243
00:12:03,400 --> 00:12:06,720
in that time, which, I also keep tabs on,

244
00:12:07,160 --> 00:12:11,760
but I want people to keep on joining
because one of the things that happens

245
00:12:11,760 --> 00:12:16,160
when you have a free account
and you're using it is occasionally

246
00:12:16,160 --> 00:12:20,640
an error happens
or a file that you upload doesn't work.

247
00:12:20,760 --> 00:12:24,160
And those kind of errors do get flagged.

248
00:12:24,160 --> 00:12:26,320
And they help me and my team.

249
00:12:26,320 --> 00:12:30,120
And I'm looking at the data
every day to see, okay,

250
00:12:30,120 --> 00:12:33,120
I can't hear the file,
I can't see the file.

251
00:12:33,120 --> 00:12:36,840
But if it fails,
I can see what codec it was

252
00:12:37,200 --> 00:12:40,040
and kind of like
what links or what was in it,

253
00:12:40,040 --> 00:12:45,240
so that I actually can learn and improve
how we deal with errors

254
00:12:45,560 --> 00:12:49,680
and also keep on adding,
oh, there's that codec from that.

255
00:12:49,680 --> 00:12:50,400
Okay, cool.

256
00:12:50,400 --> 00:12:53,000
Let's add that
I want to keep the price down

257
00:12:53,000 --> 00:12:57,000
so that we keep on bringing more people
in, by the way, process really fast.

258
00:12:57,000 --> 00:13:01,320
So I mean it immediately tells you
to download a WAV file, I guess that

259
00:13:01,600 --> 00:13:06,080
did you do that so that you could ensure
that there be a local copy for the user?

260
00:13:06,080 --> 00:13:06,760
Absolutely.

261
00:13:06,760 --> 00:13:10,280
And if you once you've done
that once, effectively what will happen

262
00:13:10,280 --> 00:13:13,280
is you press record as you just did.

263
00:13:13,440 --> 00:13:16,520
You can preview it
and go, yeah, I, I'm gonna like that.

264
00:13:16,520 --> 00:13:17,560
I'm going to process that.

265
00:13:17,560 --> 00:13:21,120
As soon as that happens, you press, okay,
it will take a copy of that

266
00:13:21,120 --> 00:13:25,200
and put it into your default downloads
folder so that you have a redundancy.

267
00:13:26,560 --> 00:13:29,880
And for my
perspective, anybody who wants to record

268
00:13:29,880 --> 00:13:34,200
anything should have a backup of that
that's not touched by any server.

269
00:13:34,200 --> 00:13:36,600
So if you haven't got it back out,
it doesn't cut it at all. Right. Yeah.

270
00:13:36,600 --> 00:13:37,680
Exactly. Exactly.

271
00:13:37,680 --> 00:13:40,920
I mean, remember this is developed
by somebody whose theory is

272
00:13:40,920 --> 00:13:43,880
there's no such thing as a backup
unless as the backup of the backup.

273
00:13:43,880 --> 00:13:46,560
Right. You don't have a copy
unless you have two copies. Exactly.

274
00:13:46,560 --> 00:13:46,800
Yeah.

275
00:13:46,800 --> 00:13:49,080
And the other thing
is, once you click on a file,

276
00:13:49,080 --> 00:13:52,120
you have a toggle between the before
and after audio.

277
00:13:52,720 --> 00:13:56,040
And most of the time,
you can just tell by even looking

278
00:13:56,040 --> 00:13:59,040
at the WAV file
that something dramatic has happened.

279
00:13:59,040 --> 00:14:01,680
Is the processing
basically faster than real time?

280
00:14:01,680 --> 00:14:04,560
Way faster than really quick,
really quick?

281
00:14:04,560 --> 00:14:05,280
Yeah.

282
00:14:05,280 --> 00:14:06,640
So how'd you go, Robert?

283
00:14:06,640 --> 00:14:08,760
If you got something,
you can play I. George. Sorry, you guys.

284
00:14:08,760 --> 00:14:12,280
So here's here's, Robert on the phone,
I don't know, I'm going to ask Gomes.

285
00:14:12,760 --> 00:14:16,320
When does the 499 limited time not.

286
00:14:16,880 --> 00:14:19,560
It's because it's normally 999. Right.

287
00:14:19,560 --> 00:14:21,720
All right. And here's
the processed version I don't know.

288
00:14:21,720 --> 00:14:26,840
I'm going to ask Gomes,
when does the port 9900 time not.

289
00:14:27,360 --> 00:14:29,440
It's because it's normally 999, right.

290
00:14:29,440 --> 00:14:30,280
Isn't that bizarre?

291
00:14:30,280 --> 00:14:33,280
It sounds all like a phone to me.

292
00:14:33,360 --> 00:14:37,080
Yeah, yeah, let's let's hear
technology, man.

293
00:14:37,080 --> 00:14:39,360
Sorry. Yeah. That's right.

294
00:14:39,360 --> 00:14:41,200
Yeah, it is freaky.

295
00:14:41,200 --> 00:14:44,160
The first time I did that,
I was not on the phone.

296
00:14:44,160 --> 00:14:48,040
I haven't done a phone one like that,
but no, but the thing is, you know, it's

297
00:14:48,040 --> 00:14:51,120
like we've got we've got a couple of radio
networks in the States

298
00:14:51,120 --> 00:14:54,200
and a radio network in Australia
using this now.

299
00:14:54,480 --> 00:14:58,440
So that phone interviews,
even if it's down to an office to,

300
00:14:58,480 --> 00:15:01,800
you know, doing tags,
hey, it's such and such

301
00:15:01,800 --> 00:15:05,160
and you're listening to suddenly
it's not a phone interview anymore.

302
00:15:05,160 --> 00:15:07,760
Suddenly
you've got them sitting in a studio record

303
00:15:07,760 --> 00:15:10,760
doing the tags,
and they can be used in proper promos.

304
00:15:10,840 --> 00:15:12,880
Take all your old phone IDs and,

305
00:15:14,000 --> 00:15:14,800
hundred percent you

306
00:15:14,800 --> 00:15:17,920
could have you could have Jim Morrison
saying goodbye.

307
00:15:17,920 --> 00:15:21,320
This is Jim Morrison sounding
like he was sitting across the room.

308
00:15:21,320 --> 00:15:23,280
So I don't think Jim Morrison
that's like it.

309
00:15:23,280 --> 00:15:25,800
I just tend not to mind.

310
00:15:25,800 --> 00:15:30,400
So as a as an experiment, it's
just because I don't really have much of

311
00:15:30,400 --> 00:15:30,880
a life.

312
00:15:30,880 --> 00:15:34,320
I about to about three weeks ago,

313
00:15:34,320 --> 00:15:39,400
I went looking for the oldest John
Lennon interviews,

314
00:15:39,400 --> 00:15:42,400
and I found a video of one

315
00:15:42,600 --> 00:15:45,120
that had so much ground.

316
00:15:45,120 --> 00:15:49,080
And, you know, it was like
the interview was in a white coat

317
00:15:49,080 --> 00:15:50,520
and all very, very. And.

318
00:15:50,520 --> 00:15:50,880
Yeah.

319
00:15:50,880 --> 00:15:54,160
And and I put it through voice reach in
and then I put it

320
00:15:54,160 --> 00:15:57,320
through a visual sharpener called Topaz
video.

321
00:15:57,560 --> 00:16:03,400
And I ended up watching this kind of like
Netflix HD video of John Lennon.

322
00:16:03,400 --> 00:16:06,400
And it's it sounded amazing.

323
00:16:06,640 --> 00:16:10,440
I mean, this has some fairly
it has a lot of use cases.

324
00:16:10,440 --> 00:16:12,360
This song

325
00:16:12,360 --> 00:16:15,480
is probably probably a bit advanced
for where you're at with the software.

326
00:16:15,480 --> 00:16:19,440
But just thinking as an audio engineer
now, like we sat to hear the

327
00:16:19,920 --> 00:16:22,080
John Lennon interview
that we're just talking about,

328
00:16:22,080 --> 00:16:24,880
can I pan left and right
or is it just all straight up?

329
00:16:24,880 --> 00:16:27,360
So right now
it's all up the center and in.

330
00:16:27,360 --> 00:16:31,920
And so for example, right now if you also
if you put in a multi-channel file,

331
00:16:32,360 --> 00:16:35,040
at this point in time,

332
00:16:35,040 --> 00:16:38,560
if it's only stereo,
we will, will mono it.

333
00:16:38,800 --> 00:16:43,240
If it's a multi-channel
more than two channels

334
00:16:43,240 --> 00:16:46,520
at this point in time,
we'll politely tell you we can't do it.

335
00:16:46,520 --> 00:16:49,520
Or if we think we can do it,

336
00:16:49,560 --> 00:16:52,200
we're only doing the first channel.

337
00:16:52,200 --> 00:16:53,600
Right.

338
00:16:53,600 --> 00:16:54,720
We're working on that.

339
00:16:54,720 --> 00:17:00,600
That's another reason why I study the data
so much, is so that we can learn from

340
00:17:01,080 --> 00:17:04,800
how the system deals
with, different kinds of files.

341
00:17:05,360 --> 00:17:06,240
It's for spoken.

342
00:17:06,240 --> 00:17:08,160
This is for dialog only. Yes.

343
00:17:08,160 --> 00:17:09,640
Speech.

344
00:17:09,640 --> 00:17:12,400
Yeah. Yeah,
but I can see, like Peter Jackson.

345
00:17:13,360 --> 00:17:14,200
I'm surprised he hasn't

346
00:17:14,200 --> 00:17:17,200
been on the, on the hornpipe saying,
can I have a go?

347
00:17:17,280 --> 00:17:21,280
We already have a couple
of post-production studios who have said

348
00:17:21,280 --> 00:17:26,080
we used this for boom mics from a film
when one of the mics failed.

349
00:17:27,000 --> 00:17:28,560
So it's like.

350
00:17:28,560 --> 00:17:31,960
And one of the reasons
why we've got a pro account which has

351
00:17:32,760 --> 00:17:36,800
13 hours, and then we'll end up
with an enterprise account.

352
00:17:37,040 --> 00:17:40,160
The other thing that we're doing with
this is we're also about to launch

353
00:17:40,160 --> 00:17:43,160
an API version of Voice region,

354
00:17:43,440 --> 00:17:46,920
so that you can actually plug
a voice region into your own funnel,

355
00:17:47,000 --> 00:17:50,160
and kind of create your own service
out of it.

