Indigo Renderer Forum

Hi. I tried to create what should have been a fairly straightforward animation of a light shining through a rotating prism, but I got very severe variations in exposure from frame to frame resulting in a pronounced flicker in the animation. I've attached images of three adjacent frames for reference. Each frame is rendered at 10,000 spp, but I rendered at up to 50,000 spp with the same results. Any idea what's causing this? Thanks in advance.

Make sure you use camera or linear tone mapping. Reinhard will automatically adjust the exposure per frame and will give flickering.

Enslaver, I used camera tone mapping.

pvbeeber wrote:Enslaver, I used camera tone mapping.

Please double check, since such a results is very unlikely for a constant tonemapper (or do you animate your emission power too?!).
Use MLT+BiDir for faster caustic rendering with "Large Mutation prob." = 0.5 and "Max Change" = 0.005 in the advanced settings. You camera should be inside the foggy material box btw. for faster rendering.
Also latest Indigo 3.4.17 build has some fix for bidir and participating media, but I don't know whats that about...

If you could share the scene we could take a deeper look
(two frames in pigs format, so everybody can give it a try)

MLT is quite a lot less uniform than without, so that could cause a lot of the problem. Maybe you can get away with not using MLT, just straight bidir; I'm pretty sure that will result in a much more stable rendered image frame-to-frame.

Zom-B, I made sure that camera tonemapping is used. I also had similar results with Reinhard tonemapping, which makes even less sense to me. I'll try rendering again with the settings you recommend to see if I get better results. I've also uploaded the Blender file if you'd like to take a look.

Lycium, I'll also try rendering without MLT to see if that helps.

Thank you both for your responses.

pvbeeber wrote:I've also uploaded the Blender file if you'd like to take a look.

Sorry no Blender (experience) here at all, therefor I asked for the two PIGS files

need to work on the material properties some?

Sent from my iPhone using Tapatalk

Sent from my computer using Google Chrome running Microsoft Windows on an Intel CPU with an Nvidia GPU driving the primary display and an AMD FirePro for the secondary.

Zom-B wrote:

pvbeeber wrote:I've also uploaded the Blender file if you'd like to take a look.

Sorry no Blender (experience) here at all, therefor I asked for the two PIGS files

Sorry, hadn't had my coffee yet. I've attached the two PIGS files.

Incidentally, I did get much better results when I placed the camera inside the foggy material (or, rather, when I expanded the foggy area to include the camera). Also, when I tried rendering the original scene without MLT (Bidir or path), I got nothing but darkness.

You've already figured out that placing the camera into the fog volume provides more reliable results.

I modified you file:

1. I placed a black box, which is a little smaller than the fog volume inside the fog. That black box has the surface normals pointed inwards.

2. I also subdivided the prism to round off the edges for some added realism. To add some dispersive effects I changed the material from a simple glass to the Lead Crystal Glass from the online material database.

3. I removed the long black cylinder with the disc lamp. Instead I used a simple disc mesh with the surface normal pointing into the lighting direction. I used an IES file as a light source attached and a little IES file that describes an IES laser so to speak. Now it renders blazing fast! Heck, I always wanted to make an IES laser

4. I turned off the Glass Acceleration. It is meant for think architectural glass or glass with coplanar faces.

I have attached a zipped file with everything in it. You just have to adjust the paths to the external material and the IES file.

Have fun!

Headroom wrote:1. I placed a black box, which is a little smaller than the fog volume inside the fog. That black box has the surface normals pointed inwards.

Nice trick since you should save the calculation for Indigo geting a ray out of the medium, since its directly killed, isn't it?!

2. I also subdivided the prism to round off the edges for some added realism. To add some dispersive effects I changed the material from a simple glass to the Lead Crystal Glass from the online material database.

Headroom wrote:3. I removed the long black cylinder with the disc lamp. Instead I used a simple disc mesh with the surface normal pointing into the lighting direction. I used an IES file as a light source attached and a little IES file that describes an IES laser so to speak. Now it renders blazing fast! Heck, I always wanted to make an IES laser

A old school method that works, but there get so many rays killed in that black tube that its performance loose we don't want

Headroom wrote:I have attached a zipped file with everything in it. You just have to adjust the paths to the external material and the IES file.

Have fun!

I would have too, if ya please could also upload a pigs file

Well.....I did create one but when i open it the Render button remains greyed out. I was under the impression that I can just open an pigs and click on Render. Is that not how it's supposed to work ?

I've attached it anyway. Perhaps it's a mac GUI think that I can't open it. or I am particularly dull tonight