2019 fail?

[kris]

So we have a taste, ahead of tomorrow's release.

https://help.autodesk.com/view/3DSMAX/2019/ENU/?guid=GUID-40222876-7E08-42F8-A912-F6CDC597A3FD

https://area.autodesk.com/blogs/the-3ds-max-blog/meet-3ds-max-2019/

(etc)

Is it just me, or is this a major anticlimax?  Since when is something so trivial as spline booleans a major feature?  How is a material that only works in Arnold (does anyone use that?) and scanline worth anything at all?

Where's the list of major bugfixes?

Where's the list of core improvements that will affect stability and usability, like multithreading and double precision?  

Where are the things that make this worth a full version update?

Kris.

[dgorsman]

Multi-threading isn't a magic wand (it's not vanilla, it doesn't make everything better just by adding it), and implementing it isn't as simple as clicking a toggle.  I'd make the argument that adding multi-threading would make the program less stable, more complex, and more prone to unsolvable bugs.

Double-precision math is irrelevant for pretty pictures (even 3D printed ones).  It would only come into play for physically precise simulation such as stress analysis, which you wouldn't be using 3DSMAX for.

[Anonymous]

I think that as a subscription Autodesk doesn't rely any more in mayor releases to deliver features and they are scattered across the year. Perhaps they should stop labelling it as such, as Adobe does, to avoid false expectations.

Another problem is that being a program that covers a vast field important features, as the liquid simulation, can be irrelevant to many of their users. It is not about this feature is not important, but that the specific field you work is not the one enhanced.

[kris]

@dgorsman While I agree that I may have unfairly used multithreading as a blanket term, I mean it thusly: there are modifiers and other features within max that are single threaded, to the immense detriment of my time.  Now that speed is coming in the form of additional cores (I have 10 physical in this machine), sitting around and waiting for a single core to do something for 20 minutes is immensely frustrating.  Obviously there are parts of max which do not need to be multithreaded because they are millisecond single thread operations.

As far as double precision, we don't all peddle fluffy pieces of art.  Have you ever tried importing geo-located data?  When your import comes in 6000 km from the orgin, it's a total nightmare.  How about something that is scaled large (like a factory that is 6km long?  How about large scale point clouds?  Ever dealt with a scene that is more than 3km across, but you need to have features at the cm scale in the foreground?  Or trying to line something up for an engineering purpose, and you end up with a gap because max rounds the value to some arbitrarily weird value.  Yeah, there are workarounds, but they take enormous amounts of time to bend over backwards to fuss with because of the single precision of max.   

Kris.

[spacefrog_]

Hey Kris, why don't you join the beta ?

The problems like the one you mention come up regularly in real world use cases and i think the beta would give your voice and input some addional weight.  Besides that, i don't think that Max 2019 is a "fail", far from it. Of course many features were already introduced through the various Max 2018 updates. I think this is one side of Autodesk trying to implement some continuous delivery scheme, but of course as usual missing the essential parts, like for example some benefit for the user ...

[kris]

@spacefrog_  I was on the beta, onceuponatime...  It was an immensely frustrating experience.

I can't say the point updates were exciting.  I've found stability has been more flaky since I installed .3 and .4...  Though, with max, that's really hard to qualify.

Kris.

[spacefrog_]

---

@kris wrote:

@spacefrog_  I was on the beta, onceuponatime...  It was an immensely frustrating experience.

---

Understood, but you might reconsider.  I'm on the beta since 2009, and i have to say that since then things have changed to the better. Of course some limits still apply, but nevertheless i think your input would be a valuable addition

Of course i can't say anything about you getting an invite or something

BTW: i find Max 2018.4 exceptionally stable, main bug for me being the dialog focusing bug ( have to click 2x on some dialogs to accept mouse and keyboard input),

[-niels-]

---

@kris wrote:

How is a material that only works in Arnold (does anyone use that?) and scanline worth anything at all?

---

Hey, i use Arnold and i like it!

Just hope there are more updates for it then a wood material.

---

@kris wrote:

https://area.autodesk.com/blogs/the-3ds-max-blog/meet-3ds-max-2019/

Where's the list of major bugfixes?

---

Nearly at the bottom of that blog post you linked is a link to the release notes, is that what you were looking for?

[Anonymous]

meanwhile in houdini 😕

https://vimeo.com/244246886

[-niels-]

---

@Anonymous wrote:

meanwhile in houdini 😕

https://vimeo.com/244246886

---

Drooling at that viewport quality, would love to see what i'm doing without having to use an activeshade window in Max as well...

