All posts related to V2
User avatar
By JorisMX
#349079
Nice looking tennisball tom!

I can't state this often enough about newer maxwell releases though:
PLEASE update the manual. Or atleast add a PDF which explains the new functions and limitations.

I spent like 10 minutes with 2.6 and it crashed on me 2-3 times.
Lesson learned: Maya Hair and FIRE are not very stable. Atleast not on my workstation.

All renderings I've seen so far look great, as a matter of fact the new website looks great too!
Reliability and stability still seem wobbly to me. And I've really enjoyed using maxwell for I guess 3 years now.

All in all I believe maxwell is going in the right direction and I'm very curious where maxwell-core will take us in the near future!
By pavel59
#349085
OK, here some more information:
What actually happened to me: at the same time, while I was working on the final rendering of a project, Rhino 5 suggested me to update to a newer release, which I did and soon after, maxwell also informed me there was a new release available.
And - as I mentioned befoe - I was pretty happy with the news.

I downoaded both the core release and teh new plugin for Rhino, installed it, restarted the session and ... the images were darker, slower to render or, at least to clear from the noise in the initial stage, but the worst thing: the textures were all messed up.

before I'd been using, successfully, patch 2.5.1/march 2011 and Rhino plugin 2.5.13 August 2011 (x86 and 64).
previous Rhino beta was: rh50Beta_x64_en-us_20111025_2249

now the sw has been updated to Rhino 5.0 (5.1.2011, 1103, 03/11/2011). I didn't download the new release, it self-updated online.

Before this happened, I never had issues with texture mapping.

The model is complex and big as it's a complete bowling venue, but I was rendering it on separate stages, so far the one on which I'm having the problems is pretty simple.
Just a section of the venue, with simple walls and a sitting area.
the texture issue happens on simple planar surfaces.

I have no time to post images now, I will try to explain it simply.

Let have a wall with an inset, a kind of recess of approx. 15m in a 23m long wall.
All along the walls I have a decorated wall which is simply recalling the bowling lanes, running horizontally. So fr kind of teak deck if you like, a simple texture of maple wood with darker horizontal lines running from left to right.
I can hardly image anything simpler than that.

when I restarted my rendering, I found the textures all messed up. The wall is composed by 5 segments, three of them run parallel to each other (just one is recessed a few meters inside) and two other ones connect them, so far they're 90°, facing each other.

Simplest way to texture it is to select all together, apply a box texture, capped, using the bounding bow option.
Then adjust the U and V sttings and it's done.

No, the three main surfaces had the texture apparently turned 90à (lines running vertically) but unevenly spaced, while the two perpendicular surfaces kept the correct original setting.
All my attempts to fix it failed.
When I slected the texture gizmo, to see what was going on, I was shocked to see that - in place of a unique texture gizmo, planar or box, basically the shape and size of the selected wals, there were three overlapping squares, much smaller than any of the walls, laying horizontally quite far from the surfaces. Let say in a negative quadrant, maybe -15000x, -5000y, 0z. Assuming my model sits on 0,0,0.

After several tweakening, I was finally able to get a decent result by selecting alll the surfaces together, applying "surface texturing" on them and all appeared correct again.
After that now the texturing appears to work properly.
But I know I did all what I could reasonably do to correct it, without success.
I deleted the texturing and re-applied, it was always doing the same thing.

And, this never happened before to me. I'm not really a novice, and this model is definitely a basic one compared to my usual works.

----

about rendering:

All the values appear to work differently from before.
The image starts much darker and the noise takes much longer to clear.

before, the image started rendering pretty dark on the first pass, then - accordingly to my setup - it started to clear quite fast on the second and third pass, and also the noise started to clear after a few passes.

Now, the same image, with the same settings, keeps much darker, and the noise is being cleared progressively on each single pass, but much slower than before.

It looks like if before it was able to clear 60% of the noise in the first few passes and taking its time to compute the whole image, now it clears only a little % on each pass and you cannot predict the result until it's finished (yes, you can still see it on the preview image but the large final one takes longer to clear).
Pretty disappointing.

Also, the change of sliders limits and light values is really disappointing. Why are you doing this on every one or two releases ?

At least, this is what happens on my PC.

BTW, after a couple of days, reloading the file, the texturing issue seems not to happen anymore.
Don't ask me why, I didn't change anything, I just focused on something else which was more urgent than that, and left it partially unsolved. Now it works.

But it didn happen and it's not nice when it does that way. maybe it's a problem with Rhino releases, but it was crazy.

I can confirm the other issue, regarding the darker and noiser images, even on other models. I tested on a yacht I'm designing and it was the same. Much darker, slower and noisy.
And even the final image was not as good as before. It seems that i have to re-tune the lighting.
BTW, this happens on IBL and with or without emitters in the scene.

On the yacht model, I had to completely change the emitters values to get some results.

I will let you know as the tests proceed.

Paolo
User avatar
By JorisMX
#349101
I didnt have the patience to read all of the above however I would never ever update anything whilst in production unless you are fighting bugs that have been solved.

If you are so keen to try the newest version do it on your homeworkstation or a machine thats not tied into production!

Sometimes updates (esp windows updates) can happen by themselves and it can be very frustrating when it disrupts your workflow. With Maxwell I have only seen windows pop up to inform and ask you if you want to download the update or not. In Maya this can be turned off. Not sure about Rhino. The plugin and maxwell update installation will have to be done manually which should save you from updating when you don't want to

I made a similar mistake a few years back and it cost me a couple night shifts to fix all my issues. But in the end I had to accept that I learned a lesson and that it wasnt the software that went wrong but my impatience due to the heavy workload.