356
00:17:50,360 --> 00:17:54,080
The only person I can say, or part
of the industry suffering from this is

357
00:17:54,080 --> 00:17:57,840
if you're into audio dialog replacement,
because you won't be needed anymore.

358
00:17:57,840 --> 00:17:59,760
Well, well, yes or no.

359
00:17:59,760 --> 00:18:01,280
I mean, there's a couple of things.

360
00:18:01,280 --> 00:18:07,720
Firstly, a lot of the situations
that voice region is useful to a user

361
00:18:08,120 --> 00:18:11,800
are the users who would never
go to a dialog editor in the first place.

362
00:18:12,320 --> 00:18:14,520
They don't even know what the term means.

363
00:18:14,520 --> 00:18:16,520
All they know is their audio is terrible.

364
00:18:16,520 --> 00:18:18,240
They need a solution. What?

365
00:18:18,240 --> 00:18:21,120
What I'm not selling them is a hammer
and a nail.

366
00:18:21,120 --> 00:18:25,320
I'm I'm I'm selling them
a nail hammered into a wall.

367
00:18:26,320 --> 00:18:27,080
Yeah.

368
00:18:27,080 --> 00:18:29,880
Okay. They don't need the process.
They don't understand the process.

369
00:18:29,880 --> 00:18:32,880
They will never know who to go to
or who to ask for.

370
00:18:33,200 --> 00:18:36,200
So therefore, I'm not really risking
anybody's career.

371
00:18:36,360 --> 00:18:39,880
The other thing that's important
here is the content creator

372
00:18:39,880 --> 00:18:42,960
market is like 150 million times

373
00:18:42,960 --> 00:18:45,960
bigger than the audio pro market.

374
00:18:46,200 --> 00:18:48,920
And, I feel like

375
00:18:48,920 --> 00:18:52,200
I'm giving it to like 1% of it.

376
00:18:52,200 --> 00:18:56,760
Maybe so I don't I'm not losing sleep,
feeling like I'm

377
00:18:56,800 --> 00:18:58,640
taking anybody's job away.

378
00:18:58,640 --> 00:19:00,600
I'll tell you someone who could
he could benefit from this.

379
00:19:00,600 --> 00:19:03,880
And as a podcast, I was watching earlier
and I'll tell you exactly

380
00:19:03,880 --> 00:19:07,600
what it's called,
I think, racing news three, six, five.

381
00:19:09,480 --> 00:19:12,520
And they, they do a Formula One podcast

382
00:19:13,680 --> 00:19:17,520
every time there's someone beaming in

383
00:19:17,680 --> 00:19:21,200
to them, like a guest or something,
or the person's on location,

384
00:19:21,800 --> 00:19:24,480
they compressed the shit out of it
to a point

385
00:19:24,480 --> 00:19:27,720
where it's so gated
that half the stuff is missing.

386
00:19:27,720 --> 00:19:30,160
It's like, what the hell did you like?
Woods have gone. Yeah.

387
00:19:30,160 --> 00:19:31,800
It's like, yeah.

388
00:19:31,800 --> 00:19:34,800
It's like,
come on, I've sent them a message saying

389
00:19:34,920 --> 00:19:37,680
your gates count is too heavy
with the gating.

390
00:19:37,680 --> 00:19:39,600
I cut off half your dialog is gone.

391
00:19:39,600 --> 00:19:42,600
Well, you raise an interesting point
that I was going to make

392
00:19:42,680 --> 00:19:45,480
was that, you know,
they're they're fed content creators.

393
00:19:45,480 --> 00:19:47,880
There actually is a direct correlation

394
00:19:47,880 --> 00:19:51,880
between the quality of your audio
and how much people believe you.

395
00:19:52,160 --> 00:19:52,520
Yeah.

396
00:19:52,520 --> 00:19:55,920
It's actually been studied
in at the University of Queensland

397
00:19:55,920 --> 00:19:59,200
and that one in California,
when we talked about this before,

398
00:19:59,200 --> 00:20:04,720
you know, it's like you can draw a line
almost in terms of how listener perception

399
00:20:04,720 --> 00:20:08,560
of the information you're giving them
falls as your audio quality gets more.

400
00:20:08,640 --> 00:20:12,960
Doesn't matter
if you've got an ethics three camera

401
00:20:12,960 --> 00:20:18,360
with a Sigma lens, anamorphic
perfect lighting, and your sharp focus.

402
00:20:18,360 --> 00:20:21,280
If your audio sounds like you're recording

403
00:20:21,280 --> 00:20:25,120
on a Logitech
C nine to a webcam in a kitchen,

404
00:20:26,200 --> 00:20:26,760
which you

405
00:20:26,760 --> 00:20:30,440
probably, it's just a random side note

406
00:20:30,600 --> 00:20:36,280
the highest selling webcams
still in 2026 is the Logitech C9 two.

407
00:20:36,440 --> 00:20:38,320
I've got one.

408
00:20:38,320 --> 00:20:39,800
I think I'm using it right now.

409
00:20:39,800 --> 00:20:40,960
Literally, literally.

410
00:20:40,960 --> 00:20:46,200
It's like I did
I did, I have, a report software

411
00:20:46,240 --> 00:20:50,360
that I pay for a lot for every month
that tells me how many things sell on

412
00:20:50,360 --> 00:20:52,000
Amazon per product.

413
00:20:52,000 --> 00:20:56,240
This thing just on Amazon
sells like at least 20,000 a day.

414
00:20:56,240 --> 00:20:57,240
It's insane.

415
00:20:57,240 --> 00:20:58,560
It's not a bad camera. Lenses.

416
00:20:58,560 --> 00:21:00,960
I just see nine, two, two.

417
00:21:00,960 --> 00:21:03,360
Oh, you've got the posh version
anyway, so.

418
00:21:03,360 --> 00:21:06,080
But yes, good audio is confidence.

419
00:21:06,080 --> 00:21:08,680
And that's kind of one of the things
that from my perspective,

420
00:21:08,680 --> 00:21:13,080
I wanted with voice
region is to give people confidence

421
00:21:13,080 --> 00:21:16,920
because, you know, if you look amazing
but you sound is really rough.

422
00:21:17,480 --> 00:21:20,040
It does lower your self-esteem.

423
00:21:20,040 --> 00:21:22,640
And you think,
do I really want to put that out there?

424
00:21:22,640 --> 00:21:25,760
What I want is somebody to feel like
they can just literally

425
00:21:26,040 --> 00:21:29,280
pick up their phone, record
something, hit process, and

426
00:21:29,840 --> 00:21:32,840
send it out into the world,
know that it sounds amazing.

427
00:21:33,200 --> 00:21:36,200
Without knowing how it was achieved.

428
00:21:36,960 --> 00:21:38,640
We'll focus on the content. Yeah.

429
00:21:38,640 --> 00:21:42,240
And let us look after your
your polish, I agree.

430
00:21:42,240 --> 00:21:45,480
So if you like what you've come up with
for people who are doing content,

431
00:21:45,480 --> 00:21:48,520
instead of worrying about all the shit
that they have no idea what it is,

432
00:21:49,080 --> 00:21:53,080
just do what you do best and let the
you know machine do the work for you.

433
00:21:53,080 --> 00:21:54,280
The other bit for you.

434
00:21:54,280 --> 00:21:58,640
The other thing is that, and I
this is something which I, denied about

435
00:21:58,640 --> 00:22:02,000
was how large a file

436
00:22:02,000 --> 00:22:05,320
I wanted to enable people to play with.

437
00:22:06,120 --> 00:22:08,880
So you can upload anything to voice
region,

438
00:22:08,880 --> 00:22:12,720
up to 1.5 gig in size.

439
00:22:13,280 --> 00:22:16,440
So now, 1.5 gig.

440
00:22:16,440 --> 00:22:20,520
If you're at 724, you know,

441
00:22:20,520 --> 00:22:24,840
something like that with video
you're talking, I mean, I ran

442
00:22:25,800 --> 00:22:28,600
one of the webinars
I did with kind of like Andrew Shepp.

443
00:22:28,600 --> 00:22:31,680
So somebody I ran the entire webinar,

444
00:22:31,680 --> 00:22:35,320
720p after I ripped it down off YouTube
through Voice Region.

445
00:22:35,320 --> 00:22:39,360
And that was like two hours and it cleaned
up at vocals, something chronic.

446
00:22:39,360 --> 00:22:40,440
It was amazing.

447
00:22:40,440 --> 00:22:42,600
But 1.5 gig I think is enough.

448
00:22:42,600 --> 00:22:46,200
And if you just doing audio you could
that's like four hours of audio

449
00:22:46,200 --> 00:22:46,800
right there.

450
00:22:46,800 --> 00:22:49,360
And if you're doing
four hours of podcasting,

451
00:22:49,360 --> 00:22:51,640
you're not going to have an audience,
right?

452
00:22:51,640 --> 00:22:53,200
Okay. They're gone.

453
00:22:53,200 --> 00:22:53,800
They're asleep.

454
00:22:53,800 --> 00:22:55,560
Yeah. We should know. Yeah.

455
00:22:55,560 --> 00:22:57,720
Yeah. That's that's right.

456
00:22:57,720 --> 00:22:59,160
Yeah. About not having an audience.

457
00:22:59,160 --> 00:23:02,160
I mean, yeah.

458
00:23:02,680 --> 00:23:04,800
I think yeah yeah yeah yeah yeah.

459
00:23:04,800 --> 00:23:07,280
So where do you think this leads Gomez?
Where do you think? What?

460
00:23:07,280 --> 00:23:11,440
What's the next horizon
if content creator focus is added

461
00:23:11,440 --> 00:23:13,280
to what waves is already doing?

462
00:23:13,280 --> 00:23:16,160
I mean, not not specifically,
but in a broader sense.

463
00:23:16,160 --> 00:23:20,160
Do you think as an industry in general,
that focus is going to move

464
00:23:20,160 --> 00:23:21,000
in that direction?

465
00:23:21,000 --> 00:23:22,360
Not really.

466
00:23:22,360 --> 00:23:24,440
I mean, let me give you some perspective.

467
00:23:24,440 --> 00:23:29,960
So we there are,
three product managers, waves

468
00:23:29,960 --> 00:23:34,760
who deal specifically in live sound,
because we have one of the most popular

469
00:23:35,280 --> 00:23:38,160
digital live
consoles in the world with the LP one.

