Mentalray EOL

News concerning 3D DCC business
Bullit
Moderator
Posts: 2621
Joined: 24 May 2012, 09:44

Re: Mentalray EOL

Post by Bullit » 21 Nov 2017, 22:27

but any of that was at best completely drwafed by other pretty extra-outrageous decisions in regards to essentially the only other long standing general purpose DCC
As such? give some examples. In my opinion XSI would be dead well before 2010 if wasn't for ICE.

luceric
Posts: 1251
Joined: 22 Jun 2009, 00:08

Re: Mentalray EOL

Post by luceric » 21 Nov 2017, 23:18

Softimage doesn't have much to do with the market some of you talk about, Max Viz, Design and VRay. We were fighting Maya. 80% of the Softimage seats were in games and post (places like The Mills, Hybride, Animal Logic), with only the remaining 20% being freelancers. It's a character animation product and we only cared about that market, regardless of if you could use it for viz and realistic renderings.

A typical Softimage studio had from 100 to over 300 seats, few of them doing any rendering. The purported exodus due to rendering quality looks like message board bias. I have no doubt a few people left, but that's on thousands of seats. The message boards are full of butterfly that download every product and compare specs all day. Nobody is going to put themselves out there to defend mental ray or any "meh" market leader, they'll just use it quietly. In fact, most people do not post at all.

User avatar
rray
Moderator
Posts: 1774
Joined: 26 Sep 2009, 15:51
Location: Bonn, Germany
Contact:

Re: Mentalray EOL

Post by rray » 21 Nov 2017, 23:19

So it could be seen from this angle too: Would we even have gotten ICE had the XSI devs not to fight Maya?
softimage resources section updated Jan 5th 2024

User avatar
mc_axe
Posts: 415
Joined: 12 Mar 2013, 18:44

Re: Mentalray EOL

Post by mc_axe » 22 Nov 2017, 00:31

MR had its pick in the 90s with only rival being pixar and maybe lightwave for commercials right? MR was somehow legendery for preey long time. I think that many of the starters hobbyists had in mind that MR is crazy and it wasnt so simple to accept that anyone can beat that, so i can see alot of discussions that take place in 2006? that ppl still cant let go MR. Personally I had no knowledge of 3rd party solutions at all, to compare softwares i was looking in to the default package and what it offers. As a Maya fan at 2003 or so, i found a comparison on paper from a magazine and Softimage was just too good, MR bullets killed me that day, esp Brute force GI and a faster GI finalgathering that even mortals like me could use in one PC, i was so sad i had to migrate to access that kind of quality. Before that i was on 3dsmax rendering with 1 core.And before that 3ds msdos lol.

I felt preety satisfied when i actually got a nottice that i could install MR in new demo of Maya my first rendering plugin, but it was too late i alrdy had fallen in love with SI mental ray ice UI everything, i never looked back to 3ds cause it felt years behind in everything else. But that was not so smart cause i entirelly missed the V ray bandwagon, and other solutions (that till GPU rendering got my attention).

Now i can see that VRay is definately the MR killer and a software that changed rendering in total.
I dont feel that VRay contributed alot in the demise of Softimage (maybe it was a very small factor) even if it was pain in the bum to chage the software to accept other solutions.
That is because Maya was also without Vray till 2009 aswell, and it appears that Vray was hot since 2003 so Maya should be also demised?

But both SI and Maya was way better at other things, both got much cheaper after 2000. And non died for that reason, of course Max got huge crowds with Archviz.
I never felt hanging without a good rendering solution i was working at the time 2004+ in a small studio in small projects with 5 nodes renderfarm and it felt crazy to me. But this is just my POV.

User avatar
Mathaeus
Posts: 1778
Joined: 08 Jun 2009, 21:11
Location: Zagreb, Croatia
Contact:

Re: Mentalray EOL

Post by Mathaeus » 22 Nov 2017, 00:37

luceric wrote: 21 Nov 2017, 23:18 Nobody is going to put themselves out there to defend mental ray or any "meh" market leader, they'll just use it quietly. In fact, most people do not post at all.
Well they don't use it anymore, or not for long. According to nVidia decision, it seems that your silent majority quietly chose to *do not* proceed with MR, as soon as it was removed from Maya. They used it because it was available, that's all. Does not tell anything about 'quality'. Considering that they have some project already done with MR which needs some update, now it seems that they left it completely, two or three years ago.