I hope you solve your issues, Pavel!
User avatar
By JorisMX
#349104
I have to correct myself about the documentation!

I was browsing around these Forums when somebody posted a Link to the online documentation.
It also reads very nice on th iPad I got as a gift from a client. Time to sit on the couch with cup of coffee and read all about the new functions yay!
Last edited by JorisMX on Sun Nov 06, 2011 2:34 pm, edited 1 time in total.
User avatar
By Bubbaloo
#349105
pavel59 wrote:What actually happened to me: at the same time, while I was working on the final rendering of a project, Rhino 5 suggested me to update to a newer release, which I did and soon after, maxwell also informed me there was a new release available.
And - as I mentioned befoe - I was pretty happy with the news.
I too did not get past this first paragraph. Upgrading one software, let alone two!, during a project!?! That's begging for trouble. I always try to time the upgrade with a brand new project. It's so much easier that way.
User avatar
By JorisMX
#349106
Here's a question about long term backwards compability:

Last week I was trying to get rid of heaps of old data I generated with mxwl versions dating back about over ago. I was wondering how to get these scenes to stay compatible when a client calls me in maybe 6 months or so.

Do I have to keep track of version numbers and then reinstall 2.4 or 2.5 in order to correctly read the old mxs files?
By JDHill
#349113
@pavel59: there is probably no need to post any images from Rhino, as V5's texture mapping has been broken in various ways throughout the alpha/beta process; at the moment, Rhino Render does not even render mappings as shown in the viewport, in many cases. If you are doing production work, you are taking a risk by doing it with non-production software.

As I have posted in the sticky in the Rhino forum (please use that forum for plugin-related questions; I do not check this one very often), Maxwell for Rhino is a Rhino 4 plugin -- Rhino 5 support is provided strictly as a convenience for those who wish to participate in the Rhino 5 alpha/beta process, allowing them to render in Maxwell without having to use Studio via OBJ export from Rhino. Being a Rhino 4 plugin, I cannot pursue any V5-specific workarounds, and except in very obvious cases, any potential plugin bugs found while using Rhino 5 need to be reproducible on Rhino 4 before they can be considered as such.

I can't comment much on anything but the plugin; just quickly though, I know people had previously complained of issues when combining sun light and emitters; the emitters would apparently get too high of a priority in the calculation, with the sun light fully resolving only at high SLs. So judging from the description of your scene, perhaps changes made regarding this (I don't know that there were any, I am just guessing) are affecting how your particular scene resolves, with it being darker, overall, at lower SLs.
User avatar
By Bubbaloo
#349116
simmsimaging wrote:
Bubbaloo wrote:
Can you do a pack n go on the file?
Sure - here it is:

http://www.simmsimaging.com/upload/file ... st.MXS.zip

I got rid of the emitter with the image map - it wasn't necessary. Sorry about that.

b
It's missing the particle file. :mrgreen:
By pavel59
#349126
OK, I'm aware about production/non production software. I'm also aware I was taking my risks.
However:
When Rhino5 runs out of date and informs about a new update there's very little to do, if you want to keep using it.
And Rhino5 is considered a pretty stable platform now ....well, please don't induce me in polhemics I left in Italy a fw years ago. I'm one of the earliest users of Rhino since 1995, and I actually love it, but I'm not dumb and I can recognize when something start having some issues. And I was quite critical and had some personal arguing with Bob Mc Neel, time ago, about some [bad] strategic decisions. A development lasting more than four years now ? I definitely prefer to not say one word more about that.
BTW, for what concerns me, some features included in Rhino 5 improved my workflow, that's why I took my risks.

And I'd not been screaming this time, yes I was on a project but those issues didn't compromise it so much. Just enough to disappoint me but not to the point of make me screaming.

This time it was just the case that both the sw, Rhino and Maxwell alerted me about the update.

And about Maxwell, maybe it was due to Next Limits excellent improvements in the recent past, which induced me to be so confident and not careful enough.

Again, nothing serious happened.

I'm just pointing the fact that the new release doesn't appear to be much faster - if any - than the previous one. And the final results doesn't seem too different, it's just the way it behaves which is different from what I'd been used to. And that's a bit irritating. But I will survive to it and will get used to the new settings. Hoping that this will not change again over the next six months ... that would be really bad.

Paolo
User avatar
By simmsimaging
#349128
Bubbaloo wrote:
simmsimaging wrote:
Bubbaloo wrote:
Can you do a pack n go on the file?
Sure - here it is:

http://www.simmsimaging.com/upload/file ... st.MXS.zip

I got rid of the emitter with the image map - it wasn't necessary. Sorry about that.

b
It's missing the particle file. :mrgreen:
I dont think Max generates a separate particle file?? I just pointed it back to the .mxs file and that seemed to do the trick. Is there supposed to be a separate .bin file or something like that?
User avatar
By Bubbaloo
#349130
simmsimaging wrote: I dont think Max generates a separate particle file?? I just pointed it back to the .mxs file and that seemed to do the trick. Is there supposed to be a separate .bin file or something like that?
Hmmm, I just quickly looked at the Maxwell Particle object, and there was no particle file referenced. It was blank. I'll take another look. I'm only used to working with RF particles, and it always works with a reference to an external BIN file.
By pavel59
#349175
It's definitely the latest Rhino 5 to mess up with the texturing.
I will post this info on the proper section as well.

For the rest, once got used to the new settings (multilight layers, plus additional setting to IPL and physical sky) it's pretty OK, and maybe the final result a little better than before for the same rendering time.

Paolo
  • 1
  • 3
  • 4
  • 5
  • 6
  • 7
render engines and Maxwell

"prompt, edit, prompt" How will an AI r[…]