470
00:23:39,640 --> 00:23:41,720
There are, ten

471
00:23:41,720 --> 00:23:46,040
product managers who deal specifically
with the plug in market

472
00:23:46,040 --> 00:23:49,240
and coming up with ideas for plug ins,
developing plug ins, etc.

473
00:23:49,840 --> 00:23:51,360
is that's what a product manager does.

474
00:23:51,360 --> 00:23:56,080
A product manager comes up with
the concept for something, fleshes it out,

475
00:23:56,080 --> 00:24:00,000
and then works with the development team
to all the way through to make it happen.

476
00:24:00,600 --> 00:24:05,960
There's one product manager, waves,
who deals with the content creator market.

477
00:24:05,960 --> 00:24:07,280
That's me.

478
00:24:07,280 --> 00:24:12,680
And for a company who has, you know,
we have about 5 million

479
00:24:12,680 --> 00:24:15,840
plug in users who have purchased
plug ins over the years.

480
00:24:15,840 --> 00:24:20,160
And in our database,
our focus is still very,

481
00:24:20,160 --> 00:24:24,480
very much on looking after, the market
that is our heritage.

482
00:24:24,880 --> 00:24:31,040
However, my goals are set on
how can we take advantage

483
00:24:31,040 --> 00:24:34,080
of the processing
that we have, the skills that we have

484
00:24:34,640 --> 00:24:39,440
and the technology that we have to make
a content creators life easier

485
00:24:39,880 --> 00:24:45,240
so that they can focus on content and
not focus on trying to learn a new skill.

486
00:24:45,840 --> 00:24:46,880
And that does two things.

487
00:24:46,880 --> 00:24:51,560
Number one, it helps us retain
the people in the industry like yourselves

488
00:24:51,560 --> 00:24:56,520
who, voiceovers,
production experts, etc..

489
00:24:57,000 --> 00:25:01,240
And means there are less people out there
on Fiverr saying,

490
00:25:01,240 --> 00:25:06,240
hey, I can to you a radio ad,
but also means that,

491
00:25:06,360 --> 00:25:09,600
the audience gets a better product
regardless.

492
00:25:10,160 --> 00:25:13,200
Is there a user feedback, functionality?

493
00:25:13,200 --> 00:25:17,160
I remember kind of in the earlier days
of this type of technology, I'd

494
00:25:17,160 --> 00:25:22,400
watch YouTube videos
that occasional audio randomness,

495
00:25:23,360 --> 00:25:24,280
you'd be just listening to

496
00:25:24,280 --> 00:25:27,560
the dialog, and then 1 or 2
words were just, you know, this.

497
00:25:27,640 --> 00:25:29,640
Yeah. I mean, yeah, yeah,
it wasn't that way.

498
00:25:29,640 --> 00:25:33,440
So if that, that like it was a discord
or one of those.

499
00:25:33,840 --> 00:25:38,160
So I used to crunch the crap
out of things.

500
00:25:38,520 --> 00:25:39,000
Yeah.

501
00:25:39,000 --> 00:25:42,400
So can you, can you give feedback
like if you get a, if you get bad output,

502
00:25:42,400 --> 00:25:46,320
which I would imagine you've done a lot
to prevent from ever happening.

503
00:25:46,320 --> 00:25:47,480
But can you give feedback.

504
00:25:47,480 --> 00:25:52,440
So so will we with the there was a support
page and a contact form,

505
00:25:52,440 --> 00:25:56,480
and I get a report from Tech Support
literally

506
00:25:56,760 --> 00:26:00,480
six days a week on
anything that comes from them.

507
00:26:00,480 --> 00:26:01,600
And we escalate it.

508
00:26:01,600 --> 00:26:05,880
So if it's something that's really easy
and it's a user issue and 80% at the time,

509
00:26:05,880 --> 00:26:08,320
a problem that comes to us
is a user issue.

510
00:26:08,320 --> 00:26:11,080
So something they didn't quite understand.

511
00:26:11,080 --> 00:26:12,480
That we clear up.

512
00:26:12,480 --> 00:26:15,560
But if there's a real problem with audio,
then I will contact them

513
00:26:15,560 --> 00:26:19,840
directly and say,
are you happy to let me listen

514
00:26:19,840 --> 00:26:23,200
to the audio file or,
and give you an explanation?

515
00:26:23,200 --> 00:26:27,800
Because at this point in time,
we're still in the point where

516
00:26:28,400 --> 00:26:31,560
we want to make sure that everybody
who has the chance

517
00:26:32,160 --> 00:26:36,680
to really get an idea of why they file
didn't work should know why.

518
00:26:36,720 --> 00:26:40,440
And one of the other things
we're about to do is we're about to open

519
00:26:40,440 --> 00:26:43,440
a whole new set of social media channels

520
00:26:43,760 --> 00:26:46,880
specifically for the content creator
market, because we can't

521
00:26:47,360 --> 00:26:52,520
we can't really run tutorials on voice
region on, you know, waves, YouTube.

522
00:26:52,520 --> 00:26:54,400
And then somebody will be scrolling
and they'll go from all

523
00:26:54,400 --> 00:26:55,320
that was really useful.

524
00:26:55,320 --> 00:26:59,080
And then I'll the next one is Andrew Ships
talking about high pass and low

525
00:26:59,080 --> 00:27:00,200
pass and we lose them.

526
00:27:01,160 --> 00:27:02,000
Yeah, exactly.

527
00:27:02,000 --> 00:27:03,880
You know,
we're opening a set of new channels.

528
00:27:03,880 --> 00:27:08,880
We just employed a whole new marketing
department to deal specifically with this,

529
00:27:08,880 --> 00:27:14,160
so that we can actually give content
creators the attention they deserve

530
00:27:14,680 --> 00:27:18,280
without affecting anything else in waves,

531
00:27:18,280 --> 00:27:23,320
or lessening the attention
that we give to our heritage audience.

532
00:27:23,320 --> 00:27:24,400
Did you say discord?

533
00:27:24,400 --> 00:27:27,960
Was that the the discord used
a bit crunch everything to?

534
00:27:27,960 --> 00:27:31,760
Yeah, I was going to say because
it used to do this with also discord

535
00:27:31,760 --> 00:27:35,840
when I put it in the system
that would try to clean up the audio,

536
00:27:36,240 --> 00:27:36,960
it would just like

537
00:27:36,960 --> 00:27:40,560
you're talking about a secret, oh,
discord, discord or sorry descript. Yes.

538
00:27:40,640 --> 00:27:46,400
Oh, descript discord studio
toggle called Sound Studio sounds.

539
00:27:46,440 --> 00:27:48,480
So do you remember
the one that you sent us, George,

540
00:27:48,480 --> 00:27:49,920
whether it was a real estate thing

541
00:27:49,920 --> 00:27:52,320
where it all of sudden went
and started doing gibberish.

542
00:27:52,320 --> 00:27:54,120
So it sounded like I.

543
00:27:54,120 --> 00:27:56,880
What I would like to do is get that
that plug in,

544
00:27:56,880 --> 00:28:00,040
and then I want to get the chef
from Sesame Street and run.

545
00:28:00,040 --> 00:28:01,440
That's really just

546
00:28:02,440 --> 00:28:04,760
a new legacy.

547
00:28:04,760 --> 00:28:08,640
So from what I mean,
obviously I haven't managed to get anybody

548
00:28:08,640 --> 00:28:13,560
from descript.com to talk to me about this
for obvious reasons, probably.

549
00:28:13,560 --> 00:28:17,840
But I have I have actually genuinely
I mean, you guys know me.

550
00:28:17,840 --> 00:28:19,760
I'm pretty genuine an open book.

551
00:28:19,760 --> 00:28:20,640
I have reached out

552
00:28:20,640 --> 00:28:23,960
and said I would love to talk to you
about helping you improve that,

553
00:28:24,400 --> 00:28:27,800
because from my perspective,
what I want to do is I want to make sure

554
00:28:27,800 --> 00:28:30,800
that the content creators
who are using anything

555
00:28:30,920 --> 00:28:33,920
get the best and feel that's really good.

556
00:28:35,400 --> 00:28:36,960
What I feel like

557
00:28:36,960 --> 00:28:40,040
the audio process is doing is,

558
00:28:40,800 --> 00:28:43,200
number one,
I feel like it's giving the users a bit

559
00:28:43,200 --> 00:28:46,880
too much control without explaining
what the toggle is doing,

560
00:28:47,560 --> 00:28:50,880
but when it's on full,
which I think most people just put it on

561
00:28:50,880 --> 00:28:55,240
to full without really,
I think it bit crunch is to a point

562
00:28:55,240 --> 00:28:58,440
where, as you said,
some words just disappear.

563
00:28:59,240 --> 00:29:02,760
Now, one of the interesting things
about this, when I was investigating

564
00:29:02,760 --> 00:29:06,200
this kind of problem was
I found out around the same time

565
00:29:06,480 --> 00:29:09,360
that I created video.

566
00:29:09,360 --> 00:29:12,360
I generated video with dialog.

567
00:29:12,600 --> 00:29:15,720
The dialog with most video models

568
00:29:15,720 --> 00:29:19,520
comes out in the codec as 96 K,

569
00:29:20,120 --> 00:29:24,000
but is a bit crunched to Helen back.

570
00:29:24,960 --> 00:29:28,200
If you drag that video onto voice region,

571
00:29:28,400 --> 00:29:32,400
it makes it sound like they're
talking into a boom mike or a lapel.

572
00:29:33,080 --> 00:29:35,440
So suddenly the AI video

573
00:29:35,440 --> 00:29:38,440
that was given away by the audio,

574
00:29:38,520 --> 00:29:41,880
irritatingly to some people,
really impressive to others,

575
00:29:42,040 --> 00:29:45,240
now looks really realistic
because the audio sounds realistic.

576
00:29:45,240 --> 00:29:50,400
I think all AI videos should always
put Waldo somewhere. Yes.

577
00:29:53,880 --> 00:29:55,240
You mean the Waldo watermark?

578
00:29:55,240 --> 00:29:58,320
Yeah, that that's how you know it's
AI because Waldo

579
00:29:58,680 --> 00:30:03,040
is somewhere in the video
semi is saying is you brought up Waldo.

580
00:30:03,360 --> 00:30:06,440
I'll share something which I share
with the team every week at Waves.

581
00:30:06,800 --> 00:30:09,960
Once a week
when we have our, engineering meeting,

582
00:30:10,440 --> 00:30:14,960
I create a new Where's Waldo with
but with me in it.

