mti 3.2.x

Announcements, requests and support regarding the MAYA exporter
User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: mti 3.2.x

Post by CTZn » Thu Jan 19, 2012 2:15 pm

ior wrote:Found an annoying bug:
When set any key on an object on rotateY (that is what I tested) the object disappears from the render.
Please confirm this behavior and try to fix it (I'm stuck with the animation :( ).
Keying an object rotate Y is working fine here, however nothing is happening if the object is under a keyed group. Transforms must have a shape as direct child to be considered.

I don't intend to hide the fact that this is a major flaw MtI has currently (hierarchies management). It will be fixed by the year I am sure.
ior wrote:Another thing.
When exporting the animation there should be an option to stop exporting, or else, if someone makes a mistake you need to kill maya.
Stopping still or animation export will be possible in the next update (it's implemented already).
ior wrote:For indigo renderer the result might be to power off the computer? Or there is an option to stop the animation and prevent indigo from popping up for each frame exported?
Animation rendering in Indigo can be stopped if it is in console mode (mti options > renderer > display > console). Closing the current console will stop the whole process. Another method is to kill the process cmd.exe, this will work with the Indigo UI as well. Note that cmd.exe will hang around idle if you interrupted the sequence otherwise.
ior wrote:About splitting the indigoShader into multiple material nodes, that would be a nice option to have.
This will happen when I have successfully hooked up the lightdata they require for viewport feedback. Actually I could split them already, only that I thought that releasing the two changes at once would have a bigger effect ;)

I'll try to have a release ready the next week. Thank you for your contributions ior !
obsolete asset

ior
Posts: 166
Joined: Tue Aug 24, 2010 11:47 am

Re: mti 3.2.x

Post by ior » Sat Jan 21, 2012 10:29 am

CTZn wrote:
ior wrote:Found an annoying bug:
When set any key on an object on rotateY (that is what I tested) the object disappears from the render.
Please confirm this behavior and try to fix it (I'm stuck with the animation :( ).
Keying an object rotate Y is working fine here, however nothing is happening if the object is under a keyed group. Transforms must have a shape as direct child to be considered.
Well, it is a bug. The thing happening is that when you make a key (rotateY for example) on an object it will not disappear as I said but will shrink or enlarge the object to the size of 1 and invert the rotation.
Because I has using large objects...
CTZn, in your test you may have not tested with objects larger or smaller than XYZ=1.
Please confirm this.

There is another bug involving sky witch appears on the opposite side of the sun.

I uploaded two scenes to show it. Render the scenes as they are and see.
Attachments
Bugs 3.2.2.2.rar
(20.19 KiB) Downloaded 313 times

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: mti 3.2.x

Post by CTZn » Sat Jan 21, 2012 12:02 pm

CTZn wrote:The scaling is fixed for the next release, it's a two liner really. The orientation will require a deeper investigation.
Fixed bug when a keyed transform has no rotation.

Update by the week I assume.
obsolete asset

ior
Posts: 166
Joined: Tue Aug 24, 2010 11:47 am

Re: mti 3.2.x

Post by ior » Sat Jan 21, 2012 2:26 pm

The sky is clear in the view from above and dark in the bottom view and both views are with Reinhard tonemapping.

I always saw the sky as a globe in indigo.
Attachments
error sky inverse - MtI 3.2.2.2.mb - top view.png
error sky inverse - MtI 3.2.2.2.mb - bottom view.png

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: mti 3.2.x

Post by CTZn » Sat Jan 21, 2012 3:45 pm

What you are showing is normal. The Reinhard tonemapping is adapting to luminosity like an organic eye, with no timed transition. It will multiply the input to match the fixed amount of light received by the film in total. It is an auto-exposure system in istelf.

Because the sun on its own is so bright in the image, the Reinhard tonemapping system has to lower the whole luminosity to match the parameters. The sky results darker then, because it is much less powerfull than the sun alone as an emitter.

To obtain the exact same luminosity from the bottom, the system must in turn amplify the film sensibility by a great amount. In addition, instead of coming mostly from one point (the sun), the light is coming from all directions equally, changing the perception drastically.