And, if Softimage really competed against Maya, did team really believed that best connection to MR and only MR was enough to fight against all internal or external Maya renderers.

Bullit
Moderator
Posts: 2621
Joined: 24 May 2012, 09:44

Re: Mentalray EOL

Post by Bullit » 22 Nov 2017, 01:03

luceric wrote: 21 Nov 2017, 23:18 Softimage doesn't have much to do with the market some of you talk about, Max Viz, Design and VRay. We were fighting Maya. 80% of the Softimage seats were in games and post (places like The Mills, Hybride, Animal Logic), with only the remaining 20% being freelancers. It's a character animation product and we only cared about that market, regardless of if you could use it for viz and realistic renderings.

A typical Softimage studio had from 100 to over 300 seats, few of them doing any rendering. The purported exodus due to rendering quality looks like message board bias. I have no doubt a few people left, but that's on thousands of seats. The message boards are full of butterfly that download every product and compare specs all day. Nobody is going to put themselves out there to defend mental ray or any "meh" market leader, they'll just use it quietly. In fact, most people do not post at all.
My point was not exodus but lacking appeal, ICE brought appeal again.
Yes that is the problem with Softimage, for a long time only focused in small part of the market that ultimately is judged by the quality of its final image. I can guess the discussions to advance with ICE vs continuing in same route in Softimage HQ.*
XSI never had that many users since it arrived late to the party compared to Maya. Also didn't followed Maya when Maya dropped its price. So it was a product that came from the aristocracy so to speak, not from the bottom users. That is why PR acts like Softimage Foundation mostly failed. If Foundation when was released it had Vray available history would have been different.

If Vray does not matter why ultimately Softimage got Vray and Softimage developers had to do hard work to get API/SDK fixed for Vray?

*Someone should make a book on histories of these applications.

Bellsey
Posts: 688
Joined: 19 Apr 2010, 11:50
Location: London, UnitedKingdom
Contact:

Re: Mentalray EOL

Post by Bellsey » 22 Nov 2017, 01:11

MR was both a blessing and a curse for Softimage (XSI). MR had alway been around in Soft, even in the Soft3d days and while PRman was the main choice for film, many used Mental Ray for everything else. If you knew what you were doing, it was a great renderer.
The MR integration in XSI was second to none, and we have Halfdan to mainly thank for that I(Luc-Eric can correct me here no doubt). The amount of work and code put into that was insane.
The integration was so good though, that I think many believed that MR was the one and only renderer for XSI. Max on the other hand had a very open API for plugins, so it attracted alot of plugin makers and renderers. And this was where Vray started to gain ground. Compared to MR its was kinda simple to to get good results and it was also very fast. MR was still seen by some to be too complex and in some ways it was.

When I joined Soft in '07, there was alot of effort made to change the perception of other renderers in XSI, and at Siggraph we pushed alot on this with name drops like Vray, Maxwell, 3Delight, and Arnold. Nvidia bought MR then things seemed ok but it wasn't long after that some people started to loose confidence and certainly from a Soft view many started to jump to Arnold. And outside Japan, Soft's market was mainly commercials and short form TV. Within a few years everyone had gone to Arnold and Redhsift, those on MR did so because they had long running projects to maintain.

Despite that, MR was still being used a fair bit. There were some big names who still used MR for some stuff, even though they had Arnold and Redshift.

luceric
Posts: 1251
Joined: 22 Jun 2009, 00:08

Re: Mentalray EOL

Post by luceric » 22 Nov 2017, 03:18

Mathaeus wrote: 22 Nov 2017, 00:37
luceric wrote: 21 Nov 2017, 23:18 Nobody is going to put themselves out there to defend mental ray or any "meh" market leader, they'll just use it quietly. In fact, most people do not post at all.
Well they don't use it anymore, or not for long. According to nVidia decision, it seems that your silent majority quietly chose to *do not* proceed with MR, as soon as it was removed from Maya. They used it because it was available, that's all. Does not tell anything about 'quality'. Considering that they have some project already done with MR which needs some update, now it seems that they left it completely, two or three years ago.

