I will try to make an animation but i dont have any clue; Should i make it of small resolution with very good quality or with big resolution and small quality?
İs there any configuration file for animation like we had on beta? I was thinking about using one of gi_qmc_lc files since it is a inside render and that configuration brins out the colors very lively…
First off don’t use qmc for animations, its just too slow. The best thing to use is IR+LC. Both methods have the option to save your solutions which is critical with animations. I’m not sure how involved you want to get with animation, but there are several priceless tricks that you can use if this is something you want to really work with. If you really want to get into it let me know and I’ll post a link to a good tutorial. Its for max, but with some explanation you’ll be able to make it work. As for .visopt files there is one there for animations, so I’ll zip those off and email them to you. As for resolution, its not really the best to decrease the size bc in the end it still looks bad no matter what the initial quality is. I usually go with either 640x480 or 720x480 with a .9 pixel aspect ratio. Those aren’t too big either so you should really stick with that size.
dalomar; i calculated the IR for every 5 frame but it doesnt use them when i try and take a render. The first render comes ok but the second and the rest comes dark again as it was before i made a 300mb irmap file
Should i make a irmap file for every frame?
It rendered small png files and a 200 irmap
Another problem is I cant seem to find taking the final file as a avi when render ends in VFS
You don’t need to make an IR map for every frame. Make sure the IR map is set to Incremental Add to Current Map and remember to clear your IR map from memory before you begin the sequence. The tutorial I sent you said to use Multiframe Incrememental, which we don’t have right now, but its the same thing. Once you calculate the IR map, make sure that you set the mode to file and grab the IR map that you calculated. Thats all I can really tell you without looking at your file. I’m not sure exactly the scene your trying to do this with is like (interior,exterior), but I would recommend trying it out on a few simple boxes with a very simple environment (just GI maybe) so that you can get the workflow down. That way things are simplified and the feedback is quick. Also it is lite on memory which may come into play (although I highly doubt it).
At this point I am not sure if its the calculation process or the render process thats causing the issue. To make sure you have a valid IRmap you might want to load it into the Irradiance Map viewer. Chances are you don’t have that tool, so send me an email and I will find a way to get it to you.
As far as the final file it will only be saved as individual images for each frame, so inorder to make a avi file from it you will have to use some sort of video editing type software. Don’t worry you could download Premiere from Adobe’s website, and getting a movie from it is pretty simple.
You also might want to do multiple single frame renders from various camera angles when incrementing the IR map. Since the IR Map is tracing and calculating only from the camera view youll need to increment the IR Map to get all the views youll need for the animation. I like to do north, east, south and west renders to make sure the IR map has enough light information for the object or objects in the animation.
The thing with that is that you have to merge the maps, and you can;t do that w/o the IR viewer. If the incremental doesn’t work, then that is definitely the way to go though. To be completely honest I don’t really think its and IRmap problem…unless its not getting loaded right.
i thought the incremental add to map would do just that it seems to do that for me, or am i getting my light from when the LC regenerates? which brings up a slightly off topic question of why is it that IR has an add to map but LC only allows you to save the post render with the auto save feature? The both trace from the camera correct? so they would in theory work the same way?
well i found a ir viewer; and merged the file but still the problem occurs; no matter from which scne a start from; the first render is always fine but the rest is darker.
by the way; when i render for irviewer it only brings out one vrmap file. Was it supposed to bring more than one for merging?
When you go through the animation that will be one whole IR map for the whole path. If you want to save individual IR maps from key spots, then merge them together that is fine too. At this point I don’t believe it the IRmap it may be something else that is happening down the line. Post a bit more about your process and maybe that will shed some light on things…I’ll get a chance to look at your file later today.
@Travis…I didn’t see your post before, but you ask a very interesting question. I think the reason why LC doesn’t have an Incremental Add function is due to the nature of how LC works vs. IR. IR uses the camera view to determine sample placement, then once the position of those samples it determined vray subdivides the hemisphere around that point to determine illumination. In other words the process for gathering the illumination for a given point is the same every time(subdividing the Hemisphere). With the LC samples there may be numerous bounces that contribute to a given sample. Once the illumination information is calculated this is only stored as the final result, not all of the bounces that make up the final result. Without being to seperate this information, it would be doubtful that LC could make use of it. Also an IR sample is valid for any camera position that can view it because its only dependance on the camera is the initial sample placement. With LC the sample is very dependent on the camera because the rays are traced from the camera, once the camera position changes the bounces to/from that sample change as well, which might throw off the calculation. This would explain why when calculating LC for animations the calculation process goes through the whole sequence as one rendering instead of taking rendering a succession of frames as IR does. This would allow LC to calculate for all of the viewing angles of a particular sample…Thats my best guess.
ok i finished one avi but it took the whole render except for the first frame darker. So i simply used color correction but the shadows came out too dark.
Just saw your reply Damien. Ya know, I posted that some time ago and have since realized that for the type of renderings that I do, LC is not that much of a helper.
Please don’t get me wrong, there are times and places where it can show a difference. I have definitely noticed that if you are rendering an interior space that is closed, LC is your best friend in the world. However.. when rendering a ring on a flat plane it doesn’t seem to improve the over all quality of the GI that much, and I will stress “that much” because there is some noticeable difference. I tend to rely on really tweaking out the IR to do the work, especially for Animation.
At the heart of what you just explained is really what are the illumination components of a given render. Most exteriors and studio shots rely more on direct lighting and primary bounces, so what gets used for secondaries becomes less important then the other illumination components. As soon as you start moving towards interiors you will rely more and more on secondary bounces, to the point where that may be where the majority of the illumination actually comes from. As a general rule of thumb, with simple secondary bounce conditions you will find less of a discrepancy between LC and QMC, both in render times and quality. As reliance on secondary bounces increases, the difference between the two will increase as far as time goes. As far as quality, thats a bit more of a conditional thing since direct illumination and primary bounces will mask quite a bit of what is by nature a blurry result.