583
00:30:15,400 --> 00:30:18,400
Using Google's now open on a pro,

584
00:30:18,840 --> 00:30:24,000
and I share it in the teams chat
every week and the product managers like,

585
00:30:24,000 --> 00:30:27,640
okay, where's MPX, where's MPX,
where's it put little.

586
00:30:30,160 --> 00:30:33,520
It's funny, I wear pajamas that, stripy.

587
00:30:34,000 --> 00:30:37,120
I wore I, we were bananas in pajamas.

588
00:30:37,520 --> 00:30:39,960
Did this bananas did that.
Does it make here?

589
00:30:39,960 --> 00:30:41,360
We were in a meeting.

590
00:30:41,360 --> 00:30:43,080
We were in a in the today.

591
00:30:43,080 --> 00:30:45,240
And I was wearing a collared shirt and,

592
00:30:45,240 --> 00:30:48,520
and, and I had three people say,
you look like you're a prisoner.

593
00:30:49,200 --> 00:30:50,680
And it's like I might.

594
00:30:52,080 --> 00:30:54,960
So because we all are prisoners.

595
00:30:54,960 --> 00:30:58,640
Well, for me, for me as a user of waves,
the hardest thing for me to wrap

596
00:30:58,640 --> 00:31:01,640
my head around when I hadn't looked at it
yet was just understanding that,

597
00:31:01,640 --> 00:31:04,600
well, that's, you know,
so that's like number one, it's like

598
00:31:04,600 --> 00:31:07,760
you have to get out
of the ecosystem of Waves Central

599
00:31:08,200 --> 00:31:13,080
when your brain clicks and goes,
oh yeah, first it's the web app.

600
00:31:13,080 --> 00:31:19,200
Okay, I used video I did on voice
region was to the large waves market

601
00:31:19,600 --> 00:31:23,760
saying and it literally is all about,
hey, we've just released this thing.

602
00:31:24,000 --> 00:31:26,880
Just so you know, it's
not a plug in, don't it?

603
00:31:26,880 --> 00:31:28,920
That's kind of the idea of the video.

604
00:31:28,920 --> 00:31:33,240
It's it's not, you know,
we're not we're not degrading or saying,

605
00:31:33,240 --> 00:31:35,280
you know,
I uses couldn't understand stuff,

606
00:31:35,280 --> 00:31:38,400
but we wanted to make it clear
that for the first time ever, we're

607
00:31:38,400 --> 00:31:41,400
releasing an audio process
that isn't a plug in.

608
00:31:41,880 --> 00:31:45,360
You know, it's like,
I wouldn't have wanted to release this

609
00:31:45,360 --> 00:31:48,880
without letting the bigger
audience know first were about to.

610
00:31:49,000 --> 00:31:51,840
Is it also mobile friendly?
I didn't forget, oh, absolutely.

611
00:31:51,840 --> 00:31:52,400
To the point.

612
00:31:52,400 --> 00:31:57,920
The in the next push in the next few days,
we're adding a little banner that says,

613
00:31:57,920 --> 00:32:01,320
hey, do you want to kind of shortcut this
so it looks like an app on your phone?

614
00:32:01,640 --> 00:32:04,000
Teleprompter works on the phone.

615
00:32:04,000 --> 00:32:08,040
You can audio record,
to a script on your phone.

616
00:32:08,040 --> 00:32:09,240
You can drag a video

617
00:32:09,240 --> 00:32:12,240
or an audio file from your phone
straight in and press process.

618
00:32:12,680 --> 00:32:15,880
I'm not letting people on mobile
phone record video

619
00:32:16,040 --> 00:32:17,160
when we release that,

620
00:32:17,160 --> 00:32:20,760
because a mobile phone is perfectly good
at recording video by itself

621
00:32:21,280 --> 00:32:22,080
and you're holding it.

622
00:32:22,080 --> 00:32:24,720
Just yeah,
record the video with your standard app.

623
00:32:24,720 --> 00:32:27,320
Upload that. That's easy.
So here's a question.

624
00:32:27,320 --> 00:32:32,760
Then this just occurred to me
if I were a voiceover artist

625
00:32:33,920 --> 00:32:36,920
and I was putting together my travel rig

626
00:32:37,800 --> 00:32:41,760
and I had a interface like,

627
00:32:42,880 --> 00:32:46,640
you know, maybe something that some
certain podcast may have put together.

628
00:32:46,840 --> 00:32:48,280
Yeah, sold for a while.

629
00:32:48,280 --> 00:32:51,880
But I needed something to record onto.

630
00:32:52,560 --> 00:32:54,240
Could I use it just as a recorder?

631
00:32:54,240 --> 00:32:56,720
Like, I guess the question is,
do I have to process?

632
00:32:56,720 --> 00:32:58,560
Can I record on to there
and then just down?

633
00:32:58,560 --> 00:32:59,480
You don't have to process.

634
00:32:59,480 --> 00:33:01,320
You can just do it that way
if you want to. Yeah.

635
00:33:01,320 --> 00:33:05,360
So so so does it burn your hours
if you use that that teleprompter

636
00:33:05,360 --> 00:33:09,240
and the recorder
or is that all without using the hours

637
00:33:09,240 --> 00:33:11,480
that the, you know, the,
the only thing the,

638
00:33:11,480 --> 00:33:15,880
the only thing that burns you in
minutes is the processing processing.

639
00:33:15,880 --> 00:33:17,480
So that's why I recorded a call.

640
00:33:17,480 --> 00:33:20,040
It's a teleprompter and a recorder
right there. Yeah.

641
00:33:20,040 --> 00:33:22,640
And that's honestly
and it's actually intentional.

642
00:33:22,640 --> 00:33:25,880
And one of the reasons why I did
that is because

643
00:33:26,400 --> 00:33:29,160
of my heritage and my background from,

644
00:33:29,160 --> 00:33:32,440
you know, not just 20 years in music
production in the States, but,

645
00:33:32,640 --> 00:33:35,640
you know, 15
before that radio production in Australia.

646
00:33:36,200 --> 00:33:38,080
One of the things
that when you've got headphones

647
00:33:38,080 --> 00:33:40,480
on, you've recorded a script
and you listen back, it's

648
00:33:40,480 --> 00:33:44,960
when you listen back that you realize,
yep, sounds good, but I'm in a hotel room.

649
00:33:44,960 --> 00:33:47,320
You know what? I'm
just going to click process in the room.

650
00:33:47,320 --> 00:33:48,880
That's exactly right there.

651
00:33:48,880 --> 00:33:51,120
I love that
you integrated the script reader.

652
00:33:51,120 --> 00:33:53,760
That was first. I was like, oh, I had to.

653
00:33:53,760 --> 00:33:57,720
I mean, it's like, I mean,
I've got two users who are on pro accounts

654
00:33:57,720 --> 00:34:01,400
because they
they are professional audiobook readers

655
00:34:01,840 --> 00:34:06,280
and they love me to death
because I got them to beta test this.

656
00:34:06,280 --> 00:34:10,760
And they were like, wait,
so I can have literally chapters in this?

657
00:34:11,080 --> 00:34:14,120
I just have it
right there in front of me,

658
00:34:14,120 --> 00:34:17,120
scrolling really short
and just do it at my own speed.

659
00:34:17,280 --> 00:34:19,920
It's like, yeah, and just pause
if you want to.

660
00:34:19,920 --> 00:34:22,560
This is going to be controversial,
but I guarantee you could

661
00:34:22,560 --> 00:34:25,800
literally just record an entire audiobook
on your phone full stop.

662
00:34:25,800 --> 00:34:29,880
If you if you give this really good audio,
like really good audio,

663
00:34:29,880 --> 00:34:32,200
will it still touch it
or will it go, that's good.

664
00:34:32,200 --> 00:34:33,160
Interesting question.

665
00:34:33,160 --> 00:34:38,600
I'm so glad you asked it now because of
the fact that this is a very large model

666
00:34:38,960 --> 00:34:42,600
that's trained on high quality
files, it's trained on

667
00:34:43,000 --> 00:34:47,160
not the finest quality,
it's trained on quality files.

668
00:34:47,520 --> 00:34:53,120
So when somebody from a studio puts
in a file that's recorded in, you know, a

669
00:34:53,120 --> 00:34:57,480
Neumann and, in a soundproofed

670
00:34:57,480 --> 00:35:00,680
room, acute compressor preamp.

671
00:35:00,680 --> 00:35:04,080
And exactly if you put that into it,
it will degrade the file.

672
00:35:04,320 --> 00:35:06,400
Interesting enough. Yeah. Okay.

673
00:35:06,400 --> 00:35:10,080
That makes sense, because what it does
is it's still trying to go, okay,

674
00:35:10,080 --> 00:35:13,800
I have a quality,
but I'm trying to get this file to this.

675
00:35:13,800 --> 00:35:16,640
I don't know what to do with this,
so it's going to degrade it for you.

676
00:35:16,640 --> 00:35:19,280
But only if you hit process. Right.

677
00:35:19,280 --> 00:35:19,960
You can still read.

678
00:35:19,960 --> 00:35:22,600
Of course. Yes, I understand, yeah.

679
00:35:22,600 --> 00:35:25,920
At some point you decide
this model is baked.

680
00:35:26,120 --> 00:35:28,800
Let's build a new model and start
all over. We're on.

681
00:35:28,800 --> 00:35:30,800
We're on the third model. Yeah.

682
00:35:30,800 --> 00:35:33,600
Because as I said earlier

683
00:35:33,600 --> 00:35:38,120
on, we're also about to, enable
this is an API

684
00:35:38,120 --> 00:35:41,400
that people can use as a feature
to put into their own website.

685
00:35:41,400 --> 00:35:43,360
And charge four or whatever.

686
00:35:43,360 --> 00:35:47,120
So what this will end up being is,

687
00:35:47,520 --> 00:35:50,920
they'll be the initial model

688
00:35:50,920 --> 00:35:54,040
that will end up being kind of
like one of the most affordable ones.

689
00:35:54,040 --> 00:35:57,360
And then as we build

690
00:35:57,360 --> 00:36:01,240
new models
that will be more dynamic or brighter

691
00:36:01,240 --> 00:36:05,920
or this then will we'll decide
how we want to move forward

692
00:36:05,920 --> 00:36:09,560
with, the, the consumer app,
the one that we just released.

