So I have done this particle pass rendering before with 1.1 version and everything rendered fine with matte objects.
now same file I am rerendering with 1.5.1(only change i done is used “overscan” script to cover more area out of camera) and you can see that strange matte/shadow line appears on particles.
I still havn’t figure out why it’s coming . I made matte object turned off for both “cast shadow” and “receive shadow” but still getting that matte line.
matte object is just high resolution tidal wave in background. and particles are in front of that wave plane.
Do I have to increase subdivision value for matte object? I tried 2 but same result. Also increasing subdivision increases rendering time but i am still getting the line.
Is it just me or I am getting this strange matte lines when particles are very near to camera. I also got this same issue before for one of dust scene.
I will let Darcy take a look and answer this, but it sounds like the motion blur bug he addressed recently.
It is quite possible that it is related to a bug we discovered and fixed about a week ago (the fix is still not available to you, but we are working on it).
I am surprised that it worked ok in 1.1. because I was under the impression that the bug has been there since the beginning.
Does it look right when not rendering motion blur? Does it render ok when rendering with Camera MultiPass Motion Blur (I know it is slower, but for the sake of testing…).
I wonder why it is in 1.5 coz I havn’t noticed in 1.1 either.(though I maybe wrong, I should check with 1.1 again, I was lately noticing it with 1.5 beta only) I always thought I am doing something wrong with version 1.5.1 or was thinking that I didn’t understand new matte object at all and I was doing mistake when rendering with high dense mesh as matte objects.
So far I was under impression that it happens with points only coz whenever I tried with voxels, it worked properly. But then with voxels, I might havn’t used motionblur becoz of pretty more rendering time.
I havn’t checked without motionblur yet. So I will check this without motionblur. I hope it’s just motionblur issue and nothing to do with points rendering.
I just checked it without motionblur and it happens to be still there but now as very hardline. It happens if I use matte object whether low poly or high poly.
Can you zip a single PRT frame and a Max scene and send them over?
I am not sure this is what I thought it was (the bug I mentioned was related to motion blur with matte objects, not points with matte objects. Difficult to tell by just a picture.
just rendered this with version 1.1 posted here for max 2010 and it rendered absolutely fine with motionblur.
So I guess it’s a bug/issue in 1.5 beta only. I also experienced this bug/issue since the beginning of the 1.5 beta but I thought I didn’t understood new matte system completely.
This seems to be an issue with Matte calculations of the Light - the darker portion is getting shadows cast from the matte object even though the particles are in front of it relatively to Spot01. Moving Spot01 closer and reducing its cone angle fixed the problem (as a workaround), but we are looking into what is causing this imprecision.
We will keep you posted.
Uhm no, you weren’t doing anything wrong.
It was a bug and it was simply amplified by the distance and the falloff angle of the light. We managed to fix it, but we were unable to produce a quick patch since we were deep into code changes for a future update. The current plan is to roll out only the bug fixes (of which we have many) as a v1.5.2 in the following weeks.
It seems moving light to closer to particles, it got removed the matte issue for particular frame but for further frames when particles go down more, it still remains there.
Anyway I will wait for next release (or render out this with 1.1 version)
I just checked version 1.1.3.36690 which supports max 2010. I saw it has support for “Use all CPUs”. (As I know that only 1.5 supports multithreading and faster matte objects rendering?) Or Did I miss something?
That has been there since 1.0.0 and controls the number of threads to use when sorting particles. Sorting and (partially) loading particles have always been multi-threaded.
In 1.5.0, we improved multi-threading by applying it to material evaluation when loading PRTs (multiple particles evaluated in parallel), culling in the PRT Loader, evaluation of KCMs, most calculations of Voxel drawing etc.
We also moved the CPU control over to the Preferences rollout because nobody is ever changing the threading settings unless debugging or running Krakatoa on many copies of Max. But if you install 1.0.0, you will find the same option.
We have posted a Beta 1 of 1.5.2 which should address the matte problems you were experiencing.
Please try it out and let us know if the issues have been solved!