And, if Softimage really competed against Maya, did team really believed that best connection to MR and only MR was enough to fight against all internal or external Maya renderers.
You're working too hard to defend your false thesis. Softimage was competing against Maya as a character animation tool. Animation seats do not render at all and mental ray or not has nothing to do with competing with Maya
There are hundreds of animation seats to one rendering seat.

luceric
Posts: 1251
Joined: 22 Jun 2009, 00:08

Re: Mentalray EOL

Post by luceric » 22 Nov 2017, 03:36

Bullit wrote: 22 Nov 2017, 01:03 If Vray does not matter why ultimately Softimage got Vray and Softimage developers had to do hard work to get API/SDK fixed for Vray?
Softimage added shaderballs, and we wanted to have opengl and direct X real-time renderers, and the mental images APIs and shaders were changing drastically.

Halfan wanted to get it of old code, and make it easier to integrate drops of Mental Ray, and Softimage wanted to have the option to remove mental ray from XSI.

None of this really has anything to do with trying to battle in the offline rendering market. At best, it's maintenance work for existing clients, with the bonus of potentially getting third parties.

Softimage had at that point absolutely zero interest in fighting in the offline rendering market. Softimage is not Lightwave or Modo, or Max. It thought of itself as a pipeline tool in games or high end post, and the freelancers were just legacy gravy.

Bullit
Moderator
Posts: 2621
Joined: 24 May 2012, 09:44

Re: Mentalray EOL

Post by Bullit » 22 Nov 2017, 05:09

Well i know it was Softimage and later Autodesk when Softimage went in that pushed for Vray to Softimage promising and giving assurances to Vlad that they wouldn't drop it. No wonder that Vlad had a justified tantrum when everything crashed down and his investment were for nothing. I also remembering before Vlad complaining of workarounds to port Vray to Softimage and Softimage then making changes to answer that and make his work easier.

luceric how do you explain ICE, it wasn't certainly for character animation work. It was a decision to branch the application to more areas.

User avatar
Mathaeus
Posts: 1778
Joined: 08 Jun 2009, 21:11
Location: Zagreb, Croatia
Contact:

Re: Mentalray EOL

Post by Mathaeus » 22 Nov 2017, 10:17

luceric wrote: 22 Nov 2017, 03:18

You're working too hard to defend your false thesis. Softimage was competing against Maya as a character animation tool. Animation seats do not render at all and mental ray or not has nothing to do with competing with Maya
There are hundreds of animation seats to one rendering seat.
I'm just considering the *facts*, that today, Softimage and Mental Ray are EOL. While for example, V-Ray is still alive, more than ever. Something obviously went wrong with MR.
So what do you think, what went wrong, exactly with Mental Ray? - just as "today" EOL. We were already talking about SI EOL enough, I think.

Bellsey
Posts: 688
Joined: 19 Apr 2010, 11:50
Location: London, UnitedKingdom
Contact:

Re: Mentalray EOL

Post by Bellsey » 22 Nov 2017, 15:00

I don't really see any correlation between the Mental Ray EoL and Softimage EoL.

Luc-Eric is right, much of the core evangelising for Softimage was around animation and pipeline. Everything else was really a bonus.
It allowed us to position Softimage very well to Maya/Max people as an augmented solution. In my time (XSI 6.x onwards), I rarely did any modelling or texturing demos. All the main ones were around animation, MOTOR< GATOR, Delta refs etc. There was some rendering stuff too, but more enhances to existing features.

ICE, yes it wasn't for animation work. If memory serves it started as the replacement for the particle system but in the process became much more.
Imo, the focus was really about the next evolution if you will to creating tools. While MEL allowed for a technical to user to easily create tools for Maya, ICE allowed the same but in a visual way allowing more artist types to experiment as well. It certainly helped turn heads back in the direction of Softimage. Houdini aside, it paved the way for others to kinda follow, or improve their own node based systems. I look at UE4's blueprint and see so many parallels to ICE.
Also, as a side note remember how painful it was to really render anything from ICE?

Where did it go wrong for Mental Ray? Well its kinda simple really, like anything, unless you evolve and more with the times, you die. And post acquistion, nvidia doidn't seem to really push it. iRay was looking interesting, but other GPU solutions were emerging which I think just seemed better and more adaptable.

luceric
Posts: 1251
Joined: 22 Jun 2009, 00:08

Re: Mentalray EOL

Post by luceric » 22 Nov 2017, 16:05

Iray and other related tech are not dead. Just the mental ray product

User avatar
FXDude
Posts: 1129
Joined: 19 Jun 2012, 21:59

Re: Mentalray EOL

Post by FXDude » 23 Nov 2017, 17:47

Hi sorry for the delayed response, wrote that yesterday evening, as I'm on the clock.