693
00:36:09,600 --> 00:36:13,400
You get old model, it says this is a voice
that sounds like it was recorded

694
00:36:13,400 --> 00:36:14,920
with the UAD seven.

695
00:36:14,920 --> 00:36:16,120
Yeah. I don't want to do that. Yeah.

696
00:36:16,120 --> 00:36:20,160
There's a couple of reasons
I don't want to go that exact number

697
00:36:20,160 --> 00:36:22,080
one, it's a huge claim.

698
00:36:22,080 --> 00:36:26,320
Yeah, but number two,
one of the things that happens in

699
00:36:26,520 --> 00:36:30,560
if you're creating a tool
that content creators use is

700
00:36:30,800 --> 00:36:35,680
you run a risk of giving people who

701
00:36:36,840 --> 00:36:39,720
are focusing on reporting their content
and creating content

702
00:36:39,720 --> 00:36:42,480
and don't want to be an audio engineer
or a video editor,

703
00:36:42,480 --> 00:36:43,960
you're asking them to make

704
00:36:43,960 --> 00:36:47,960
too many decisions without the experience
knowledge to back it up.

705
00:36:48,320 --> 00:36:51,320
And that becomes what I call
feature creep.

706
00:36:51,320 --> 00:36:52,120
Yeah.

707
00:36:52,120 --> 00:36:56,320
You're asking them to make a decision
on control, on something

708
00:36:56,720 --> 00:37:00,320
that, they may not be able to hear.

709
00:37:00,320 --> 00:37:05,160
Well, you know, even that is like
but understanding what you're hearing,

710
00:37:05,160 --> 00:37:09,840
whereas my goal is still I don't want
to give you a hammer and a nail.

711
00:37:09,840 --> 00:37:12,240
I want to give you a nail
that's in a wall.

712
00:37:12,240 --> 00:37:16,160
I would imagine 99% of content
creators would ask, what is a UAD set?

713
00:37:16,400 --> 00:37:17,360
Yeah, exactly.

714
00:37:17,360 --> 00:37:17,560
Yeah.

715
00:37:17,560 --> 00:37:20,880
If you take a look at what
a lot of there's been a lot of,

716
00:37:22,040 --> 00:37:25,040
experience over the
years with different pro audio companies

717
00:37:25,480 --> 00:37:28,480
entering the content creator market
in one way or another.

718
00:37:28,560 --> 00:37:30,000
And I've studied most of them.

719
00:37:30,000 --> 00:37:33,560
And I'm not saying I or
we are perfect in any way, but

720
00:37:33,560 --> 00:37:37,360
one of the things I've learned from others
mistakes is, number one,

721
00:37:37,560 --> 00:37:41,360
never use pro audio nomenclature
when you're talking to a market

722
00:37:41,360 --> 00:37:44,880
that could be a wedding event planner
doing a vlog.

723
00:37:46,480 --> 00:37:48,920
At what point in her life or his life?

724
00:37:48,920 --> 00:37:52,760
If they come across a high
pass, low pass filter

725
00:37:52,760 --> 00:37:56,480
and the a two and 2 to 1 compression
ratio,

726
00:37:56,960 --> 00:38:00,200
a it's, you know,

727
00:38:00,360 --> 00:38:04,680
even the term compression or EQ, you just
I want to stay away from them.

728
00:38:05,200 --> 00:38:06,080
Well, even in clarity,

729
00:38:06,080 --> 00:38:10,640
if you have I think I haven't checked
in the last two months, but three models.

730
00:38:10,640 --> 00:38:11,760
Right.

731
00:38:11,760 --> 00:38:12,480
That's a user. Yeah.

732
00:38:12,480 --> 00:38:16,160
And there will be more,
you know, the clarity

733
00:38:16,440 --> 00:38:19,200
development team, and it's a team.

734
00:38:19,200 --> 00:38:23,600
The they're constantly at work
improving, tweaking.

735
00:38:24,360 --> 00:38:27,680
And obviously there has been some,
you know, questions as well

736
00:38:27,680 --> 00:38:28,560
out in the marketplace.

737
00:38:28,560 --> 00:38:30,840
It's like,
well, you made clarity a plug in.

738
00:38:32,040 --> 00:38:34,520
Is region just a website of clarity?

739
00:38:34,520 --> 00:38:36,440
No. Completely different models,

740
00:38:36,440 --> 00:38:38,920
completely different models,
completely different processes.

741
00:38:38,920 --> 00:38:41,920
Clarity can be a plug in region

742
00:38:42,000 --> 00:38:46,120
will never be a plug in
because it will melt a different products.

743
00:38:46,120 --> 00:38:51,680
I mean clarity 100 is a is is not there
to change the character of the sound.

744
00:38:51,680 --> 00:38:54,840
It's there to isolate the voice
in the character that's given to it.

745
00:38:54,840 --> 00:38:55,080
Right?

746
00:38:55,080 --> 00:38:58,080
Whereas we literally regenerate a voice.

747
00:38:58,440 --> 00:39:01,840
You change the character, you're like,
re wouldn't say

748
00:39:02,040 --> 00:39:05,640
we're not really changing the character,
which was the

749
00:39:05,800 --> 00:39:11,200
the Sonicare, like like this, this, this
was originally recorded on, whatever.

750
00:39:11,200 --> 00:39:14,200
Like a, like a carbon
microphone wire recording.

751
00:39:14,320 --> 00:39:14,640
Yeah.

752
00:39:14,640 --> 00:39:17,280
And now it sounds like it's
coming out of a SM seven.

753
00:39:17,280 --> 00:39:18,360
Yeah. Yeah.

754
00:39:18,360 --> 00:39:20,600
You know,
like the character of the recording.

755
00:39:20,600 --> 00:39:22,280
Not not the person.

756
00:39:22,280 --> 00:39:24,560
Would rebuilding it
be a better way to say it?

757
00:39:24,560 --> 00:39:29,680
We're renovating, renovate,
kind of regenerating in a lot of case,

758
00:39:29,680 --> 00:39:33,480
because it's like we're
we're kind of looking at the in a way,

759
00:39:33,480 --> 00:39:38,800
I guess the DNA of vocal tone, and going,
okay, what?

760
00:39:39,360 --> 00:39:41,640
Let's put some work in and improve this.

761
00:39:41,640 --> 00:39:45,680
So in the end, you know, how well
you know, how you go to various websites.

762
00:39:45,680 --> 00:39:46,440
We won't name them.

763
00:39:46,440 --> 00:39:50,320
And then you've put your voice in there
and there's the whole idea of a one shot.

764
00:39:50,400 --> 00:39:51,840
Oh, the voice model.

765
00:39:51,840 --> 00:39:54,160
Yeah. No we don't do that. Cloning. Yeah.

766
00:39:54,160 --> 00:39:55,640
So this is not a clone.

767
00:39:55,640 --> 00:39:58,640
Like, just to be clear, it's not learning.

768
00:39:58,640 --> 00:40:02,400
This person's voice
the that someone uploads to it,

769
00:40:02,520 --> 00:40:04,800
and and those people
have nothing to worry about.

770
00:40:04,800 --> 00:40:07,160
Like, oh, now
you've uploaded my voice to some website.

771
00:40:07,160 --> 00:40:08,800
I don't want my voice uploaded to.

772
00:40:08,800 --> 00:40:10,680
No, we don't we don't retain.

773
00:40:10,680 --> 00:40:14,520
It's like we it goes through our process.

774
00:40:14,520 --> 00:40:15,920
It goes out the other end.

775
00:40:15,920 --> 00:40:18,200
It doesn't even stay in our cloud.

776
00:40:18,200 --> 00:40:20,640
It goes back to your dashboard

777
00:40:20,640 --> 00:40:23,640
and after 14 days,
we just delete them from the system.

778
00:40:24,200 --> 00:40:25,480
You know, I had one

779
00:40:25,480 --> 00:40:29,440
case named a year ago because I know
everything's happening so fast.

780
00:40:29,440 --> 00:40:32,440
So whenever I say I tried this thing,

781
00:40:32,840 --> 00:40:36,320
you sort of have to say when you know,
because, like, everything's changing.

782
00:40:36,920 --> 00:40:39,840
But the client had sent
me some audio from, like,

783
00:40:40,800 --> 00:40:41,360
Catholic

784
00:40:41,360 --> 00:40:45,920
priest or something, and it was all this
archival audio and it was recorded

785
00:40:45,920 --> 00:40:50,040
very on a bad technology,
probably like a 12

786
00:40:50,040 --> 00:40:54,280
bit 32kHz or worse recorder

787
00:40:54,280 --> 00:40:58,360
that was built in to the,
I don't know, something pretty bad.

788
00:40:59,240 --> 00:41:02,680
And they sent it to me
and the audio was pretty bad.

789
00:41:02,680 --> 00:41:06,960
And then I threw it into one of these
AI tools that regenerate, essentially.

790
00:41:07,480 --> 00:41:10,800
And when she listened to it back, she was
she was kind of

791
00:41:11,160 --> 00:41:17,080
I think she didn't know what to expect
because it was a reconstruction.

792
00:41:17,080 --> 00:41:17,560
Right.

793
00:41:17,560 --> 00:41:21,520
And so she thought
it didn't sound authentic.

794
00:41:22,120 --> 00:41:25,880
So then I was like, well,
I don't know what to tell you.

795
00:41:25,880 --> 00:41:28,880
This is probably what it did sound like,

796
00:41:29,520 --> 00:41:34,440
but you've heard it for the last 30 years
sounding like it comes through a tin can.

797
00:41:35,240 --> 00:41:38,640
You know, out of us
it's the it's exactly the same thing

798
00:41:38,640 --> 00:41:42,560
that happens with,

799
00:41:42,560 --> 00:41:46,440
I, I was talking to the guy, producer,
composer guy

800
00:41:46,440 --> 00:41:50,640
called Greg Wells, a few months ago,
and we were talking about,

801
00:41:51,720 --> 00:41:54,720
a soundtrack that he's just composed.

802
00:41:55,560 --> 00:41:59,000
And one of the biggest problems
with composing a soundtrack

803
00:41:59,280 --> 00:42:02,320
and making the film director
and the company happy with

804
00:42:02,320 --> 00:42:05,720
it is when they're actually
doing the edit of the film.

805
00:42:05,720 --> 00:42:10,160
They use a temp soundtrack,
which is usually an existing soundtrack

806
00:42:10,160 --> 00:42:14,000
from something else that kind of fits,
but they spend so many months

807
00:42:14,000 --> 00:42:16,720
on the cutting room floor
with the temp soundtrack.

