In this episode of the Confluence podcast, hosts Randall Stevens and Evan Troxel sit down with Phillip Miller from Chaos Group. The discussion navigates through Phil's extensive background in the 3D technology and AEC industry, highlighting his contributions to Autodesk, Adobe, and NVIDIA. Phil introduces Chaos Group’s latest product, Envision, a standalone real-time ray tracing tool designed for the architectural industry. He breaks down the journey of its development, the intended user personas, and its innovative features. Tune in to get a behind-the-scenes look at how Envision aims to revolutionize architectural visualization and animation workflows.

Episode Links:

• Chaos Group - Learn more about real-time ray tracing and visualization tools: https://chaos.com
• Chaos V-Ray - Visualize anything imaginable with V-Ray: https://www.chaos.com/vray
• Chaos Envision - Effortless storytelling & unmatched realism with Envision: https://www.chaos.com/envision
• NVIDIA RTX Technology - Details about ray tracing hardware capabilities: https://www.nvidia.com/en-us/geforce/rtx/
• Autodesk 3ds Max History - Evolution of 3D modeling and rendering software: https://www.autodesk.com/products/3ds-max/overview
• Computer Graphics History - Computer History Museum's perspective on the evolution of CG: https://computerhistory.org/blog/computer-graphics/

Watch this episode on YouTube or Spotify.

-----

The Confluence podcast is a collaboration between TRXL and AVAIL, and is produced by TRXL Media.

• Learn more about them at https://trxl.co/media and https://getavail.com
• Learn more about the Confluence podcast at https://confluence.getavail.com/podcast
• Learn about the Confluence events at https://confluence.getavail.com

Welcome to another Confluence podcast.

2

I'm Randall Stevens.

3

And, uh, as usual, uh, Evan Troxel and I

are going to have a great conversation,

4

uh, with Phil Miller from Chaos Group.

5

I'm going to let you give, you

know, more in depth of, uh,

6

of your resume and background.

7

But, uh, Phil's, Phil's been around

the world of 3D and technology in

8

the AEC business for, uh, many years.

9

Actually, I was kind of surprised,

uh, Phil and I kind of officially,

10

I think just met a couple of years

ago and I was like really surprised

11

given the pedigree and, and history

that we had never, uh, met before.

12

But, um, Uh, many years working

with Autodesk and products

13

there and now with Chaos Group.

14

And then, uh, and then Phil

also came out this fall for our

15

Confluence event here in Lexington.

16

So appreciate you coming and

participating and presenting there.

17

So welcome to the podcast.

18

Our goal here is that we're going

to get to reach a much broader

19

audience with this conversation.

20

So why don't we just kick it off?

21

Phil, just give us a little bit of your

background and we'll, we'll go from there.

22

Phillip Miller: Sure, no problem.

23

So in my past lives, I started

off actually as an architect.

24

Registered architect, processing, design.

25

Really got into visualizing it when the

first 3D software started coming out.

26

I was doing it just to relieve myself

of having to do perspectives back then.

27

Just finding the vanishing

point was a good start.

28

Anyways, I got into it a lot and I

started writing magazine articles

29

about it and that led to a book.

30

which was Inside 3D Studio and Autodesk

read it and said, Hey, do you want to

31

come out here and maybe have a job?

32

So I came out and decided that

would be a good time to have

33

a vacation from architecture.

34

And that was, you know, three decades ago.

35

Evan Troxel: Wow.

36

Phillip Miller: so Autodesk

at a really good time.

37

That was right when they were,

Transitioning from the original 3D

38

Studio that ran on DOS to 3D Studio Max.

39

And I got to be on the ground floor of

that product as it was being developed.

40

And then I was under my wing

for the next five releases.

41

So, that was, that was

a fantastic opportunity.

42

It was also this time that we acquired

Discrete Logic and merged with them.

43

Um, so, I got exposed to all of the,

uh, post production, uh, capabilities,

44

uh, that are now, well, Flame,

Inferno, those household names.

45

Um, at that time we also did a new

product, it was called Combustion,

46

um, which was, uh, basically taking

Flame and bringing it to the desktop.

47

Um, After Autodesk, um, I did a

startup where we were doing, um,

48

actually presentation software.

49

And after that, I actually was at

Adobe when, uh, they were actually,

50

had just acquired, uh, Macromedia

and needed some help in, uh, uh,

51

merging all those products, Flash,

Dreamweaver, those sort of things,

52

into what became Creative Suite.

53

Um, so, after that, my next gig in

technology was actually with NVIDIA.

54

And so I was at NVIDIA for about

nine years, where we were responsible

55

for helping all the companies

accelerate their applications.

56

We actually pioneered all the ray tracing

on the GPU back then, and brought out all

57

that technology, which eventually became

RTX, and all the software stack behind it.

58

Um, through that I actually

got to know, um, uh, Volato

59

and Peter at, uh, Chaos Group.

60

Um, and so after NVIDIA, that was

just kind of a natural place to, to

61

end up, was with these guys that had,

you know, created this, kind of like

62

the de facto, uh, rendering software.

63

And it was actually the same software

that, I mean, it was what we were trying

64

to enable originally with 3DS Max.

65

You know, we, we developed 3ds Max as

a platform for, for plugins, and it was

66

just kind of like coming full circle

to what is really the most successful

67

plugin ever made, which is V Ray.

68

Um, so,

69

that team there, I've really enjoyed

working with over these last 8 years,

70

um, because, or 7 years I should say, um,

because they remind me of the old days.

71

They remind me of the wild west when,

when software was fresh and new,

72

and, um, you simply asked yourself,

you know, what do customers want?

73

Um, and Go Build It.

74

laugh

75

Randall Stevens: That's the career I want.

76

That's the career I wanted.

77

No, like, you know, I kind of

got enamored with the, we're

78

both men, Close to the same age.

79

I think I got that same like

intoxication with all that

80

late eighties, early nineties.

81

3D graphics were just incredible.

82

And I don't, I don't know.

83

I think it's gonna be hard to, it's hard

to replicate that maybe the AI stuff

84

that's going on now is kind of that.

85

A little bit of, uh, Wild West and just

so much, so much activity around it, but

86

the, uh, you couldn't beat those early

days of, you know, what you guys were

87

doing with 3D Studio and those transition

and then on up into the Wavefront,

88

Alias and the Silicon Graphics machines

and everything that was going on.

89

That was just,

90

Phillip Miller: Well, I feel

like I could have done it twice.

91

Its because it's like, initially we were

just, trying to get it to work at all.

92

Evan Troxel: Yeah.

93

Phillip Miller: Like,

oh, particle systems,

94

how do you do that?

95

You know, character animation,

96

well how do you do that?

97

And so suddenly what we were

trying to just get to work,

98

now we're all making real time.

99

Randall Stevens: Yeah.

100

Phillip Miller: Um,

101

and so that is, that's where the game

changing is now that we, you know, what

102

used to take hours, if not days, you know,

you can actually do it 30 frames a second.

103

Randall Stevens: Yeah.

104

Evan Troxel: I'm going

105

to add my hat to the ring of dating

myself or ourselves because, uh, I,

106

I started architecture school in 92

and there was the PC lab and we had

107

3d Studio running on DOS in there.

108

And we had the Mac lab next door

where we had Electric Image running.

109

And I mentioned to you, Phil, at the,

at the Chaos booth, uh, at Autodesk

110

University that I would render in phong.

111

And you're like, well, that's a

word I haven't heard in a long time.

112

Right?

113

Like, Um, but, but that's one of

those things where I would go use 3D

114

Studio to do some ray tracing and I

would use Electric Image because it

115

was so fast at Phong rendering, right?

116

But, but those were the days like

you would type in the command

117

line to start 3D Studio, right?

118

It was, it was definitely a,

a different world back then.

119

Randall Stevens: I think I had this

conversation with you, similar Phil,

120

about, uh, when I was in school in

the late 80s, I was using, uh, BigD.

121

Do you remember that

122

Phillip Miller: Oh yeah, that was

123

Randall Stevens: guy out of

Dallas, uh, it was called BigD.

124

It was this rendering.

125

I remember it was on these like, you

know, five and a quarter floppies, you

126

know, so we're we're definitely dating

ourself, but those were really fun early

127

days when that stuff was in its infancy.

128

It was just incredible.

129

so What we want to try to dig in with

you today, Phil, is just a little, behind

130

the scenes on, I know you're introducing,

uh, a new product called Envision.

131

And, uh, so maybe you can start out,

just give us a little bit of, of

132

what it is, and then we can kind of

work backwards and talk about where

133

the origins came from or what the

problem that was trying to be solved.