Maybe i'll give Houdini a try since they say it's free for hobbyists, though it won't work with Arnold or other 3rd party renderers in their free version....

[kris]

Overall, Houdini is where it's at.  It's the closest to max, while being ultimately way more powerful.  It's missing a few things, though, and poly modelling isn't nearly as fast as max. 

What concerns me about the "dribble out improvements in subscription" isn't that there is a more constant flow.  I think that's a good thing.  What concerns me is that the ~quarterly output makes it much more difficult for the dev team to write broad stroke core changes that might take months (if not years) to execute.  Let's not forget that Max is still built on an archaic core, and there are still hundreds of bugs.  

Having taken the time to get to know modo, houdini and C4D, I can safely say that yes, max is the best generalist package.  But each one of those packages whips max into a bloody mess in several areas, and each one of them is going through significant improvements with time. 

I guess time will tell.  I'm underwhelmed with 2019, even when one considers the point releases in the past year.

Kris.

[kris]

Houdini also has the indie license, which is quite reasonable, if you fit in its criteria. (I do.)

Kris.

[Anonymous]

Unless you are in FX or animation (that then you probably would prefer Maya) poly modelling is a very basic feature to be ignored. 

My guess is that is possible yet to do structural changes with incremental releases. Adobe does that with all his products, also Algorithmic. And I perhaps the 2018 interface Update of Max is also that. This style of updates seems that come to stay as it is becoming the dominant one.

[kris]

@Anonymous I don't think it's ignored.  It's just slower since it's completely procedural on a per-operation level.  I was able to build out a complex object no problem in Houdini.  Where it failed for me was on splines, using splines as building blocks for geometry and the like.  That said, I only spent a week with it intimately, and there is a *lot* that I don't know.  

I prefer Houdini over modo, in spite of its limitations.

Kris.

[Anonymous]

Will it function with a high resolution laptop screen and Windows 10 scaling like 2017 does? I never bothered with 2018 because it was completely unusable on my laptop. Have to say I've been using Max for going on 15+ years now and seriously consider switching to something else pretty frequently these days. Seems clunky when I see what people are doing in Cinema 4D or Houdini.

Advertising "new wood shader" as one of the main improvements is pretty sad.

[Anonymous]

This video shows that OSL shaders is a lot more impressive that it sounds.

[brentscannell]

Thanks @Anonymous for posting that

Here's another video discussing OSL in 3ds Max 2019 in more detail:

https://www.youtube.com/watch?v=SJZWAqlzqS8

And the Advanced Wood material:

https://www.youtube.com/watch?v=KoVsXAZra2g

[Kranos_Orphydian]

you realize that Cinema 4d updates are not like they were several years ago. Their software is limited at any corner and utterly lacks capability in so many areas that makes it suitable only for Basic generic motion graphics like  passing and effector through  a line of text or colored ball animations. Thier viewport is slow as hell, (kinda ironic for a motion graphics app) no production render, UV tools (just more than nothing), modeling tools ( has the basic ones just for the people cant say  they dont have them),  broken open subdiv support, broken FBX exporter (compeltely unsuitable for games), crappy hair generator, their particles are a joke for a motion graphics app etc, no liquid simulation etc.

Seriously there is no other software depending and in the same time pandering to 3rd party plugins developers. Expectations are so low for their next  " big"  10 version C4D 20 that you kinda feel sorry for them.

[Anonymous]

There is too much in max, that needs fixing, to convert it into modern platform.

I doubt it's possible. It's a high level operation software with obsolete maxscript.

There are attempts to make max better, like max creation graph based on external language and data channel. But it will face structural limits.

MCG misses good tutorials and real time scene control.

Also people wants node based management system instead of schematic view, wire manager, parameter collector, etc.

Faster tools to streamline production in certain areas, dynamic booleans, baking, better uv and counting.

In today's state max is nice app for a very primitive animation, but mostly for rendering static scenes.

But even in static scenes it's missing approximate viewport feedback (mostly lighting) for photorealistic renders and native tools, like scatters, ids management, mass parameters editing, randomizes etc. VR and 4k will shift this area to game engines.

So, it's a an obsolete single threaded platform with a bunch of plugs over it, which has different interfaces and ideology.

May be it's a reason why there is no new features: it is too hard to code them this platform.

My opinion: may be it's easier to create a new platform with external plugs, instead of rewriting old one. Not just because computers will have more cores, but also because industry standards applies faster adaptation for new requirements and cross software pipelines.

[Anonymous]

I guess until appears that wonderful 3d program we will continue using Max.

Is there any program that at the end doesn't become victim of his own complexity and history?. Unless is a brand new program and in that case is normally very thin in features.