808
00:42:16,720 --> 00:42:17,200
No idea.

809
00:42:17,200 --> 00:42:20,320
So when the new soundtrack
that's being composed specifically

810
00:42:20,320 --> 00:42:23,320
for it comes up, a lot of the time
they're like, oh, that sucks.

811
00:42:23,760 --> 00:42:28,800
So yeah, I think the funniest example
of this is, is, some clients of mine

812
00:42:28,800 --> 00:42:33,560
were talking about a session and basically
the scratch track was done with an AI,

813
00:42:34,080 --> 00:42:37,080
and then they were
recording the real voice,

814
00:42:37,520 --> 00:42:41,160
and they began asking the voice talent
to read it more like the scratch track,

815
00:42:41,440 --> 00:42:44,440
which was basically saying,
please read it more like I

816
00:42:44,520 --> 00:42:47,520
yeah, it's like, yeah, like

817
00:42:47,520 --> 00:42:51,520
how to lose your soul in five seconds
and yeah,

818
00:42:51,560 --> 00:42:54,720
yeah,
I had that exact thing happen to me with a

819
00:42:54,720 --> 00:42:56,760
we talked about the session before,

820
00:42:56,760 --> 00:43:00,960
whether it was a female reading to this
short film that they wanted me to voice.

821
00:43:00,960 --> 00:43:03,840
And I was kind of thinking,
that sounds okay. Anyway,

822
00:43:04,880 --> 00:43:06,600
but they said, oh, no, no,

823
00:43:06,600 --> 00:43:09,560
the check out the three words
are not pronounced correctly.

824
00:43:09,560 --> 00:43:13,640
And I'm thinking, well, why don't
you just get a back to redo those lines?

825
00:43:13,800 --> 00:43:16,800
But it wasn't until later
I realized it was an AI,

826
00:43:17,120 --> 00:43:20,880
but they they were trying to get me like
the inflection on the on the guide track.

827
00:43:20,880 --> 00:43:21,720
Can you try and do that?

828
00:43:21,720 --> 00:43:25,080
And it wasn't until the next day,
I think, and I mentioned this before,

829
00:43:25,080 --> 00:43:29,040
another episode that I was playing it
to, to my son, who's 17.

830
00:43:29,040 --> 00:43:30,760
I said, oh, you know,
this is what happened, boy,

831
00:43:30,760 --> 00:43:34,200
because that's I, I'm like, yeah,
I can hear it a mile away.

832
00:43:34,200 --> 00:43:36,200
Usually two they can hear it a mile away.

833
00:43:36,200 --> 00:43:41,360
The can I, I'm, I'm 56, but I can hear it
a mile away because I've spent so many,

834
00:43:41,520 --> 00:43:46,080
it spent the last few years
literally deep in AI research.

835
00:43:46,680 --> 00:43:51,200
And one of the things that's really hard,
even if you're using something is good,

836
00:43:51,200 --> 00:43:52,920
is 11 laps.

837
00:43:52,920 --> 00:43:59,040
You have to make sure that words,
especially in English words

838
00:43:59,040 --> 00:44:02,040
that, the same.

839
00:44:02,440 --> 00:44:06,920
So so for example,
one, or one, they're so windy.

840
00:44:08,120 --> 00:44:10,040
Oh. And

841
00:44:10,040 --> 00:44:13,080
you have to work
out a way of spelling them phonetically

842
00:44:13,080 --> 00:44:16,040
so that they, separated.

843
00:44:16,040 --> 00:44:19,200
And there was so many other words
in the English language that I just

844
00:44:19,200 --> 00:44:23,640
does not get its perceptive head around
because it doesn't have a perception

845
00:44:23,920 --> 00:44:27,920
version of the context in between
this word, this word and this word,

846
00:44:28,320 --> 00:44:32,040
when it's spelt one way
but it's pronounced another.

847
00:44:32,400 --> 00:44:36,040
So I remember doing this
with a phone in chips in college.

848
00:44:36,040 --> 00:44:39,200
And just to get this thing to say
electronics,

849
00:44:39,680 --> 00:44:43,640
you had to put like ten ese in a row
because it had to do like electronics.

850
00:44:43,640 --> 00:44:47,400
And if you wanted to say electronics,
you had to put like IEEE.

851
00:44:47,400 --> 00:44:50,400
L like electronics. Yeah.

852
00:44:50,520 --> 00:44:52,240
Because it just yeah.

853
00:44:52,240 --> 00:44:54,840
If to like force it
to get out what you want

854
00:44:54,840 --> 00:44:58,240
because it's not always changing
but it's also a dialect.

855
00:44:58,240 --> 00:45:02,160
I mean, you know, I,
I've just heard a couple of things like,

856
00:45:02,160 --> 00:45:07,520
I heard Gomez say data instead of data,
but it spelt the same.

857
00:45:07,520 --> 00:45:10,280
But it's pronounced differently,
like dance.

858
00:45:10,280 --> 00:45:11,120
Dance, you know.

859
00:45:11,120 --> 00:45:14,120
Yeah. 2020 is an American man.

860
00:45:14,280 --> 00:45:16,680
Yeah, yeah yeah, yeah. Tomato, tomato.

861
00:45:16,680 --> 00:45:18,400
Yeah, exactly.

862
00:45:18,400 --> 00:45:21,040
It's like when you get, someone to say it

863
00:45:21,040 --> 00:45:22,720
like English is not their first language.

864
00:45:22,720 --> 00:45:24,880
And then they're selling
their writing on the script.

865
00:45:24,880 --> 00:45:28,600
Their phonetic version
of how you pronounce that word.

866
00:45:29,120 --> 00:45:31,680
But they forget that the fact is that

867
00:45:31,680 --> 00:45:33,560
their phonetics are going
to be different to yours

868
00:45:33,560 --> 00:45:35,320
because they say certain things
differently.

869
00:45:35,320 --> 00:45:36,200
The way you would say it.

870
00:45:36,200 --> 00:45:40,960
So when you read their phonetic version
of that word, it's completely wrong.

871
00:45:40,960 --> 00:45:42,960
And they come back and go, no, no, no,
that's not what I want, I want it,

872
00:45:42,960 --> 00:45:43,840
blah, blah, blah.

873
00:45:43,840 --> 00:45:45,080
It's like, well, that's what you just did.

874
00:45:45,080 --> 00:45:47,280
But, you know, because they will
pronounce it differently.

875
00:45:47,280 --> 00:45:50,560
So now this brings up something
which I think is interesting.

876
00:45:50,560 --> 00:45:55,440
I find it fascinating in
when when building the model for region

877
00:45:55,920 --> 00:45:56,760
was happening.

878
00:45:56,760 --> 00:46:01,360
The first model is a term that we realized

879
00:46:01,360 --> 00:46:06,000
early on that
we had to rectify, which is accent bias.

880
00:46:06,960 --> 00:46:07,880
If a model

881
00:46:07,880 --> 00:46:12,200
is educated
on too many of a specific accents,

882
00:46:12,200 --> 00:46:15,360
then it doesn't understand how to deal

883
00:46:15,360 --> 00:46:18,640
with other pronunciations or accents.

884
00:46:18,640 --> 00:46:21,640
So initially, early on,

885
00:46:22,040 --> 00:46:26,440
an Australian accent was an anomaly,
and some of the first tests we did

886
00:46:26,800 --> 00:46:30,080
was frustrating
because the Americans would sound amazing,

887
00:46:30,320 --> 00:46:33,000
the Brits would sound amazing,
but any time I tried it,

888
00:46:33,000 --> 00:46:36,000
it would sound glitchy as hell
and we worked out.

889
00:46:36,400 --> 00:46:40,200
It just doesn't understand
this half drunk way we talk.

890
00:46:40,560 --> 00:46:43,920
Hey, careful when

891
00:46:45,200 --> 00:46:46,240
we're not half drunk.

892
00:46:46,240 --> 00:46:49,480
It's like we tend to kind
of take shortcuts with the way we speak.

893
00:46:50,040 --> 00:46:55,520
So we had to kind of make sure that we
spent a lot of time avoiding accent bias.

894
00:46:55,520 --> 00:46:59,840
So this is only for English
at this English language.

895
00:46:59,840 --> 00:47:04,360
At this point no works with Hebrew, works
with French, German, Spanish, Japanese.

896
00:47:05,240 --> 00:47:06,800
But it is language specific.

897
00:47:06,800 --> 00:47:07,440
Basically.

898
00:47:07,440 --> 00:47:10,600
No, it's human voice specific.

899
00:47:10,600 --> 00:47:12,240
Yes. It's to understand it.

900
00:47:12,240 --> 00:47:15,240
So so if like does it work with pig Latin?

901
00:47:15,920 --> 00:47:16,640
I have no idea.

902
00:47:16,640 --> 00:47:20,480
I, I it's not high on my list
of priorities, to be honest.

903
00:47:21,440 --> 00:47:22,760
Not trained on pig Latin.

904
00:47:22,760 --> 00:47:25,040
Yet Robert Young recorded this thing.

905
00:47:25,040 --> 00:47:29,040
Like, if it's like some unknown language
that it's never, ever heard before,

906
00:47:29,400 --> 00:47:32,920
here's here's the key,
here's the key with voice region.

907
00:47:33,360 --> 00:47:35,400
If you're putting a file into voice
region,

908
00:47:35,400 --> 00:47:40,640
any in the language that you speak,
you can understand the words

909
00:47:40,640 --> 00:47:43,200
that are coming
out of the speaker's mouth.

910
00:47:43,200 --> 00:47:46,440
Regardless of the noise level,
we have a chance

911
00:47:46,440 --> 00:47:48,720
of saving that file for you.

912
00:47:48,720 --> 00:47:53,880
If there are words in that file
that you cannot comprehend yourself,

913
00:47:53,880 --> 00:47:57,400
then our chance goes down by about 80%.

914
00:47:57,640 --> 00:47:58,280
Okay, yeah.

915
00:47:58,280 --> 00:48:01,200
Is this is this forensics proof?

916
00:48:01,200 --> 00:48:05,400
Like, like, could someone go to court
and be like, here's this awful recording.

917
00:48:05,400 --> 00:48:07,920
But look, he said I murdered her.

918
00:48:07,920 --> 00:48:11,160
And then voice regions like, perfect,
like their cat.

919
00:48:11,160 --> 00:48:12,480
You hear that? 100%.