134

And, uh, we can kind of talk

through, uh, how this all came

135

about.

136

Phillip Miller: Yeah, sure, no problem.

137

So Envision is a brand new

independent product, meaning it,

138

you install it and you launch it.

139

It's not a plugin.

140

Um, and, um, it was built around the

concept of real time ray tracing.

141

So we already had another product like

this called Vantage, but Vantage kind

142

of grew up to be more of a companion

to V Ray and is really getting

143

more into the workflows of M& E.

144

And we realized that the

architectural industry really

145

needed a focused tool just for them.

146

Um, and so that's what Envision became.

147

And so, because Vantage was

already too complicated.

148

We needed to dial this back and

really, really make it purpose based.

149

Um, so with that idea in mind, we

decided, okay, what, what would be the

150

perfect, you know, combination of things?

151

Okay, we have a real time ray

tracing environment, where

152

do we get the data from?

153

So we just made it, what we actually

started early on two years ago, uh,

154

or releasing anyways at, uh, at Chaos

after we merged with Enscape, was

155

what we called the Chaos Bridge, which

was basically getting the data, uh,

156

out of Enscape and over into V Ray.

157

But we were actually doing it

with the idea that we'd eventually

158

I'll bring it into a centralized

application, which is now InVision.

159

So the idea behind InVision that

it's a, it's like hub and spoke in

160

the sense that it's the hub and has

spokes to 10 different creation tools.

161

So wherever you find Enscape or

V Ray, uh, you can then, uh, take

162

whatever you, wherever, whatever you

left off, bring it all into InVision,

163

render ready and start assembling it.

164

So, at its heart, it's

really an assembly tool.

165

Um, you know, you're

bringing in data elsewhere.

166

Uh, that could also be assets, like we

have Chaos Cosmos, you can start adding

167

models, you can bring in your own models.

168

Um, but then also you can start, you know,

cloning and scattering them and arraying

169

them, uh, bringing in animated people

and everything else, uh, to, you know,

170

build up your scene very, very quickly.

171

Um, you don't do any mesh editing.

172

There are plenty of great

modeling tools out there.

173

No need to reinvent that wheel.

174

So, uh, we also built it on the idea

that everything was going to iterate.

175

Um, what I hated the most as a

designer was, uh, having to start over.

176

Like you would, you would, uh, you

would, uh, build out the building, get

177

all your mapping coordinates right,

and then the walls would change.

178

You have to do it all over again.

179

And so we, we came up with the idea

that all the decisions made within

180

Envision would be one layer above

the reference, what was coming in.

181

And so that the underlying reference could

change and all the, everything you did

182

to it could, could, would then propagate.

183

So it's a very powerful

assembly tool in that regard.

184

It's designed with the idea

that things are going to change.

185

Randall Stevens: Yeah, as you were, as

you were, saying that, Phil, it reminded

186

me, and you'll, you'll remember what

the terminology was, but in those early

187

DOS days of 3D Studio, the modeler

was actually different than the scene,

188

wherever you, do you remember that?

189

It wasn't combined, right?

190

So when you were, when you were saying

that, I was thinking, it's like, yeah,

191

you model over here, but then that comes

into a, you know, kind of common scene.

192

So,

193

Phillip Miller: that was, 3D Studio was

originally, had five different modules.

194

It had the shaper,

195

had the shaper, the the

modeler, the keyframer,

196

the material editor.

197

Um,

198

and originally, that was because it had to

fit within the original amount of memory.

199

Evan Troxel: Yeah.

200

Randall Stevens: that makes sense.

201

Phillip Miller: Each one was

sitting in its own memory bucket.

202

Um, and then they, uh, they

were able to break through that

203

with, with, uh, with more RAM.

204

Evan Troxel: No one will

ever need more than 640K

205

Randall Stevens: Ever.

206

What would you do with it?

207

Uh, so, uh, give it a, give an

example, uh, Phil, just so that

208

everybody, or, or, you know, if you

want to share your screen, you're

209

welcome to show, show something.

210

We,

211

Phillip Miller: I don't

have anything here yet.

212

Randall Stevens: edit stuff together.

213

But, uh, uh, yeah, maybe you

can just describe, um, you know,

214

what, what is, what's a story?

215

What's a use case, right?

216

Of,

217

of this stuff kind of coming

218

Phillip Miller: Yeah, well the first thing

is that, um, so we built it around, uh,

219

internally we call it the Levina Renderer.

220

Um, which is Bulgarian for avalanche.

221

Um, the idea was we were going

to teach the world Bulgarian

222

one code word at a time.

223

Evan Troxel: It's a slow process.

224

Phillip Miller: Right?

225

But, but, um, but seriously, it was built

with the idea that, um, to be fully ray

226

traced and to use ray tracing hardware.

227

Right?

228

The idea being that rays are free.

229

Um, ray tracing is much different

than a game engine, which is a

230

rasterizer, in the sense that with

a game engine, Pixels are free.

231

You can go as big a resolution

as you want to and not slow down.

232

Um, but you pay for every triangle.

233

And so that's why you always hear

about working within your polygon

234

budget and having to, you know, reduce

and reduce and reduce in order to

235

fit, in order to keep that real time.

236

Ray tracing is the opposite, right?

237

In ray tracing, um, you pay for

every pixel that you're rendering.

238

So as the rendering gets bigger, it

gets slower, but polygons are free.

239

So now you can just, it'll

just eat polygons and with

240

no, with no performance loss.

241

So we were planning early on, like

when we started, uh, this render about

242

six years ago, that, um, um, that

ray tracing hardware would be there,

243

rays would be free, and that we would

hit a tipping point where for a given

244

resolution that you like, say full

HD, um, it will now be real time.

245

And at that point you can just

make as big of a model as you want.

246

It doesn't matter.

247

It'll stay real time.

248

And so we hit that tipping

point a couple of years ago.

249

Um, and so now even on commodity

hardware when you buy a laptop,

250

it is ready to go with full time

ray tracing at HD resolution.

251

Which is really revolutionary.

252

Um, it's kind of like when games

went real time 3D many years ago.

253

It's just like the world wasn't the same.

254

Um, that's kind of what we

have now with ray tracing.

255

So, what type of environment is it?

256

Well, the first idea is that you're

fully immersive, as big as possible,

257

fully ray traced all the time, meaning

you have a high level of quality.

258

And we don't have to cheat, we don't

have to bake, we can just have the real

259

lighting, we can bounce the light, we

don't have to reduce the geometry, we

260

can take the original complexity in,

no matter how badly it was modeled.

261

And by the way, by bad I mean that

you've got overlapping geometry,

262

Randall Stevens: Coplanar stuff.

263

Phillip Miller: know, all stuff.

264

that is messy and not

supposed to be there.

265

We don't care.

266

We'll just keep rendering it.

267

Um, and so that, so the time to get

into this environment is super fast.

268

One export from the creation tool and

you're immediately into the ray tracing.

269

Um, and then you just keep assembling.

270

Because it doesn't care about size.

271

The files can get really big

without, in ways that, uh, the

272

creation tools can never handle it.

273

Um, and that's, that's what's fun,

when you start counting faces into

274

the billions, if not trillions.

275

Um, then, um, the amount of detail, you

know, it's like rendering, what makes

276

something look real, it's all about

the shading and everything, but it's

277

also about how much detail you have.

278

All the stuff that makes it look right.

279

And,

280

um, so it's that combination that suddenly

really, really opens everything up.

281

And, um,

282

Randall Stevens: at ArtVision over

the years we called it clutter.

283

The more clutter, we were, it

was actually part of the uh,

284

content collections were always

285

Phillip Miller: oh yeah

286

Randall Stevens: yeah everybody does the

building but you, where's all the stuff?

287

You need the desk full, you

need every, you need the shelves

288

full, you need stuff, clutter.

289

Yeah,

290

Phillip Miller: Oh yeah, in fact,

we used to do these real and

291

rendered comparisons at NVIDIA.

292

You know, can you tell

which is a photograph?

293

Well, you'd study it, and

it's like, well, what's wrong?

294

Because you just wouldn't model it wrong.

295

Evan Troxel: Right?

296

Phillip Miller: And, um, you

know, all that stuff, like you

297

said, the clutter, the junk.

298

Um, that's what makes it look real.

299

Um, and that's, that's also what,

uh, scattering is great for.

300

Uh, put in collections of

stuff and just scatter them.

301

And, uh, you know, maybe there's

a, you know, uh, uh, a wrapper

302

in there or a cigarette, but,

or something like that too.