In fact, the bottom and sky (without the sun) have about the same luminosity. But in all cases the film is restituting the same amount of luminosity, as specified by the Reinhard tonemapping parameters.

Please note that the Reinhard tonemapping system is in cause for renderings turning black suddently. I do not recommend to use it during the lighting stage, only punctually.

Wouldn't the Camera tonemapping system be a better option by default ? For the linear tonemapping I recommend the actual default value for exteriors (0.035). Dunno wich you would prefer as defaut, camera is nicer to start with I think

The whole sky is a globe indeed.

Attached an image of the camera editor exposing tonemapping settings.
Attachments
Indigo_Editor_camera.jpg
slightly reviewed
obsolete asset

ior
Posts: 166
Joined: Tue Aug 24, 2010 11:47 am

Re: mti 3.2.x

Post by ior » Sun Jan 22, 2012 6:41 am

Off course, you are right, didn't noticed because the image from the top is brighter in the center and that does not make sense or does it?

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: mti 3.2.x

Post by CTZn » Sun Jan 22, 2012 8:14 am

It does, it is the vignetting effect, a kind of penumbra projected by the camera aperture (iris) onto the film. Unticking the vignetting option will totally remove this penumbra, wich is nonetheless a significant feature in Indigo's real camera simulation.

The camera aperture is altered by the f/ value, wich does also affect the depth of field and amount of light reaching the film.

While Maya does support most if not all camera models, I discovered their basic properties myself by using Indigo :)
obsolete asset

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: mti 3.2.x

Post by CTZn » Tue Jan 24, 2012 12:17 am

This was very usefull ior, the keyframes rotation and scaling issues are now fixed for everyone. Thank you for the example.
obsolete asset

ior
Posts: 166
Joined: Tue Aug 24, 2010 11:47 am

Re: mti 3.2.x

Post by ior » Tue Jan 24, 2012 9:53 am

CTZn wrote:This was very usefull ior, the keyframes rotation and scaling issues are now fixed for everyone. Thank you for the example.
Grate. I am pleased with the fast fix :) . I was needing it :wink: .

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: mti 3.2.x

Post by CTZn » Tue Jan 24, 2012 9:56 am

It will be coming this week.
obsolete asset

ior
Posts: 166
Joined: Tue Aug 24, 2010 11:47 am

Re: mti 3.2.x

Post by ior » Wed Jan 25, 2012 2:57 am

One more!

when you start to export the animation after you have done a single render, MtI will export the animation starting at the frame you previously rendered, although the name of the frame is correct, the image is not.

Can you fix this as well during this week.

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: mti 3.2.x

Post by CTZn » Wed Jan 25, 2012 3:59 am

I would if I could reproduce it.

I animated the camera over the 24 frames (3 keys), rendered frame 12 single then ticked animation, set rendering limits and rendered again.

The frame #1 (*scene name*.1.png) was correct. Let me know if that happens again, just in case. I dare suggesting that you loaded *scene name*.png as first frame ? Because still images and the ones from animations don't have the same naming convention, the later include the frame number (only when animation is ticked).
obsolete asset

ior
Posts: 166
Joined: Tue Aug 24, 2010 11:47 am

Re: mti 3.2.x

Post by ior » Wed Jan 25, 2012 5:56 am

Sorry, my mistake. It is working fine.
Forget about it.
I have to rethink before posting :oops: .

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: mti 3.2.x

Post by CTZn » Wed Jan 25, 2012 8:54 am

It's fine as long as we understand each other :)

And again, any oportunity for me to explain MtI better is a good oportunity.
obsolete asset

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Re: mti 3.2.x

Post by CTZn » Wed Jan 25, 2012 6:09 pm

I'm glad to announce a performance increase of about 5000% in exporting duplicate instances with MtI 3.2.5 (shipping the next 48h). One thousand duplicates were exported in 6s, versus 3 per second with 3.2.2.2 (clearly, a counter-performance).

In 3.2.2.2, adding more duplicates would increase the time required for each to be exported. This is not the case anymore.
obsolete asset

Post Reply
34 posts

Who is online

Users browsing this forum: No registered users and 11 guests