920
00:48:12,480 --> 00:48:14,400
And it's like this. Yeah, absolutely.

921
00:48:14,400 --> 00:48:16,080
Yeah. We're not making up. We're not.

922
00:48:16,080 --> 00:48:19,080
There's no point in the system
where we make stuff up.

923
00:48:19,280 --> 00:48:21,680
We regenerate what the system understands.

924
00:48:21,680 --> 00:48:25,480
So a word that is really garbled
because there was a technical,

925
00:48:26,720 --> 00:48:27,280
a word that

926
00:48:27,280 --> 00:48:30,280
is really garbled is still going
to stay really garbled.

927
00:48:30,440 --> 00:48:31,560
Yeah, yeah.

928
00:48:31,560 --> 00:48:34,840
So one of the early cases
and because, I mean, one of the things

929
00:48:34,840 --> 00:48:38,920
which we did is we spent a lot of time
throwing files at it

930
00:48:38,920 --> 00:48:42,680
that were really, really challenging,
like two people

931
00:48:42,800 --> 00:48:48,000
filmed on an iPhone, but they're 30ft away
on the edge of a cliff with wind.

932
00:48:48,400 --> 00:48:50,680
Once talking, one's further away.

933
00:48:50,680 --> 00:48:53,000
She's yelling back at him with wind.

934
00:48:53,000 --> 00:48:53,880
Yeah, you've got the wind.

935
00:48:53,880 --> 00:48:56,880
And we learned very quickly the

936
00:48:56,960 --> 00:48:59,440
if you're listening to a file
and you're listening to the audio

937
00:48:59,440 --> 00:49:02,040
and you can't understand
what they're saying,

938
00:49:02,040 --> 00:49:05,560
then our chance of getting
there is a lot lower.

939
00:49:05,560 --> 00:49:06,800
It's not every case.

940
00:49:06,800 --> 00:49:09,240
There are some. There we go. Oh,
that's what she said.

941
00:49:09,240 --> 00:49:12,000
But most of the time it's at that point

942
00:49:12,000 --> 00:49:16,800
where the audio dialog frequencies
are so mixed down

943
00:49:16,800 --> 00:49:20,680
into the noise frequencies
that we can't regenerate

944
00:49:20,680 --> 00:49:23,480
because we have nothing there
to work with.

945
00:49:23,480 --> 00:49:27,480
So what it does not do,
it does not interpolate.

946
00:49:27,960 --> 00:49:30,720
No, it's not using context.

947
00:49:30,720 --> 00:49:34,000
It's not, it's not,
it's not, it's not there's no kind of

948
00:49:34,040 --> 00:49:37,040
decoding the language
and re encoding the language.

949
00:49:37,120 --> 00:49:41,200
It's just decoding the voice,
the human voice in there and then

950
00:49:41,200 --> 00:49:45,080
performing the voice in the examples
I have on the landing page on the website.

951
00:49:45,080 --> 00:49:46,920
A very intentional.

952
00:49:46,920 --> 00:49:52,000
They're all bad quality,
but they're also all very if you listen

953
00:49:52,000 --> 00:49:55,000
to them, you can understand the words,
but you're like, oh, that's terrible.

954
00:49:55,120 --> 00:49:56,240
Yeah.

955
00:49:56,240 --> 00:49:58,960
The and the reason for
that is exactly what I just said.

956
00:49:58,960 --> 00:50:01,960
It's like there is no point in the system
where,

957
00:50:02,040 --> 00:50:05,200
we just guess because that's,

958
00:50:05,920 --> 00:50:09,800
you know, we already have enough of that,
which in other AI systems,

959
00:50:09,800 --> 00:50:13,200
it's like what I did was we
we created a model

960
00:50:14,120 --> 00:50:17,520
that can bring back
nearly irretrievable audio.

961
00:50:17,920 --> 00:50:20,960
But it's also not going to happen
in every single case.

962
00:50:20,960 --> 00:50:25,040
There are some cases, obviously,
where people put in audio and go, yeah,

963
00:50:25,040 --> 00:50:25,640
it's, you know,

964
00:50:26,760 --> 00:50:27,320
so is a

965
00:50:27,320 --> 00:50:32,240
future possible pro version or an advanced
version would be to literally regenerate

966
00:50:32,240 --> 00:50:36,280
a word that was totally garbled
so the user could type in the word foxy

967
00:50:36,280 --> 00:50:39,280
and they will regenerate the word
foxy from that.

968
00:50:39,560 --> 00:50:41,400
It's been it it's it's been discussed.

969
00:50:41,400 --> 00:50:45,280
One of the things that comes up with this,
that is a big problem for me,

970
00:50:45,840 --> 00:50:48,360
because as a product manager,
one of the things that you're thinking

971
00:50:48,360 --> 00:50:49,920
isn't is something possible.

972
00:50:49,920 --> 00:50:52,920
You're thinking, is this possible in a way

973
00:50:53,280 --> 00:50:58,200
that the result will be something
that the user will take for

974
00:50:58,200 --> 00:51:01,560
granted, is something that happened
and can move on with their life

975
00:51:01,840 --> 00:51:03,360
without worrying about it?

976
00:51:03,360 --> 00:51:05,840
And my answer at this point in
time is still no,

977
00:51:05,840 --> 00:51:08,360
because what you're doing
is you're saying, okay, right.

978
00:51:08,360 --> 00:51:12,320
So we have to, first of all, clone
a certain amount of the voice

979
00:51:12,320 --> 00:51:15,480
in between, around the word
that's illegible.

980
00:51:15,840 --> 00:51:20,320
We then have to make sure that, okay,
they're typing in a word.

981
00:51:20,320 --> 00:51:21,840
That's what the word was.

982
00:51:21,840 --> 00:51:25,160
We have to get that voice to
then say that we then have to get it

983
00:51:25,160 --> 00:51:29,400
in the right intonation,
the right emotion, the right,

984
00:51:29,400 --> 00:51:34,520
the right quality to fit the rest
of the regenerated audio around it.

985
00:51:34,800 --> 00:51:38,560
And if you're creating one word, it's
actually a lot harder than it sounds

986
00:51:38,560 --> 00:51:39,360
to create one.

987
00:51:39,360 --> 00:51:41,400
That's average quality.

988
00:51:41,400 --> 00:51:46,440
That's, even with the rest of it than it
is to create one that's perfect quality.

989
00:51:46,800 --> 00:51:49,920
Yeah, it's a different product, really.

990
00:51:50,280 --> 00:51:53,280
It's I mean, it's adding
a feature to the product.

991
00:51:53,480 --> 00:51:55,080
I'm going to drop in an old school
solution.

992
00:51:55,080 --> 00:51:57,920
Remember in the day when if there was.

993
00:51:57,920 --> 00:52:01,080
And it can only work
if it's a small like one word problem,

994
00:52:01,560 --> 00:52:04,560
but you would get a male or a female,

995
00:52:05,120 --> 00:52:07,200
not the same voice to do it.

996
00:52:07,200 --> 00:52:08,160
And then you would cut it

997
00:52:08,160 --> 00:52:11,320
in and it's so fast that the mind doesn't
actually perceive it.

998
00:52:11,320 --> 00:52:14,120
Yeah. Who were the people that were best
at doing that?

999
00:52:14,120 --> 00:52:15,720
It was always the musicians

1000
00:52:15,720 --> 00:52:18,680
because the musicians
could follow the pitch of the person.

1001
00:52:18,680 --> 00:52:21,560
Yeah, I done it a few times.

1002
00:52:21,560 --> 00:52:25,000
I'll be honest to save things I can't
remember and I never got paid for it.

1003
00:52:25,000 --> 00:52:27,160
Like like The Voice.
I'm sure Opie could do it.

1004
00:52:27,160 --> 00:52:29,520
And I know Paul Davies and I used to do it
all the time.

1005
00:52:29,520 --> 00:52:30,360
It'd be 105.

1006
00:52:30,360 --> 00:52:32,320
Yeah, yeah, I used to do it.

1007
00:52:32,320 --> 00:52:35,240
Change. Promise.
I've done this, Robert. Where?

1008
00:52:35,240 --> 00:52:38,640
He's like, the audio is like chopped
the end of a word off.

1009
00:52:38,640 --> 00:52:38,960
And I've

1010
00:52:40,200 --> 00:52:41,040
recorded the end of that.

1011
00:52:41,040 --> 00:52:42,120
The rest of it, I mean.

1012
00:52:42,120 --> 00:52:43,320
And you would.

1013
00:52:43,320 --> 00:52:45,840
You can't pick it in the old days
if you actually,

1014
00:52:45,840 --> 00:52:48,000
if you dropped in
and you dropped the end off

1015
00:52:48,000 --> 00:52:50,320
like an ING off
the end of a word or something.

1016
00:52:50,320 --> 00:52:51,000
Yeah.

1017
00:52:51,000 --> 00:52:55,120
My I know my daughter can can match pitch
and I can supernatural way when my

1018
00:52:55,200 --> 00:52:57,200
when my girlfriend
first introduced herself.

1019
00:52:57,200 --> 00:53:02,200
My girlfriend's name is Firoozeh
and my daughter immediately picked it up

1020
00:53:02,200 --> 00:53:06,240
and I said, I said, hey, honey,
do you want to do a drawing for fear?

1021
00:53:06,240 --> 00:53:11,400
Issa and my daughter immediately says,
it's not fear as a dad, it's Firoozeh.

1022
00:53:11,920 --> 00:53:14,800
And she like immediate,
like a tape recorder

1023
00:53:14,800 --> 00:53:18,080
played back, you know, in her brain
the way it's supposed to sound.

1024
00:53:18,080 --> 00:53:19,800
And I was like, Holy crap.

1025
00:53:19,800 --> 00:53:23,640
Yeah, you know, my 12 year
old can playback audio from her brain.

1026
00:53:23,880 --> 00:53:26,040
That's the key.
That's the key to her being a mimic.

1027
00:53:27,360 --> 00:53:28,920
Yeah, you can mimic stuff.

1028
00:53:28,920 --> 00:53:32,160
Like I remember going to,
Because my other half,

1029
00:53:32,160 --> 00:53:35,000
she speaks Italian, and it was like,
you should learn Italian

1030
00:53:35,000 --> 00:53:38,440
so we can when we go there, we can both,
you know, communicate.

1031
00:53:38,440 --> 00:53:39,600
And I was like, oh, all right. Okay.