303

Um, but as long as it's randomized and

everything, um, suddenly it looks real.

304

Evan Troxel: I spent

305

a lot of time doing a deep dive because I

did a stint in visual effects and modeling

306

and animation and all these things.

307

And, and I did a deep dive into

the Star Wars universe because,

308

Phillip Miller: yeah.

309

Evan Troxel: Why did it look so real?

310

Because everything had scratches on it.

311

Everything

312

had dents and dings.

313

Everything had dirt, right?

314

It had,

315

and it was like grunge maps and, you

know, changing the specular highlights

316

with grunge maps and, and adding really

tiny chamfers on, you know, like as an

317

architect, you're designing stuff and

you draw a street and a sidewalk and, The

318

curb needs a curved edge on it, right?

319

It needs a chamfered edge because

that just that change in specular

320

makes it look so much more real,

321

but all of the rendering engines penalized

us as for for reality back then, right?

322

Because the more polygons, the

worse it got right back then.

323

So it's really interesting to hear how.

324

I mean, we, we went through

severe training to reduce

325

polygon counts for, uh, for,

326

decades, right?

327

And now you're saying,

328

we don't have to worry about that anymore.

329

Yeah, it was like, it, it

just became part of you.

330

It was, uh, just reduce, reduce, reduce.

331

You can't get as close

to reality as you want.

332

It's really interesting, this

kind of paradigm shift to now say,

333

okay, we hit this tipping point.

334

There's no more penalty.

335

And I'm curious to hear how, like, now

there's a behavior shift that is possible.

336

Yeah.

337

And I'm just wondering, like, like,

I'm sure there's a big sigh of relief

338

from, from people out there, but at the

same time, like, they've been trained

339

to do it a certain way, uh, do you

think there's gonna be a hard habits to

340

break, or do you think it's just gonna

be super easygoing because this is how

341

they wanted to do it in the first place?

342

Phillip Miller: We haven't

heard anyone complain.

343

Evan Troxel: Yeah.

344

Phillip Miller: Um, because, and

most people are just like, just

345

the speed of getting into it.

346

Like there's not, you don't have the

speed bump, um, before, you know,

347

having to get things real time.

348

It's just like going immediately

from your design tool into, um,

349

into real time is what making

people just feel so much better.

350

Um, because it's, it's literally,

you can, you can actually get it

351

into InVision quicker than you

can't load the original CAD file.

352

Evan Troxel: Yeah.

353

Phillip Miller: it's, um, that's

the sort of thing that, um, that

354

really changes things for people.

355

Um, so anyways, the other thing

that happens in Envision, so we

356

really wanted to try to, first of

all, had to be an animation tool.

357

Um, because unlike Vantage, which we

had before, which plugged into the DCC

358

tools, Uh, Digital Content Creation, those

are the, basically that's an animation

359

tool, like Mac, Cinema 4D, Maya, so on.

360

Um, they, wherever you animate

is where you're gonna render.

361

Whoever, whoever owns the

animation owns the final render.

362

Evan Troxel: Yep.

363

Phillip Miller: Softmash

taught us that a long time ago.

364

Fine, model an alias, you're

going to end up rendering here.

365

Um, and so, with that in mind,

you know, Vantage had already,

366

like, all the animation was going

to be done in like Max and Maya.

367

Why try to add animation to Vantage?

368

With Envision it was the opposite.

369

The CAD tools had no concept of time.

370

You know, and putting animation

into a, into a design tool

371

is a really unnatural act.

372

Because who owns what,

at what state, you know?

373

And so,

374

we had to really own all the animation.

375

So our challenge was,

how can we make it easy?

376

Because, animation's hard.

377

Evan Troxel: Yeah.

378

Phillip Miller: Um, and, like, One of

the things we were like, can we actually

379

do an animation tool that doesn't

have a keyframer, or need a keyframer?

380

Because that's when it starts exploding,

in terms of like seeing, seeing

381

all your keys and dope sheets and

function curves and everything else.

382

Um, and so we came up with the

idea of, uh, animate a scene state.

383

So, we have a concept of doing design

variations in InVision, where, by

384

which, you, at any point, you can

decide, this is a variation, and it

385

can be of anything that you've changed.

386

You could have moved objects, you

could have hidden them, um, you could

387

have changed materials, material

assignments, uh, uh, all the lighting,

388

the environment, all that other stuff.

389

All that becomes one big, uh, scene state.

390

Now, the idea originally was, like,

people could flip between designs.

391

Um, so you could actually look at all your

design permutations really quick, because

392

all we're doing is flipping parameters,

we're not changing our geometry.

393

Um, and that then led to, okay,

well, what if we just treated

394

it as a giant animation key?

395

Randall Stevens: Yeah.

396

It's like a big storyboard.

397

Phillip Miller: Yeah, exactly.

398

That's a really good, that's a

399

really good comparison.

400

Um, so the idea is, is that it's also

WYSIWYG because you see it, right?

401

You see everything rather than curves

and functions and, and parameters.

402

Uh, you just see what it's going to be

as an end state and you just go, I want

403

to go from here to here and boom, it

just, we just transition everything.

404

Um, and so like these, like these,

uh, phasing animations that became

405

so popular in do it yourself shows,

406

Randall Stevens: Yeah.

407

Phillip Miller: now suddenly became

super easy to do, because you just

408

move everything up and then

move them down and boom, and

409

everything gets animated very, very

410

Randall Stevens: Is there a style of

animation then for those transitions?

411

Cause like, I've always wondered when

I watch those shows, it's like, okay,

412

how did the, you know, sometimes

they'll flip up into place or, you

413

Phillip Miller: yeah.

414

yeah.

415

So we, we got inspired by that too.

416

And we're putting in, uh, uh, what

are called, uh, uh, basically phasing

417

controllers to allow people to, to allow

them to come in and be more playful.

418

Like

419

they could come and

bounce and wobble or, or,

420

or spin in place and all that other place.

421

And so you want to automate that.

422

Yeah, you know what you want and and give

people the ability to look like a Disney

423

animator without Really knowing anything

424

about it so That became super powerful

both from viewing designs and then

425

suddenly becoming a way to animate.

426

Now you can animate other ways too,

you can put things on paths, there will

427

eventually be behaviors where you walk up

and click on door to door opens and so on.

428

But, in the meantime, we have this

really powerful way of doing animation

429

that people are finding very intuitive.

430

Um, uh, when I asked one of our

people from Augusta University to

431

do a facing animation, he said,

He'd never thought about it before,

432

and he had it done in 10 minutes.

433

Evan Troxel: Mm hmm.

434

Phillip Miller: you know, normally

that would be an afternoon.

435

Randall Stevens: Right.

436

Evan Troxel: for

437

sure.

438

Yeah.

439

Phillip Miller: Um, you know,

it was just, it was just obvious

440

to him how it might work.

441

Um, so the other thing is like, we tried

to be very, very purpose based, like,

442

so we wanted to focus on just what a

designer needed without making it general.

443

So like 3ds max, which I originally

did, it's incredibly general.

444

You, there are no rules.

445

You can do whatever you

want, anytime you want.

446

Well, that also makes it hard

because you don't know what to do.

447

Evan Troxel: Yeah.

448

Sky's the

449

limit.

450

Yeah.

451

Phillip Miller: yeah, you're

like, where do I start?

452

Where do I end?

453

Am I on the right path?

454

Well, you're defining your path.

455

So we decided to put people on

rails, um, and make them really

456

successful or simple things.

457

So my, my example is, um, you won't be

able to bounce a ball down the stairs,

458

um, in Envision very easily, actually not

at all, but, um, you can have a person

459

walk up the stairs in a couple of clicks.

460

Randall Stevens: Right.

461

Phillip Miller: So it's kind of like

462

the opposite, like normally that would

take an afternoon to have a person walk

463

up the stairs, and now you can have

them walk up the stairs very easily.

464

Randall Stevens: Yeah.

465

I, uh, you know, just talking about the

complexity, I'm, I'm, I'm a big believer,

466

Simplicity too, and it is hard, right?

467

The simpler to make it, you either got to

constrain it and then try to simplify it.

468

But, um, back in the nineties, I,

for a lot, I was teaching, uh, at

469

the, in the architecture school and

teaching a bunch of computer classes

470

and 3d studio all through the nineties

was the main piece of software.

471

And I started that in the early nineties.

472

And then by the early, by

473

I always, in hindsight, because I've told

this story before, it's like, I don't know

474

if it was me or the software, I'm going

to, I'll say it was the software, but I

475

found that it was like taking me longer

and longer to get the students up to the

476

point where they could just do anything.

