Rendering motion blur in Arnold - 3ds Max Tutorial

very interesting tools for controlling motion blur for both realistic renderings and special effects. I've got a very simple animation of a fast moving flying saucer that will be a good candidate for showing off the motion blur effects, especially, in this close up view here. Let's park the timeline on frame 14 Increasing the number of samples is the only way to remove grain from the motion blur, and in this case, because I'm doing some test renderings, I'm actually going to knock the number of samples down to two just to speed those test renders up. Let's do a render with no motion blur just for comparison. Click the Render button and here it is with very clearly defined edges. Let's enable motion blur, scroll down in the Arnold Renderer tab until you find the motion blur roll out. Here it is. The default is to respect the physical camera settings. You can control the motion blur effect from the physical camera itself and I cover that in a course called 3ds Max Cinematography for Visualization. Here, we want to explore the unique properties of Arnold and see what it can do that the physical camera cannot. One of these three general switches has to be enabled to get any motion blur. I want to use Transform Keys. So, disable Respect Physical Camera Settings and enable Transform Keys. Transform Keys or Deform Keys will set the number of time sub-samples during the period the virtual shutter is open, and this object isn't changing shape, so we don't need any deform keys at all. For simple motion, such as this, we only need two transform keys. One when the shutter opens and one when it closes. If the object were moving in a very obviously curved trajectory, we might need to increase the number of transform keys to better resolve that curvature, but here the object is moving in a nearly straight line. So, we only need two transform keys. Well, that is the Motion Blur Time Span and the default length is .5 frames. The default behavior is to center the samples on the current frame or precisely at the current frame. Let's do a render with just default parameters, see what that looks like. Here's the result with a physically plausible shutter time of half a frame. For demonstration purposes, to make these different options more obvious on the screen and also for a cool special effect, I'm going to set the length of the shutter to be two frames, and now we should see a much more exaggerated motion blur. Do another rendering. With that highly exaggerated motion blur length of two frames, we can better visualize these other options. Over here, we have a pull down list, and it's defaulted to Center on Frame. In the rendering, we can see that there is blurring both ahead and behind the current frame. So, if we compare that to the viewport, the saucer is right about here, and we're getting motion trails behind it and ahead of it. If we switch this over to Start on Frame and do another rendering, we'll see something a little bit different. We see blurring only ahead of the current frame, so here's the location of the saucer on frame 14 and we have trails kind of ahead, while preceding the current frame, and the third option over here is End on Frame, which is what I really want and that's going to cause motion blur only after the current frame. We've got our saucer right here in the center of the frame and some trails behind it. So, that's the most basic way to control the shutter, but that is just the beginning because the length of the shutter and where it begins and ends, are not the only factors 'cause in the Shutter Behavior section below it, there is a function curve here, and that 2D graph allows us to vary the density of samples for the motion blur over the course of time, With a start value of negative two and an end value of positive two, the samples are distributed with half of them before the current frame, and the negatives down here, and half of those samples after the current frame, in the positive zone, but our motion blur time span is set to End on Frame, and that means there are no transform keys after the current frame, and so, there is no motion between the current frame and the next one. So, the shutter stays open and half of the samples are concentrated on that period. The result is that the current frame dominates the rendering and the sub-sampled motion blur before the current frame is diminished in intensity. And we can change that balance by adjusting the end value here. We can set end to a lower value, like .05, and that will cause it to be a bit less distinct. If we want it to be very, very distinct in that leading edge there, we can set the end to a high value, like five frames, and now, we're concentrating the vast majority of the samples on frame 14, and the time period before frame 14 is not getting very much sampling. Okay, so that's one way to do this. There's another way, which is using the function curve itself. We'll set the shutter behavior back to Auto, so the extents of this function curve will match the parameters in the motion blur time span. Now, we can isolate the effects of the shutter function curve, itself, and by default, this curve is disabled or grayed out, and that's because the default shutter type here is box. Shutter type gives us another way to adjust the weighting of samples. So, by default, with the box filter here, samples are spread out equally over the shutter curve extents, and during the period defined by the shutter behavior, start, end, or the Auto mode here, the samples are all weighted equally, and the equal weighting of samples, called the box filter, gives us a technically physically correct result, but we can change this up. First of all, let's see what it looks like with just the box filter. The physically accurate box filter gives us pretty well defined edges here at the trailing and the leading edges. And we can soften that up by using a different filter. So, let's make a clone of this rendered frame window so we can compare it. Got this, set this off to the side, and switch the shutter type over here to triangle. Okay, and do another rendering, see what that looks like. see what that looks like. With the triangle filter, With the triangle filter, the weighting of samples is not constant the weighting of samples is not constant across the function curve or the shutter period. across the function curve or the shutter period. Samples will increase linearly Samples will increase linearly from the beginning of the period from the beginning of the period and reach the peak at the middle of the period, and reach the peak at the middle of the period, and then, decrease linearly and then, decrease linearly to the end of that period, to the end of that period, resulting in softer edges to the blur, resulting in softer edges to the blur, and you can kind of see that here. and you can kind of see that here. This is a little bit softer This is a little bit softer than in the box filter version, than in the box filter version, and of course, higher camera samples and of course, higher camera samples will help resolve that even better. will help resolve that even better. And then finally, we have the ability And then finally, we have the ability to create a custom curve and concentrate those samples to create a custom curve and concentrate those samples anywhere we want. anywhere we want. So, let's close the rated frame window, So, let's close the rated frame window, go over to the shutter behavior go over to the shutter behavior and set the shutter type to Curve, and set the shutter type to Curve, and now, we have an editable function curve. and now, we have an editable function curve. I'm going to change this up a little bit I'm going to change this up a little bit using the default move to, I'll take this last point using the default move to, I'll take this last point and bring it up to its maximum of one. and bring it up to its maximum of one. This third point over here, This third point over here, I'll bring that a little bit closer, I'll bring that a little bit closer, and then, this second point here I'll bring and then, this second point here I'll bring to create this kind of tail. to create this kind of tail. So, what I'm doing here is I'm saying I want lots of samples here. So, what I'm doing here is I'm saying I want lots of samples here. The number of samples or the weighting of samples The number of samples or the weighting of samples is the vertical dimension, is the vertical dimension, and the time is the horizontal dimension, and the time is the horizontal dimension, and what we're doing here and what we're doing here is we're steadily increasing the number of samples is we're steadily increasing the number of samples across nearly two frames in this case. across nearly two frames in this case. And then, just before this shutter closes, And then, just before this shutter closes, we're going to take a lot more samples. we're going to take a lot more samples. Alright, so let's see what that looks like, click Render. Alright, so let's see what that looks like, click Render. Alright, very cool. Alright, very cool. Well, you've concentrated most of the samples here, Well, you've concentrated most of the samples here, and the leading edge are just before the current frame. and the leading edge are just before the current frame. That's how to use the Arnold motion blur tools, That's how to use the Arnold motion blur tools, both for physically plausible results both for physically plausible results and special effects, such as, trails behind objects. and special effects, such as, trails behind objects.