1032
00:53:39,600 --> 00:53:42,000
So I got, you know,

1033
00:53:42,000 --> 00:53:43,960
convinced I would, go and do this thing.

1034
00:53:43,960 --> 00:53:48,600
So I went to do my first Italian lesson
in Melbourne, and,

1035
00:53:48,600 --> 00:53:52,080
I finish the thing, they go, oh, well,
you know, you were really good.

1036
00:53:52,080 --> 00:53:55,000
You're really good.
And I was just mimicking.

1037
00:53:55,000 --> 00:53:58,280
But when it came to actually building
structure and knowing what the words meant

1038
00:53:58,280 --> 00:54:00,600
and how you put them all together,
I had no fucking idea.

1039
00:54:00,600 --> 00:54:03,840
So I'm just going to ask JP,
you know, to like an AI.

1040
00:54:03,840 --> 00:54:06,960
So to learn to, to learn Italian
in Melbourne, what do you do?

1041
00:54:06,960 --> 00:54:09,400
You just go to a restaurant
on like Elm Street or something.

1042
00:54:09,400 --> 00:54:10,920
And do you know what?

1043
00:54:10,920 --> 00:54:13,840
You know where the Italian school
was like on the street? Yeah,

1044
00:54:14,920 --> 00:54:17,560
mom. I mean, yeah,

1045
00:54:17,560 --> 00:54:19,560
exactly.

1046
00:54:19,560 --> 00:54:21,440
Just order an appetizer and just. Wow.

1047
00:54:21,440 --> 00:54:22,240
It's cool.

1048
00:54:22,240 --> 00:54:23,000
Yeah, yeah, yeah.

1049
00:54:23,000 --> 00:54:26,040
So just go into your lesson,
go faster and stuff after.

1050
00:54:26,040 --> 00:54:26,320
Yeah.

1051
00:54:26,320 --> 00:54:28,200
Well, you know,
I've been getting more and more teaching

1052
00:54:28,200 --> 00:54:31,640
production for podcasters and creators,
so I'll definitely be

1053
00:54:32,160 --> 00:54:36,000
mentioning this as a tool, you know, and,
and letting people know to try it out

1054
00:54:36,320 --> 00:54:40,080
because I have, you know, I'm working
more with corporate, you know, creators.

1055
00:54:40,080 --> 00:54:42,360
That's another whole, you know, category.

1056
00:54:42,360 --> 00:54:44,840
You know, people
that do content for corporate.

1057
00:54:44,840 --> 00:54:48,720
So C-suite, you know, let's let's talk
for a second just quickly about that.

1058
00:54:48,720 --> 00:54:52,480
So for example, somebody who's done
a course that they've recorded

1059
00:54:52,480 --> 00:54:56,240
their screen and the little hover over
camera in the loom or something like that,

1060
00:54:56,880 --> 00:55:01,200
but the camera is a webcam and their
microphone is their MacBook or something.

1061
00:55:01,640 --> 00:55:05,840
It's like literally if you're lucky,
lucky that, would you if you're lucky.

1062
00:55:05,840 --> 00:55:09,840
But just literally drag the course
video onto voice region

1063
00:55:10,000 --> 00:55:12,000
hit process and you're done. Yeah.

1064
00:55:12,000 --> 00:55:13,440
No. Quick question for you.

1065
00:55:13,440 --> 00:55:15,720
Where can people go and check this out.

1066
00:55:15,720 --> 00:55:19,680
Waves.com/voice region.

1067
00:55:20,640 --> 00:55:22,200
Or just type

1068
00:55:22,200 --> 00:55:26,320
in voice region waves into your, Google

1069
00:55:26,320 --> 00:55:30,400
or whatever
you want and go off and find it useful.

1070
00:55:30,600 --> 00:55:34,920
This could be, this could be an integral
part of my, road case.

1071
00:55:34,920 --> 00:55:38,400
Well, I think I was going to suggest that
even if you're not a content creator,

1072
00:55:38,400 --> 00:55:40,680
if you're a voiceover artist going,
nah, never needed.

1073
00:55:40,680 --> 00:55:42,480
Just do yourself a favor.

1074
00:55:42,480 --> 00:55:45,840
Go and just have a play with it
because you'll be blown away.

1075
00:55:45,880 --> 00:55:50,160
And one of the reasons
that I put the teleprompter in there and,

1076
00:55:50,160 --> 00:55:53,280
is because of voiceover

1077
00:55:53,560 --> 00:55:56,960
artists like UAP, and also,

1078
00:55:57,400 --> 00:56:01,160
how often you find yourself in a position
when you're,

1079
00:56:01,560 --> 00:56:06,520
you're not with your rig and you need to
record something and you need it to get it

1080
00:56:06,520 --> 00:56:09,920
back to the station or to the studio,
and have a way to do it.

1081
00:56:10,560 --> 00:56:14,480
Now, in that case, it's not taking
a producer's job away from them.

1082
00:56:14,520 --> 00:56:18,840
What it does is it gives you, Robbo,
less work to do at the other end

1083
00:56:19,200 --> 00:56:20,880
than you would have had to.

1084
00:56:20,880 --> 00:56:26,000
And that's also why I put a a teleprompter
script reader in there with zero scroll.

1085
00:56:26,280 --> 00:56:30,360
Because if you're pasting in a 32nd
script, then you know you don't need it

1086
00:56:30,400 --> 00:56:30,840
scroll.

1087
00:56:30,840 --> 00:56:34,040
You just need it in front of your face
or additions to it

1088
00:56:34,200 --> 00:56:37,960
if you got one, you I added a restaurant,
for example, or something like that.

1089
00:56:37,960 --> 00:56:39,320
You could just jump in the bathroom

1090
00:56:39,320 --> 00:56:43,080
on your phone and record it, stick
it up, do it and send it straight back.

1091
00:56:43,560 --> 00:56:48,160
You could, but I wouldn't because I
don't think it's it gets rid of slurring.

1092
00:56:48,520 --> 00:56:49,360
I'm guessing

1093
00:56:50,560 --> 00:56:51,640
the red wine.

1094
00:56:51,640 --> 00:56:52,280
Yeah. Yeah.

1095
00:56:52,280 --> 00:56:54,760
That's not good.
That's not going to cut it. See.

1096
00:56:54,760 --> 00:56:58,560
You see that's the other thing is like
I worked very hard to try and make sure.

1097
00:56:58,560 --> 00:57:00,520
And I did a lot of the coding for the,

1098
00:57:00,520 --> 00:57:03,400
for the,
for the teleprompter here in Brisbane.

1099
00:57:03,400 --> 00:57:07,680
I worked very hard
to try and make sure that the scroll,

1100
00:57:07,720 --> 00:57:10,720
especially on a mobile
phone, was not jittery

1101
00:57:11,080 --> 00:57:13,960
and it's slow enough
that you can look at it on a,

1102
00:57:13,960 --> 00:57:17,600
you know, normal phone,
which is like a tablet these days

1103
00:57:18,040 --> 00:57:23,680
and be able to read things, without
constantly moving your eyes up and down.

1104
00:57:24,120 --> 00:57:25,080
I'm really proud of this.

1105
00:57:25,080 --> 00:57:28,560
It's like and so and if anybody out
there as well is interested

1106
00:57:28,560 --> 00:57:32,480
in using an API version of this
for your own service,

1107
00:57:32,480 --> 00:57:36,120
reach out to these wonderful hosts
and they'll get in contact with me.

1108
00:57:36,440 --> 00:57:38,720
That picture on the floor behind you.

1109
00:57:38,720 --> 00:57:41,280
Lichtenstein.
Yeah. Can you can you do me a favor?

1110
00:57:41,280 --> 00:57:42,960
Can you hang the fucking thing up?

1111
00:57:42,960 --> 00:57:46,400
Know how much it frustrates me
to see looking at that picture

1112
00:57:46,720 --> 00:57:49,080
just sitting on the floor?
Because it's really cool.

1113
00:57:49,080 --> 00:57:51,160
It is really cool,
but it's also really heavy.

1114
00:57:51,160 --> 00:57:54,240
It's actually, a wooden cargo,

1115
00:57:55,000 --> 00:57:59,320
slat crate that, was painted on.

1116
00:58:00,560 --> 00:58:01,760
So if I hang

1117
00:58:01,760 --> 00:58:05,560
this or if I hang this on the wall,
it's it'll bring the wall down.

1118
00:58:05,560 --> 00:58:06,960
So, yeah,

1119
00:58:06,960 --> 00:58:09,240
I looked at the picture of the girl
and of course,

1120
00:58:09,240 --> 00:58:11,000
the context of what we're doing here.

1121
00:58:11,000 --> 00:58:14,160
Podcasting, I thought, was what it's like.

1122
00:58:14,480 --> 00:58:17,480
It's having.

1123
00:58:17,880 --> 00:58:19,960
Yeah, I know it is.

1124
00:58:19,960 --> 00:58:20,760
But I like her hair.

1125
00:58:20,760 --> 00:58:22,960
Her hair.

1126
00:58:22,960 --> 00:58:24,320
Now I'm looking at it. Now.

1127
00:58:24,320 --> 00:58:25,840
I won't be able to unsee that.

1128
00:58:25,840 --> 00:58:26,520
You can't.

1129
00:58:26,520 --> 00:58:28,200
You can't unsee it. See it?

1130
00:58:28,200 --> 00:58:28,720
That's right.

1131
00:58:28,720 --> 00:58:31,720
When she said, sorry, man. So.

1132
00:58:32,320 --> 00:58:33,360
Well, that was fun.

1133
00:58:33,360 --> 00:58:36,600
Is it over the front? Audio. Sweet.

1134
00:58:36,800 --> 00:58:40,720
Thanks to Driver and Austrian audio
recorded using Source

1135
00:58:40,720 --> 00:58:44,560
Connect, edited by Andrew Peaches
and mixed by Robert God.

1136
00:58:44,560 --> 00:58:45,840
Your own audio issues,

1137
00:58:45,840 --> 00:58:49,440
just ask Robert, echoing tech support
for George the Tech with.

1138
00:58:49,560 --> 00:58:51,360
Don't forget to subscribe to the show

1139
00:58:51,360 --> 00:58:54,680
and join the conversation
on our Facebook group to leave a comment,

1140
00:58:54,760 --> 00:58:58,520
suggest a topic or just say goodbye,
drop us a note at our website.

1141
00:58:58,560 --> 00:59:00,400
Rodeo swedish.com.