477

And the software, right, as it got

more sophisticated, got harder and

478

harder to, you know, to lay out the

groundwork for what you were going to

479

do because there were so many choices.

480

I mean, 3D Studio is amazing.

481

piece of software, right?

482

You, like I said, you

could do anything with it.

483

And, uh, but, but that's the trade off.

484

All that power means now

you've got to build up that

485

underlying base of knowledge,

486

right?

487

And you had to understand.

488

Yeah.

489

I always just say, it's like, you know,

if you want to get into the mature

490

editor, you better understand physics.

491

It's like, okay.

492

Right.

493

Uh, so I can, you know, I have

a big appreciation for that.

494

It's kind of the polar opposite,

which is all that complexity, but, but

495

hidden.

496

from the user.

497

And it

498

should be right.

499

Phillip Miller: Yeah, exactly, you know,

one of the hardest things to do when

500

making software is to make it easy.

501

It's,

502

it's, um,

503

Randall Stevens: It's inversely,

uh, I like to think that it's, uh,

504

probably heard somebody say this.

505

It's like the simpler it is in the

front, the usually the more complex it is

506

in the back.

507

It's like inversely proportional and, uh,

you know it when you're developing things.

508

It's like, it's really

hard to make things simple.

509

Phillip Miller: And we had a

real challenge here because we

510

were trying to make Envision

the real partner for Enscape.

511

And Enscape had done a great

job at being super simple.

512

Um, and so we didn't want to spoil that.

513

Now, we don't know if it's going to be

514

the same person, but we wanted to make

sure that it was the same project that

515

was going to come over into Envision.

516

We wanted to make sure it

could start where it left off.

517

So, right from the beginning,

we had to become experts on

518

the Enscape material model.

519

Because we were kind of merging that

material model with that of V rays.

520

Um, and the idea being that it's

going to look exactly like you

521

were editing it in Enscape, but now

you have the option to promote it.

522

You have the option to take it farther.

523

Um, but you still can, can have the,

the basics there whenever you want.

524

Um, that's just one example of

what, of what we're trying to do

525

the transition because, um, uh,

the, the renderer within Envision

526

is, is, is different than Enscape.

527

Enscape is, uh, is very clever.

528

It's, uh, it's a hybrid renderer.

529

It's both rasterizer and raytracer.

530

So it's trying to give quality and

maximum speed at the same time.

531

Um, but it's more limited.

532

It doesn't do as much.

533

Um, and

534

so, um, uh, the Vantage, or I

should say the Lavinia renderer

535

that Envision uses, um, is trying to

parallel V Ray more and more and more.

536

Um, and so, uh, in doing so,

it's getting a lot of capability.

537

And then the trick was, don't expose it

all at once, because suddenly it looks

538

like, uh, you know, the cockpit of a 747.

539

Evan Troxel: Right.

540

Overwhelming.

541

Yeah.

542

Phillip Miller: You just, you just want

to have it one little piece at a time.

543

And so that's, that's what we

kind of learned when we were

544

putting this all together.

545

Randall Stevens: So, Phil, you said

this is a couple years in the making.

546

Um, can you kind of walk us

through what, how does that start?

547

How did the, how, how does the idea or

the spark for what became envisioned?

548

Can you backtrack and tell that story?

549

And then what does the team look like?

550

Like, you know, is that, was

that you there at the beginning?

551

Did that come from somewhere else?

552

And then how does that end up

evolving into ultimately become

553

a product like Envision today?

554

Phillip Miller: Sure.

555

Um, yeah, every product has its

own, you know, pedigree or genesis.

556

Um, this one is a little bit more,

a little bit more interesting, um,

557

because it really involved a lot

of different places within chaos.

558

So once we merge with within, so to

begin with, to take people back a

559

little bit, when we were just chaos

group, uh, or chaos, actually, um,

560

um, the V Ray company, so to speak, we

were always focused on photorealism.

561

And we were, we, well, we had,

we had just started off with

562

this real time ray tracing.

563

Um, it wasn't the same level

of, uh, interactivity or full

564

solution that Enscape had.

565

So in Enscape, we merged with, that

had the other side of the equation.

566

We were kind of like bookends.

567

Um, in the sense of like real time design

exploration with Enscape, and then you've

568

got complete photorealism however you

want to take it with V Ray or Corona.

569

Um, but we realized.

570

That there was this big middle.

571

When we were doing a lot of the

competitive analysis and everything, and

572

doing the merger of the two companies,

we happily realized there was no

573

overlap between Enscape and V Ray.

574

None.

575

Like, they don't compete

with each other at all.

576

That was great, and we

could have a transition.

577

But what we found out was that

there was this middle group that

578

were using dedicated presentation

tools based on game engines.

579

that were actually avoiding the need

to go to an animation tool because

580

they were enough, you know, so, um,

and, and we realized that this big

581

middle was solving an important need,

and we didn't have a solution for it.

582

And we felt that we could do it better

by having a continuum, not just a

583

point solution, but rather going all

the way from real time in the design

584

to this middle solution and then even

over to photorealism or back and forth.

585

And so that's what we

call the chaos ecosystem.

586

So we, when we started doing this

product, it was actually pretty far

587

reaching in the sense that we started

looking at how all the data was

588

exchanged, what are the workflows?

589

How can we actually, you

know, do this over time?

590

And I'm sorry, how, how can we

actually get all this stuff flowing?

591

Because that's where we can provide

something that the other guys don't have.

592

The other guys are, you're starting

from scratch, and they're dead ends.

593

We didn't like that.

594

People don't want to work that way.

595

People want to be able to start

where they left off and, and

596

have a parachute if they need it.

597

They want to have people to, uh, to

contribute from all different, from

598

the tools they want to use, bring

them in and, and it all just works.

599

And so that was, that was the overall goal

for why we needed this independent, like a

600

hub and spoke, as I, as I said originally.

601

Um, so now, how do you build it?

602

Um, so we

603

had,

604

Randall Stevens: before you get into

that, is that, was that being that, that,

605

um, scoping I'll say of the problem.

606

Was that you leading that or somebody

else on the team or, or who does that?

607

Phillip Miller: it was, so it's a

lot of it had actually started in

608

the due diligence when we were,

when, when we're, the companies

609

were forming and that we realized

610

this, this really basic need,

you know, really three years ago.

611

Um,

612

and so we, we, it was on

our minds of like, that is

613

something we want to go solve.

614

And you, you do the classic like build

versus buy, you know, comparison of

615

like, and what was available to buy

in theory, let's say, um, we didn't

616

like, because it wasn't going to

solve this overall workflow thing.

617

It would just be another point solution.

618

Um, and at the same time, we looked

at like, you know, what, so then it's

619

like, okay, you're going to build it.

620

What are you going to build it on?

621

So, um, As applications go, you

start with either, you start usually

622

with a framework for what is going

to be handling all the complexities

623

of like file, file I O, hosting,

scene graph management, everything

624

about working in a 3D environment.

625

Um, there's a, there's a shortcut

to this, which is you could actually

626

go and, and use something like a

game engine, which is what some of

627

these other solutions have done.

628

Um, but then you're kind of,

you have to live with the

629

limitations of the game engine.

630

It's, it is, it's, it is good and bad

in the sense that you get a lot of,

631

lot to free to start with, but then

you're at the mercy of that game engine

632

evolving the way you want it to evolve.

633

We took a hard look at what was around

and we realized that it wasn't gonna

634

fit what we wanted to do with scale.

635

Because most of them were, as we

said before, you're having, like

636

you said, you're having to modify

everything and reduce the complexity.

637

That's what they're designed for.

638

Um, we wanted to do the opposite.

639

So we realized that we were going

to have to have our own framework.

640

We had actually started the framework

of an independent product earlier, and

641

it was going to be, at that time, it was

an independent exploration of rendering,

642

where you would actually bring it in

and do your own rendering environment.

643

But we saw that it was more of an

experiment than anything else, but it

644

had the underpinnings of being a product.

645

And so there was the, this is kind

of like when you go to a, like, the

646

chassis, you just have the wheel,

you just have the frame of the car,

647

Randall Stevens: Yep.

648

Phillip Miller: a few

649

wheels.

650

It doesn't even have a steering wheel yet.

651

Um, that's kind of what we had, um,

652

Randall Stevens: Yeah.

653

I think, uh, you, you bring up a

good point though, because we've,

654

I've had that experience too.

655

It's like,

656

whenever you're working on something,

657

Phillip Miller: built.

658

Randall Stevens: I've had a couple

of times where there's like a

659

couple, I'll call them false starts.