__________________
For MR EOL, At first glance, or maybe a factor among other things..

Unlike something like SI (dependant on OS)
MR is highly dependant on host app, and while Maya/3DSMax are still struggling in removing MR (built-in over a very long time),

... it seems like MR was stuggling in making their plugin re-integrate everything to the same previous (inevitably compared to) level,
also built-up over a very long time, in now very unwelcoming host apps trying to remove all traces of it.

The MR forum is full of integration issues, and it seems the new MR wasn't the same as the old.

Maybe other factors, but mere fact that it was let go while previously being included,
apart the previously mentioned points, the reasons to actually buy what was previously included would have had to be extraordinary.
(and actually were quite extraordinary for some things.)

__________________
Bullit wrote: 21 Nov 2017, 22:27
but any of that was at best completely drwafed by other pretty extra-outrageous decisions in regards to essentially the only other long standing general purpose DCC
As such? give some examples. In my opinion XSI would be dead well before 2010 if wasn't for ICE.
Hum... XSI was litterally on a roll just before getting snatched.
    (I still have some job stats somewhere)

Or after perhaps getting introduced late, it was most-definitely catching up...
(in terms of usage.. it was already there in terms of features)
that's until it got snatched-up.

________________________________
If there's any doubt as to why any exodus happened...
(otherwise pretty self-obvious.. but sometimes avoided)

Part of XSI's acquisition press release ::

"The acquisition is meant to strengthen Autodesk's position in the fast-growing video games market"

Quote from the list (2016):
Softimage, which also as a product that "owned" the japan game market, was immediately put in the Autodesk Games group.
Softimage's managers were made leaders of that.
Former softimage dev immediately went on to worked on an ICE-like game middleware, project skyline,
while softimage game customers were gradually migrated to Maya.

If Arnold aquisition press release said ::
     "The acquisition is meant to strengthen Autodesk's position in the fast-growing particle rendering market "

and all Devs got repurposed to particle rendering,
then there would have been serious questions about Arnold's future, and people would naturally start looking elsewhere
while others would have went on using it because Arnold is also pretty good as it was.

--> So I think it settles the question of when Softimage actually died,
as it wasn't just Softimage game customers that were getting  'gradually migrated to Maya'  from that first moment on.

And the very first intent was pretty clearly the gradual removal of XSI from the map,
(a prominent competitor, with a brand new game-changing visual programming language)
extra dev resources was just possible extra gravy, mostly ending-up all somewhere else by the next year,


And we can understand why people started leaving right there, Kim Aldis comes to mind..
more when we found-out from the outskirts that any remaining dev went to Sigapore..
and obviously lots more when it eventually got cancelled.


Despite being a smaller user base than standard Maya
(still larger than say Houdini's which also wasn't insignificant)...
luceric wrote:A typical Softimage studio had from 100 to over 300 seats ..
80% of the Softimage seats were in games and post (places like The Mills, Hybride, Animal Logic),
.
+ MPC.. Glassworks.. Whiskey Tree.. Janimation.. RodeoFX.. NerdCorps ... PsyOps, Seed Animation Studios ... ... ...
some having to rethink entire pipelines, reshuffle entire personnel
with the remaining 20% being freelancers
like many many small shops if not a large portions (?) of VFX small shops (?)
because of it's out of the box no-nonsense nature?

In all a couple thousand users, many of which worked with XSI their entire carreers (including myself)
then having to entirely relearn new ways of doing things, in either much more convoluted, OR much less flexible solutions.
FXDude wrote:... pretty extra-outrageous decision
.. indeed.
... "mindless" is a term that could be used, or entirely devoid of consideration.


________________________________
Otherwise,
XSI would be dead well before 2010 if wasn't for ICE.
ICE... in all it's magnificence, remains but a miniscule part of it.

Forced migration might have been easier to take if Maya was even remotely as straight forward or "productive" (which it absolutely is not).

ONE example... (apart many other examples such as any you'd find in the "what were they thinking!" series up on the XSI list) ..
this is a series of videos from our very (ex-)own Raphael Fragapane.

     youtube.com/channel/UCX0OBmpsTUCeGaTDbHSb88g


I don't want to say that Maya rigging is "horrible", because that's probably relative to org size,
and how many TD's you have in your -T-D *department*, then no doubt allowing for very granular control.