660

It's kind of like, you have

to, you have to do enough to

661

learn a little bit and then you go, okay,

that's not, I'm going to start again.

662

And then finally it kind of

all starts to gel, but it is, I

663

think it's these experiments that

664

have to

665

be run

666

Evan Troxel: It's just like architecture.

667

Yeah.

668

Randall Stevens: Yeah.

669

Evan Troxel: Yep.

670

You learn a lot through

the failures, right?

671

Yeah.

672

No, not.

673

Yeah.

674

Randall Stevens: No, but especially

if you're, if it's new technology,

675

right, you have to get your hands

around, you have to get your arms

676

around, you know, just anything that's

new, you got to like experiment with

677

it and kind of get a feel for it.

678

And then sometimes, but like you said,

you're, you're, a lot of times you're,

679

you're making some low level early

choices and then, uh, you know, that

680

that's critical to, to get that right.

681

Phillip Miller: now we also wanted

to get this to market pretty quick.

682

So we wanted to use pieces of

technology that were already in

683

the company that we could reuse.

684

So from the very beginning,

it was, it was an easy choice.

685

We're going to use the exact

same renderer that's coming,

686

that was, that was in Vantage.

687

Um, so that's a big piece of

work that can come on over.

688

Um, now the file system that

goes into that is actually what

689

we were perfecting at the time

with all the different exporters.

690

And so we were going to keep

on leveraging, leveraging that.

691

Um, then we had the entire, Cosmos

library and infrastructure for

692

loading models and everything else.

693

So, hey, we'll go and use that.

694

Then we have a piece of technology that

we've been making consistent across

695

our products called Chaos Scatter.

696

So this was a really powerful

scattering tool that's only getting

697

more so, and so let's put in that.

698

Now, animated people, we don't have that.

699

That's hard.

700

Let's go buy that.

701

So we acquired Axis.

702

And so the reason for us bringing

Axis was that we would have a complete

703

character animation system ready to go.

704

They have some of the

best of breed characters.

705

And a way to actually do crowd animation.

706

Which is no small feat.

707

They also had figured out

a way to do it very simply.

708

But they had never been

fully integrated in a tool.

709

Uh, if you, if you've ever used Axis,

which is called Anima, Anima All.

710

Um, uh, it's an independent product

that, that connects to the likes

711

of Max and Maya and Cinema 4D.

712

Um, but you have this other environment

that you have to keep going to.

713

Um, so the idea was, what really

excited the Axis team was, now

714

they could finally do it right.

715

They could actually have it

fully integrated right from the

716

beginning, and you wouldn't have

to go to this external tool.

717

So now you can just draw your

splines, drop in your people, and

718

suddenly you've got people walking.

719

Um, so

720

that was an example of like piecing

together, in this case, acquiring,

721

um, and merging, and so these

kind of building blocks come up.

722

So the development itself actually,

um, was primarily done in Prague.

723

Uh, Prague is our development

center for Corona.

724

Um, so that was the team.

725

Now, uh, about halfway through

the process, we realized

726

we wanted to go faster.

727

And so we actually pulled some

of our best developers, um,

728

from Sofia, where V Ray is made.

729

Um, and so some of the, some, some

veteran V Ray developers actually

730

jumped on board as well to help

the process go a bit faster.

731

So, it's right now the development is

kind of split between the two cities.

732

Sofia, Bulgaria, and

Prague, Czech Republic.

733

Um, and, um, you know, I, I try to help

my team actually more than anything else.

734

Also the UX team got built up, um,

around this time as well to get much

735

more expertise on here to try to

make things cleaner, easier to use.

736

Um, uh, really kind of like.

737

Setting the bar for what we wanted to

do across all of our products, so we've

738

got the benefit of that as well So those

were kind of like the underpinnings, but

739

otherwise it is a one of one of the rare

times when you see a Product that has been

740

done from the ground up from scratch So

it's not based on another scene graph.

741

It's not based on any type

of other engine Just based

742

Evan Troxel: Can you speak to just

the internal team mindset of pulling

743

people from maybe a product they've

been working on for a long time, like

744

V Ray, for example, and then moving

over onto the development of a brand

745

new product and just kind of what the.

746

The sense was, was there excitement?

747

Were, were they looking forward

to working on a new product?

748

Were they feeling distracted from the, the

main thing that they had been working on?

749

Just kind of wondering from like

a, you know, like, like what's the

750

general vibe in the, in the teams as,

as you're developing a new product?

751

Is there a lot of excitement around that?

752

Phillip Miller: Well, that's

a great question to ask.

753

Um, so first, first of all, let's

start with the original team.

754

The original team had been working

on this one product, um, and then

755

we were telling them to That we were

going to need them to completely

756

refocus on this other, other problem.

757

Um, as it happened, they had just

done some early test, uh, test

758

cases with their, with their own

customers to find out that, wow, it's,

759

it's, People really like that, but

they're only going to need one copy.

760

They're not, it's not going to

be, it wasn't going to be that

761

much of a commercial success.

762

That's what we were realizing.

763

And also that the list of

things they needed to complete.

764

was pretty similar to what we needed

to complete on the other side too.

765

So, uh, they realized that

they should shift gears.

766

It was still going to be what they had

been working on from, from the chassis

767

standpoint, the foundation, but now

the purpose was going to be different.

768

It was going to be much

more, much more real time.

769

Now, what's that?

770

Randall Stevens: And reach more people.

771

Yeah.

772

Phillip Miller: Right.

773

And it's

774

Randall Stevens: I think anytime

you're working on something, it's, uh,

775

it's exciting to think that there's

more people that are going to use it.

776

Phillip Miller: you know,

you're absolutely correct.

777

Um, now a couple of these guys

had gay mentioned backgrounds.

778

They had done tools for

game engines in the past.

779

So getting into a more real time

mentality was also fun for them.

780

So they actually really got

into that sort of thing.

781

Now as we grew the team, like you

said, from the other location,

782

um, uh, the development managers

did this very carefully.

783

Uh, the, in the sense that they,

nobody was told to work on it.

784

They were recruited and so they come, they

came over willingly, um, with the idea

785

being that they could work on something

fresh and, and if it did work out, they

786

could always go back So they, so in that

regard, they, it was, uh, a, a really

787

big new opportunity for these people.

788

Um, and, um, you know, it, there's

been, uh, you know, having two different

789

teams in, in different locations,

um, it has its struggles, but, you

790

know, it, it, it's all working out.

791

Randall Stevens: Can you talk,

Phil, about what does it look like?

792

Um, how much do you

all, uh, diagram, write?

793

You know, is there a

narrative that's written?

794

How formal is the, you know, scoping

or definition before people actually

795

begin to start writing code?

796

Traditionally always like to just go

start writing code, but how much is there?

797

Uh, what does that process look

like before any codes written?

798

Phillip Miller: Yeah, so we,

first of all, we had to scope out

799

what we needed the product to do.

800

And so the, really, that came down

to it, we, we, this was actually

801

just intense working sessions

over the course of an entire week.

802

We probably worked 80 hours that week.

803

Um, uh, all in Prague.

804

Um, and, um, from my side, my, my, my,

key right hand person on this was Simeon,

805

who was the product manager on Vantage.

806

Um, and before that he

was on V Ray for Unreal.

807

So his, his background in, in,

in real time was, was superb.

808

Um, and together we worked closely

with the development team early on to

809

really chart out where we wanted this

to go, really for the next five years.

810

Because you need to plot out the ultimate

destination to make sure that you're

811

in between steps or on the right track.

812

Randall Stevens: Yeah.

813

Phillip Miller: Um, and so we did that.

814

Um, and we really identified the big

boulders that had to get done early,

815

and there was a lot of architectural

work that had to get done, found,

816

I mean, foundational work to the

architecture is how I should say it.

817

Um, which kept the, the, the,

uh, the team very busy for the

818

first, about half the year.

819

They, they were having to reconfigure

that, uh, what this, what, what this

820

vehicle was going to do, um, in that time.

821

We started specking out all of the

feature areas as best as we could.

822

And

823

Randall Stevens: And are, are

y'all doing wireframes mock ups?

824

Are you

825

Phillip Miller: we did

it, so we did it from a,

826

so we did it from a stand, the

most important thing to get right

827

was first the overall, uh, UX

framework in the sense of like where

828

things were going to be located.

829

Where could you find it?

830

And we, and for this product, we use

the paradigm of, of, uh, driving a car.

831

You've got the windshield.

832

That's your viewport.

833

That's your rendering viewport.

834

Um, you've got things on the top,

which are always there, your toolbar.

835

You've got things on the left hand

side, which are, which are how you

836

find things, sorting, searching,

uh, all that sort of things there.

837

Uh, the right side is once you've selected

something on the left or interactively.

838

Then, you've got all the modifications,

all the editing is on the right, and then

839

on the bottom, we have these big areas

we call the Dock, which is where you

840

can search for things, you can animate,

you can, you can, whole paradigms are

841

down there for when you're entering a

different world, but the whole idea is

842

that you want your windshield up all the

time, and, and, and to, uh, uh, cause

843

we wanted this as visual as possible.

844

Um, one of the goals was also set up a set

of goals for what the product needed to do

845

in the sense of kind of like a philosophy.

846

Randall Stevens: Yeah.

847

Phillip Miller: Um, how you should think

about things, you know, you shouldn't

848

have to come and ask us all the time.

849

Think about our philosophical statements

of like, how, what should be the

850

behavior, where should things be located.

851

Um, you know, kind of like, that became

an ethos that all the developers started

852

to get, and they started answering

their own questions pretty quick.

853

Randall Stevens: Yeah.

854

Phillip Miller: Um,

855

Randall Stevens: a, I think

it's an important part that,

856

uh, I know we try to do.

857

I try to do here at avail

and that is like, even before

858

I've even shown somebody.

859

A piece of software,

something new that we've done.

860

I want to tell the story, the

philosophy, here's why we did it.

861

Because then if, if you agree or can kind

of follow that, once you see it, then

862

you'll go, Oh, I know why that's there.

863

Why they did that, that way.

864

It's like, it maps back into this kind

of view or philosophy of the world.

865

Right.

866

And,

867

Phillip Miller: Yeah, I mean,

like for example, one of the

868

philosophies was, that we had to

establish was like, who owns what?

869

In the sense of like, because

we're bringing all this data in.

870

from other creation tools, but like, well,

where does what they did end and we start?

871

And we have to be consistent in a

way so that users won't get confused.

872

So that they can understand that,

okay, that's what I can do here,

873

that's what I should expect.

874

Um, when I modify that,

that's what I should expect.

875

That's going to persist,

that's not going to persist.

876

Um, that sort of thing.

877

And to do that, we had to

get very clear on ownership.

878

Of like, where, by ownership I

mean like, who owns the data?

879

Where is it going to change?

880

Where can it be modified?

881

Or is it just an adjustment to it?

882

Um, but, so anyways, so that

was kind of like the high level

883

philosophy, but then we come back

to actually creating user stories.

884

So, we use JIRA as a development

framework for assigning tasks

885

and describing everything.

886

From JIRA, they link to mock ups for

what the UX would look like, what the

887

workflow would be, and so on like that.

888

Um, and so every single area

has its own complete user story.

889

Um, sometimes they get rolled up into

larger epics for describing bigger things.

890

But, um, otherwise, you know, the, uh,

my, my team was, which was really, um,

891

Simeon, Anna, who actually came over

from V Ray for Revit, and then I helped

892

out when I could, but the three of us,

kind of like two and a half of us, were

893

writing stories constantly for a year.

894

Randall Stevens: yeah,

895

Phillip Miller: You know, because

every little detail had to be done.

896

Now, Talking about Iterate, Evan, this is

what you'd find out, like you'd design it,

897

and then it would get halfway implemented,

and then you'd realize, eh, eh.

898

Evan Troxel: Right.

899

Randall Stevens: feel right.

900

Phillip Miller: then, and so you have

to quickly change things, and that's the

901

other thing we had to get people, um, uh,

accustomed to, is that, to be flexible.

902

Because, um, because developers by their

nature, they don't mind taking direction,

903

but they really hate to change things.

904

Right.

905

Evan Troxel: like you have your

structural engineer, it's like, well,

906

the design changed again, and they're

like ripping their hair out because

907

they don't want to change anything

after they've done it once, where

908

the architect's job is to iterate

and to learn through these failures.

909

It's a very similar process

that you're describing.

910

Phillip Miller: Yeah, exactly.

911

So we, um, we're conscious of that,

you know, and we try to involve

912

them early on in the process to

catch it before they're coding.

913

But, um, sometimes we'd still have to make

914

Randall Stevens: Yeah.

915

And sometimes you can do, you can do a,

you can do all the pre work and thinking

916

and validation and, and all of that.

917

And then you do it and it

just doesn't feel right.

918

Right.

919

And then it's like, okay, you

gotta be willing to go back.

920

Phillip Miller: Yep.

921

Yep.

922

We already have a list of

things we need to modify.

923

Randall Stevens: Yeah.

924

Phillip Miller: um,

925

but, uh, you know, in some of

the things that are like some of

926

the most basic things, you spend

an enormous amount of time on.

927

Because if they're not right.

928

You're going to pay for it forever.

929

Um, like on the original 3DS

Max, the transform system

930

probably took three months.

931

Randall Stevens: Yeah.

932

Phillip Miller: Just making sure all

the transforms worked exactly the way,

933

and they haven't been modified since.

934

Evan Troxel: Yeah.

935

Yeah.

936

Randall Stevens: either personal

philosophy and having worked on multiple

937

products like this over the years?

938

Like, do you try to end up, uh,

being a Being a minimalist as a,

939

you know, like you said, you, you,

you try to think out five years,

940

which I think is a great exercise.

941

It's like, where could this go?

942

And then you're going to work your

way back to, okay, what, what's the

943

essence or what do I have to do first?

944

And what's your philosophy

around like just adding things?

945

Like, do you, do you let that go or do

you try to to not add things and then

946

wait for a customer to ask for it.

947

Cause that we do, I do, I go

through this all the time.

948

It's like, we'll imagine something,

but then I'm always wanting

949

to say, well, let's not do it.

950

And then let's wait to let a

customer tell us that they need it.

951

That way we don't add something

that, that ends up not being used.

952

And because it's always hard

because there's always one person.

953

That loves it and then you

can't rip, rip it back out,

954

but anyway, maybe you can just

give us some of your thoughts,

955

uh, around how you, how you think

about that, how you've approached

956

that.

957

Phillip Miller: okay, yeah, so

there's a couple things there.

958

Sometimes we put out things as, um,

we're not sure of how it could work.

959

Um, or if it's work and

we want to get feedback.

960

Um, and so then we, we, we, we listen

and try to modify according to when

961

people actually use it, I mean,

uh, try to actually produce stuff.

962

Um, that's one thing, adding

stuff and removing it.

963

Um, I, I know what you're talking about.

964

The, um,

965

usually, um, Usually you don't

have to, usually it evolves

966

rather than get removed.

967

It

968

gets,

969

it turns into, it morphs into

something, something else.

970

Um, but otherwise,

971

yeah, yeah, like a fine line, right?

972

Randall Stevens: Yeah,

973

Phillip Miller: Um, but the, the, When it

comes down to the overall roadmap and what

974

we're doing, I would, I would like to,

first of all, we know there's a lot to do.

975

Like I said, we've got five

years of work ahead of us.

976

We, we know that.

977

But along the way, we're

going to get surprised by what

978

people find, uh, important.

979

Like, from Autodesk University,

we came off with a list of

980

like two or three things.

981

Oh, well, people really want that now.

982

Randall Stevens: Right, right.

983

Phillip Miller: Okay.

984

Let's try to get that in.

985

Um, so we, we try to reserve capacity.

986

So we're, so we're, we're at least

a half, if not two thirds committed,

987

but we want to hold, hold back at

least that other third or something

988

to be able to react to, um, you know,

great ideas and everything else.

989

Um, of course, if we ask you,

if you, what you, if you want

990

it, you'll always say yes.

991

Randall Stevens: Well, yeah, yeah.

992

Phillip Miller: So.

993

Randall Stevens: no cost, right?

994

Phillip Miller: Yeah,

995

so,

996

Randall Stevens: give a.

997

Oh, go ahead.

998

Phillip Miller: when we do do

those exercises with customers, we

999

usually make them, make them choose.

00:51:00,203 --> 00:51:02,073

Um, you know, we give them a list of

00:51:02,073 --> 00:51:02,243

like

00:51:02,308 --> 00:51:03,308

Randall Stevens: you

rather have this or this?

00:51:03,473 --> 00:51:04,203

Phillip Miller: yeah, we,

00:51:04,948 --> 00:51:05,608

Randall Stevens: You only get to choose

00:51:05,713 --> 00:51:06,913

Phillip Miller: with, exactly.