But in basically -any- other circumstance, it really is ... ... horrible!


The hole 17 part series is about properly rigging a unicycle,

you can watch the hole thing or skip to any section, or to whenever he's not covering general theory, and actually getting things done,...
  (I downloaded them and watched at 2x speed with a player that maintained pitch... and I couldnt beleive it)...
no wonder why I had such a hard time finding node editor or indepth rigging videos...

...when "noodleing" (as he put it) in the node editor,
we see him getting totally confused, we could almost say most of the time, in an extra simple graph
while fighting (and say argh! at) NodeEditor glitches every couple seconds,
that while graphs can quickly get much(much!) messier,

... having a hard time even tracing back steps to refollow connections to nail down (almost constant) issues...

...and if not in the Node Editor, it's otherwise in a script editor window typing-in long pages of code...
that end-up being for trivial features,

... -barely- ever touching the 3dview,
  (you can scrub any video to the end with your mouse)
or almost never like once every 30 min, to see that it's not quite working
  (even if he's constantly saying "ok! now that that's working!")
or just to orbit around to go rotate something because basically nothing happened in the last 30 min,
before going-on doing or fixing things (at all times, hard to know what exactly)
in the Node or script editor for another half hour.

... the rig's "work" hierachy (not the hierarchy exposed to the animator)
ending up having dozens of nodes going pretty deep....

...same goes for constraints.. construction history .... ....... ............



Now It's worth mentioning that it has absolutely nothing to do with Raph,
who is by all means, and to the fullest extent, a seasoned and very experienced professional (at Animla Logic),
here making a series -for- proffesionals, and doing these things while talking is like a notch more difficult.

But even so.. it's just how Maya is for a host of big and small reasons.

"Like entirely normal for Maya, what's the problem?"
The problem is when you know how simple things -can- be (almost like a curse),
without implying expecting it to at-all work anything like what we may be use to.

And if Maya has one consistency, that would be it across the entirety of Maya.


I did that same rig in SI,
(or at least as far as I could tell what his rig actually did from bits and peices here and there after covering the entire series),
and took literally 15 min, and that's while making my own IK/FK switcher
with a keyable toggle that also snaps back the IK controller to current rotation,
  (never 'snapping' the rigged object out of current rotation when toggling the switch back and forth)
without using ICE, which I also could just as well while remaining just as ... -simple!-

OR I could have just used a bone which all have built-in IK/FK blending (gradual sliders)
then it would have been 7 minutes.

Like unbelievable...
or night and day between that, and say any one of pooby's tutorials
where without knowing anything about ICE ,
you can see what's happening, what he's doing and why.


So either XSI is incredibly user-friendly, or Maya is incredibly user-un-friendly, I'm not exactly sure yet, but it's definitely one of those. (or both)

Edit:: typos & specifications
Last edited by FXDude on 23 Nov 2017, 21:40, edited 1 time in total.

Bullit
Moderator
Posts: 2621
Joined: 24 May 2012, 09:44

Re: Mentalray EOL

Post by Bullit » 23 Nov 2017, 20:54

Bellsey wrote: 22 Nov 2017, 15:00 I don't really see any correlation between the Mental Ray EoL and Softimage EoL.
That wasn't the point being made. It was that Mentalray obsolescence contributed to demise of Softimage in say the perid of ~2003 to ~2007.

FxDude ICE appeared while Softimage was still with Avid.

User avatar
FXDude
Posts: 1129
Joined: 19 Jun 2012, 21:59

Re: Mentalray EOL

Post by FXDude » 23 Nov 2017, 21:55

Bullit wrote: 23 Nov 2017, 20:54 FxDude ICE appeared while Softimage was still with Avid.
YES it absolutely did! (did I suggest otherwise?) and probably a principal reason why it was snatched,
(as a yet more prominent competitor) almost at the very same moment of it's release.

As much as XSI may have been picking-up (and not going down because of mental ray.... :-? ),
an unrestrained XSI with brand new ICE would have no doubt amplified that by orders of magnitude
(as opposed to plummet from that exact point, -despite- ICE)

But Maya (and AD's bottome line) would have otherwise been mostly fine, and everyone would've been happy,
that includes Maya users because of all the new features to keep up with XSI (and vice versa), all with non-extra-inflated price tags.

Post Reply

Who is online

Users browsing this forum: No registered users and 30 guests