00:51:07,343 --> 00:51:11,443

Um, there's different ways of doing it,

but, uh, we always want to make sure

00:51:11,443 --> 00:51:15,043

that they realize that they're giving

something up if they want that thing.

00:51:15,338 --> 00:51:17,628

Randall Stevens: Now I'll give a, I'll

give a little example of something that

00:51:17,628 --> 00:51:23,588

we just released a feature, uh, in the

avail platform with, uh, we were going

00:51:23,588 --> 00:51:28,538

to make a PDF exporter, so the idea was

that you're going to take graphic, put

00:51:28,538 --> 00:51:32,878

it on a canvas of some document of some

type, and then get it out to a PDF,

00:51:33,168 --> 00:51:35,388

mainly just to facilitate redlining.

00:51:35,398 --> 00:51:36,308

That was the story.

00:51:36,688 --> 00:51:39,068

Somebody needs to get it out,

mark it up, hand it to somebody.

00:51:40,193 --> 00:51:45,393

Well, when we started coding the,

the, the document that became,

00:51:45,613 --> 00:51:47,813

Oh, you could change the size.

00:51:47,833 --> 00:51:51,233

You can, it could be, you know, as a

letter size, you know, now all of a sudden

00:51:51,233 --> 00:51:54,653

you have this, well, what kind of document

is it and what kind of size is it?

00:51:55,623 --> 00:52:02,353

So we actually had it coded up and

gave it to a customer to, um, you

00:52:02,353 --> 00:52:03,873

know, to, to look at and play with.

00:52:04,573 --> 00:52:07,443

And all of a sudden it became about.

00:52:08,148 --> 00:52:13,268

Now the document and what the

document size and like this, I call

00:52:13,268 --> 00:52:17,268

it, you know, it's the feature creep

of now it's not just about getting

00:52:17,818 --> 00:52:20,088

something out to do markup on.

00:52:20,108 --> 00:52:25,238

It's about now I'm going to

start composing a sheet, right?

00:52:25,538 --> 00:52:29,688

And all of a sudden, and at that point

I'm like, whoa, whoa, whoa, whoa.

00:52:30,558 --> 00:52:35,218

We, we, you know, you given, you give

a little bit and then all of a sudden

00:52:35,218 --> 00:52:37,168

they want to, you want to like take it.

00:52:37,783 --> 00:52:39,013

and do something else with it.

00:52:39,013 --> 00:52:42,733

So instead, you know, in that

particular case, I was like, we should

00:52:42,733 --> 00:52:45,343

just not even call this anything.

00:52:45,343 --> 00:52:46,483

Not even give an option.

00:52:46,483 --> 00:52:48,763

Just start with the simplest.

00:52:49,583 --> 00:52:53,763

Maybe that's where this leads, but

it just confuses people when you're

00:52:53,763 --> 00:52:56,953

kind of like halfway there almost

with a feature and it, and it didn't.

00:52:57,173 --> 00:53:00,613

So anyway, I've just, I've had

that example in my mind recently.

00:53:00,613 --> 00:53:04,413

So I'm just interested in how

other people think about those

00:53:04,413 --> 00:53:07,483

kinds of problems and you know,

product development's hard, right?

00:53:07,483 --> 00:53:11,303

Right.

00:53:11,518 --> 00:53:13,768

Phillip Miller: there's, you

know, the classic MVP, right?

00:53:13,768 --> 00:53:14,878

Minimal Viable Product.

00:53:15,338 --> 00:53:20,358

Um, like you, you're trying to do just

enough to, to be, to be what it is.

00:53:20,658 --> 00:53:25,218

There is danger in like providing

it half baked or, or like, or, or

00:53:25,248 --> 00:53:29,968

like instead of, um, people don't

know what to do with it then.

00:53:30,358 --> 00:53:33,458

And they start taking it in ways

that, that you don't really want to

00:53:33,768 --> 00:53:36,368

Randall Stevens: Well, yeah,

and you can't ever stop that.

00:53:36,378 --> 00:53:39,628

You can't keep people once it's in their

hands, they're going to do whatever.

00:53:39,628 --> 00:53:43,268

But I think back to what we were

talking about earlier, which is there's

00:53:43,268 --> 00:53:45,528

a philosophy around why we did this.

00:53:45,528 --> 00:53:46,608

And that's the story.

00:53:46,608 --> 00:53:47,628

That is the use case.

00:53:47,628 --> 00:53:50,218

You may try to go solve

another problem with it.

00:53:50,628 --> 00:53:54,523

Um, You know, which is fine, but

I don't know, there's, there's

00:53:54,523 --> 00:53:58,303

a fine line there between like,

this is what it's supposed to do.

00:53:58,313 --> 00:54:01,623

You're trying to use it in a way that's

not really, it wasn't designed for.

00:54:01,943 --> 00:54:04,523

We'd love to have that feedback

because that's where you now

00:54:04,543 --> 00:54:06,233

get your next best things.

00:54:06,233 --> 00:54:08,903

We're like, ah, now, now

we know what to do next.

00:54:08,903 --> 00:54:14,008

And that, that gets into this roadmap

of like, okay, We can kind of envision

00:54:14,018 --> 00:54:16,458

that this may be where we want to

take this, but we've got to like

00:54:16,468 --> 00:54:20,758

pull back now and say, uh, but then

not make it confusing for the user.

00:54:20,768 --> 00:54:23,298

Simple, simple, fewer

buttons, fewer choices.

00:54:23,698 --> 00:54:24,028

Like,

00:54:26,053 --> 00:54:27,607

Phillip Miller: Yep, mm hmm.

00:54:27,607 --> 00:54:28,243

Mm

00:54:28,298 --> 00:54:32,628

Evan Troxel: I'm curious from like a

user standpoint, do you define personas

00:54:32,668 --> 00:54:35,158

for different types of users up front?

00:54:35,538 --> 00:54:39,038

And I'm also curious kind of

about the way you see workflows

00:54:39,048 --> 00:54:40,558

happening with software like this.

00:54:40,668 --> 00:54:44,558

You've talked plenty about iterative

design process of the software, but

00:54:44,558 --> 00:54:48,708

obviously the architectural visualization

and animation is also iterative.

00:54:48,788 --> 00:54:54,158

Designs are always changing, so when

you're defining the end user, and and

00:54:54,158 --> 00:54:59,018

where this product fits in your ecosystem,

obviously you, it's a separate thing

00:54:59,018 --> 00:55:03,818

from Enscape to V Ray, and it kind

of sits in between those two things.

00:55:03,868 --> 00:55:08,368

And, it's a standalone app, so it's no

longer a plug in for something else,

00:55:08,368 --> 00:55:10,038

and so you have this hub and spoke.

00:55:10,038 --> 00:55:12,368

There's a lot of things,

moving parts here,

00:55:12,788 --> 00:55:17,978

and I'm curious how you, how you defined

your, your user in this, for this piece

00:55:17,978 --> 00:55:18,618

of software.

00:55:19,373 --> 00:55:25,193

Phillip Miller: Yeah, so we we identified

three personas that we wanted to address

00:55:25,753 --> 00:55:30,923

Um, now it's not like those are, it's

not like they're different skins or

00:55:30,923 --> 00:55:34,923

anything like you don't walk in and like

change the product for that one persona.

00:55:35,383 --> 00:55:37,903

It's more like it's the same

product serving them all,

00:55:37,983 --> 00:55:42,613

but you want to make sure that the way

they use the product is the way that

00:55:42,913 --> 00:55:45,353

will, will work the way they need it to.

00:55:45,753 --> 00:55:48,143

And so the workflows become different.

00:55:48,263 --> 00:55:52,053

And so the three, those three personas

defined really three different workflows.

00:55:52,743 --> 00:55:59,838

Um, and, um, We always use that as

a test for, okay, that person, like

00:55:59,838 --> 00:56:04,128

for example, the Enscape person, the

Enscape person needs it to just move

00:56:04,128 --> 00:56:09,228

over, come over real easily, he can't

be confused, he has to begin where he

00:56:09,228 --> 00:56:17,268

left off, we can't intimidate them,

it has to be familiar, he can increase

00:56:17,268 --> 00:56:19,078

it over time, but he doesn't have to.

00:56:19,358 --> 00:56:23,698

So that's like one workflow, but then

also like, why is it coming over here?

00:56:24,113 --> 00:56:24,463

Right?

00:56:24,893 --> 00:56:28,093

Well, he's probably coming over

because he wants more animation, or

00:56:28,103 --> 00:56:31,203

his file got too big, or he needs to

coordinate between different products.

00:56:31,763 --> 00:56:35,143

Um, you know, all those sort of things,

or maybe he's, you know, wants to

00:56:35,143 --> 00:56:38,323

add a lot more architecture than is

appropriate to have in the design file.

00:56:39,413 --> 00:56:44,023

Those are all four things that, that

we are really common for that, for

00:56:44,023 --> 00:56:46,013

that persona to, to be caring about.

00:56:46,878 --> 00:56:50,448

Um, and so that's what we want

to make sure is all possible.

00:56:50,858 --> 00:56:54,638

Um, and then that finally gets back

into how we describe the product.

00:56:54,838 --> 00:56:58,968

We, we describe it in, in those

terms for, for those use cases.

00:56:59,538 --> 00:57:02,538

So, we, we always talk about it

from the standpoint of like the

00:57:02,538 --> 00:57:04,138

workflow, is that workflow working?

00:57:04,468 --> 00:57:09,308

Um, and then of course in the eventual

marketing and, and even documentation,

00:57:09,678 --> 00:57:12,898

uh, we're, we're trying to speak to

those different, different personas.

00:57:13,058 --> 00:57:13,258

Mm

00:57:14,038 --> 00:57:18,618

Evan Troxel: So when you're Actually

testing this out with those people.

00:57:18,638 --> 00:57:22,288

Are you going through a real

architectural workflow process with them?

00:57:22,298 --> 00:57:26,348

Where, because this is a standalone

app, it sounds to me like you can

00:57:26,348 --> 00:57:28,548

swap out geometry whenever you want.

00:57:28,558 --> 00:57:28,858

And this

00:57:28,868 --> 00:57:32,488

kind of sits in a layer separate from the

geometry, I think is what you were saying.

00:57:32,498 --> 00:57:32,758

It's kind

00:57:32,758 --> 00:57:37,008

of a, a metadata kind of a

analogy in my mind, at least.

00:57:37,018 --> 00:57:37,238

So

00:57:37,658 --> 00:57:40,763

if, if there, are they actually going

through that and able to test it?

00:57:40,873 --> 00:57:44,853

to test that out and, and because that

to me is what the design process is.

00:57:44,863 --> 00:57:49,773

Like I remember we had viz guys and those

viz guys would complain every single time

00:57:49,773 --> 00:57:51,833

we wanted to update the model and, and,

00:57:51,833 --> 00:57:54,873

and it's like, and, and to me it's

normal and to them it's a headache.

00:57:54,873 --> 00:57:55,063

Right.

00:57:55,063 --> 00:57:59,643

And so I imagine you're trying to

create something that allows for that

00:57:59,643 --> 00:58:01,313

to happen very easily since it is

00:58:01,313 --> 00:58:05,633

separate from the main modeling program,

knowing that that happens in architecture,

00:58:05,973 --> 00:58:10,373

but still kind of like taking away

the friction in that process is that.

00:58:10,718 --> 00:58:11,858

Is that accurate to say?

00:58:12,288 --> 00:58:12,828

Phillip Miller: Yeah, it is.

00:58:12,828 --> 00:58:15,588

I mean, like, for example, the one

persona is like, okay, it's the

00:58:15,588 --> 00:58:18,598

person that was actually doing the

design, wanting to take it farther.

00:58:19,528 --> 00:58:22,048

And then it's the other very

common situation where he's

00:58:22,048 --> 00:58:23,718

handing it off to a specialist.

00:58:24,318 --> 00:58:27,078

And now that specialist needs to

suddenly understand that file.

00:58:27,178 --> 00:58:30,988

And, and, what we wanted to do

was minimize how many questions

00:58:30,988 --> 00:58:32,598

got asked back and forth.

00:58:33,798 --> 00:58:38,208

Uh, so not only did it evolve, but,

but, um, the design intent came over.

00:58:39,103 --> 00:58:43,233

Um, uh, but design intent, I mean, like

all the materials, all the, uh, all the

00:58:43,243 --> 00:58:47,313

entourage, everything else, you know,

came over as, as it, as it started.

00:58:47,743 --> 00:58:51,263

Um, but then we needed a ways for

that person to be able to communicate

00:58:51,283 --> 00:58:53,543

back to the person that handed it off.

00:58:54,093 --> 00:58:58,103

Um, so that became the second

persona, which is really kind

00:58:58,103 --> 00:59:02,683

of like the specialist, um, uh,

that workflow had to happen too.

00:59:03,968 --> 00:59:07,858

Now suddenly we don't care too much about,

well, is he, we don't have to worry about

00:59:07,858 --> 00:59:11,578

the complexity for this person because

he's going to be spending a lot of time

00:59:11,578 --> 00:59:13,718

in that product and, and, and knowing it.

00:59:14,098 --> 00:59:18,188

Um, but, um, but he's got this

communication and this iteration

00:59:18,248 --> 00:59:22,698

thing that he has to accommodate

because he doesn't know the file as

00:59:22,698 --> 00:59:24,288

intimately as the person who made it.

00:59:24,948 --> 00:59:26,228

Um, so

00:59:26,673 --> 00:59:29,213

Randall Stevens: is the

architecture of that fill a pull

00:59:29,213 --> 00:59:32,313

or a push from the, from the host

00:59:33,223 --> 00:59:34,093

into the hub?

00:59:34,513 --> 00:59:35,213

It's a push.

00:59:35,653 --> 00:59:41,813

So if somebody is doing new work,

they've got to push the update to the

00:59:42,823 --> 00:59:43,463

Phillip Miller: Right, right,

00:59:43,763 --> 00:59:46,883

they have to save, I mean, uh, if

they're using the same file name,

00:59:46,893 --> 00:59:49,713

it'll just automatically update, um,

00:59:49,953 --> 00:59:54,353

Randall Stevens: So if you, so if,

if you launch Envision, it's going

00:59:54,353 --> 00:59:58,113

to pull from whatever the source

is, as long as it's always the same

00:59:58,208 --> 00:59:59,228

Evan Troxel: Like a link, a linked

00:59:59,578 --> 00:59:59,708

Randall Stevens: Got

00:59:59,708 --> 00:59:59,848

it.

00:59:59,968 --> 01:00:00,238

Yeah.

01:00:00,373 --> 01:00:02,063

Phillip Miller: it's a

reference file, yeah.

01:00:02,088 --> 01:00:02,328

Randall Stevens: Yeah.

01:00:02,918 --> 01:00:03,188

Cool.

01:00:03,448 --> 01:00:03,678

Yeah.

01:00:03,683 --> 01:00:04,923

Phillip Miller: Yeah, so, but,

01:00:04,978 --> 01:00:07,933

um, yeah, that, but, it's

really the workflow is where

01:00:07,933 --> 01:00:09,143

the personas come in the most.

01:00:09,203 --> 01:00:13,863

It, to, uh, and then we use that

as the, as the real test cases for

01:00:13,873 --> 01:00:15,543

making sure that things are working.

01:00:16,358 --> 01:00:16,568

Randall Stevens: Yeah.

01:00:17,258 --> 01:00:21,408

So is the, is the product out yet or

where, where are you in the cycle?

01:00:21,683 --> 01:00:23,633

Phillip Miller: Well, we're

going to start beta next Tuesday.

01:00:23,633 --> 01:00:25,798

Yeah,

01:00:25,868 --> 01:00:29,198

Randall Stevens: by the time this is

out, you know, how will people be able

01:00:29,198 --> 01:00:31,248

to see this or where can they learn more

01:00:31,343 --> 01:00:36,673

Phillip Miller: Yeah, so starting

on November 12th, or maybe 13th, the

01:00:37,133 --> 01:00:40,223

people will be able to come to Chaos.

01:00:40,393 --> 01:00:44,993

com and sign up for the beta, download

the software, they'll have a beta

01:00:44,993 --> 01:00:47,613

license, it'll keep running for

the length of the beta and a little

01:00:47,613 --> 01:00:51,213

bit thereafter, and they're free

to use it as much as they'd like.

01:00:51,703 --> 01:00:52,758

Um, we even,

01:00:52,963 --> 01:00:57,103

we even produced independent

exporters, um, uh, to, to reach it

01:00:57,103 --> 01:01:01,273

as well so that people don't, and

people using SketchUp and Rhino for

01:01:01,273 --> 01:01:06,553

example, can, they don't need to ins

to install like B Ray or, or Inscape.

01:01:06,613 --> 01:01:10,423

They can just, uh, be exporting

directly from that, from that product.

01:01:10,933 --> 01:01:11,443

Um.

01:01:11,843 --> 01:01:12,453

right away.

01:01:12,833 --> 01:01:14,693

So those will be available