DISCUSSION: Computer can't render a large scene

Discussion in 'It's About the Art!' started by carmen indorato, Aug 22, 2016.

  1. carmen indorato

    carmen indorato Extraordinary

    Had a disappointing night last night.
    After hours of trying to set up a render I had to quit frustrated and angry as sin.
    I was reminded of the old days using Poser 4 when I did not have a powerful enough computer to handle the complex scenes i always seemed too eager to try and set up.
    Last night I had a cave and ground background with a modified V4 as a young Satyress using the Creature Creator set, with a nice braided hair strolling on the set playing a Greek fantasy lyre and in the background emerging from the cave was our Sparky created from Harry Kirrin being seduced by the lyre music.
    Everything was going great until the Kirrin was placed on set and i was trying to pose it coming over a slight hillock toward the seductive Satyress.
    I realized the powerful computer I had so proudly thought I had finally gotten after years of angry frustration, was in fact NOT powerful enough to set up and render my idea......AGAIN!
    I wish I knew how some people can set up the incredibly rich and complex scenes I see in the community.
    I am getting bored with simple planar sets and want to use forest and mountain scenes with multiple figures but alas seems forbidden.
    So I am going to stay away from Poser for a while until I can figure out where to go from here.
     
  2. Miss B

    Miss B Drawing Life 1 Pixel at a Time CV-BEE

    Hey Carmen, I don't know about others, but I often do multi-pass rendering. IOW, I render sections at a time, and sometimes have to save the sections as separate scenes. I did that not too long ago because I needed different lighting on a group of NapalmArsenals DayLily textures, than I needed for the partial view of a building the DayLilies were set near. In my case it wasn't a matter of not having enough resources, but just how the final render looked when I did it all as one scene.

    Needless to say, the two renders would have to be composited together in Photoshop, or whatever your preferred 2D app happens to be. Try it out next time you want to set up a complex scene, and render them out as .PNGs. That way the portions that don't include the background of the intended scene, don't conflict when you try putting them together.
     
  3. Rae134

    Rae134 Renowned CV-BEE Contributing Artist

    Unless mine are one or 2 characters and a HDRi for a background and MAYBE a couple of props (depending on what they are), I have to do multi renders like Miss B suggested above.

    (my Kittytaur above Miss B's post is about 5 or 6 renders)
     
  4. carmen indorato

    carmen indorato Extraordinary

    Hi
    Sorry to plug up this thread with this. I assumed because it was a harry Horse thread I was ok posting problems rendering him. I worked on a pure knee jerk action and apoligize to all.
    I am not the sharpest nail in the bag with Poser as many have come to understand about me over the past several years. I have come a way since Poser 4 but no way near as sophisticated a user as i should be by now despite working in PoserPro 2014.
    So complex stuff like you mention though i have heard of, I have no idea how to do.
    Also, Why .png aren't .tiff better?
     
  5. carmen indorato

    carmen indorato Extraordinary

    Also I forgot to add it is nt the rendering that is the problem but setting up all the characters, clothes sets and everything else that is becoming the problem. I had no problem with all of everything until I included the Kirrin. Seemed it was just pushing the envelope too far. I still think a fully rigged harry or in this the harry Kirrin is a bigger file that the MilHorse 1.
    Maybe all the textures are higher res and the transmapped hair and tail just add to the size of the file? But I even had problems moving it in wireframe.
     
  6. Rae134

    Rae134 Renowned CV-BEE Contributing Artist

    I don't use Poser so only guess why PNG over TIF is size/prefference (I know that DazStudio saves tiffs with "clear" backgrounds, but I don't know if Poser does).
    These are the separate layers I rendered and assembled in Photoshop as it just wouldn't do it in one go (I have a crappy computer). (I only did it small as an example but this is a thumbnail so the full image maybe bigger just click on it :p)
    workflow.jpg
    It could be you just had too much for your computer to handle, that happens to me sometimes. Try hiding the mane/tail/fetlocks when moving it around the scene to see if that helps (sometimes hiding other stuff in the scene can help too.)
     
  7. Miss B

    Miss B Drawing Life 1 Pixel at a Time CV-BEE

    TIFF files hold a lot of data, and are "heavy", and once you bring a .TIFF file into your 2D app, you're going to have to finalize the render, and export it again in it's final format. I don't know about this forum, but a lot of forums don't accept such heavy images because it limits the amount of storage space.

    Also, since I don't believe I've ever worked with a .TIFF file while rendering in 3D software, I'm not aware whether it has an alpha channel. The alpha channel is what gives you the transparency needed if you're doing multi-pass rendering, so you can compose the different renders on separate layers until you're happy with the way it looks, or if you're using a 2D image, such as the ones offered here in the store, as the background of the scene. If you don't have the alpha channel when you export your render, which happens if you render to .JPG, then you'll have a problem getting rid of the solid background color of the .JPG render.

    I happen to export my renders as .PNG, even if I don't need the transparency, simply because they are not "lossy" like a .JPG which can lose some details. Then again, it depends on how detailed the render is. That's usually only noticeable on large areas that have solid colors, and you wind up with "artifacts". That doesn't happen with .PNGs.

    TIFF files are "containers", so to speak, and can actually contain more than one document. I once opened a .TIFF file at work years ago to find it had over 100 pages of a document, each page in .PDF format. It was a nightmare.
     
  8. Rae134

    Rae134 Renowned CV-BEE Contributing Artist

    PNG files: can only be saved compressed (“lossless” – reduces file size), and can hold an alpha channel

    TIF files: can be opened with almost every image program (it’s an extremely common format), can be saved compressed or uncompressed, can store “layers” within (great for use in Photoshop for example), can hold all color, color depths and color groups (like RGB and CMYK), can save 16-bits per channel scans (your 48-bit scanner setting), can store IPTC metadata (captions etc.)

    I find for this kind of thing the PNG files are enough. Tiffs are great for Photos as it keeps the metadata (time, date etc). If you are using your renders for printing then you may need to convert them to CMYK tifs but thats all I'd use it for. (I'd never use Jpegs for working in because you loose info everytime you save, I might save a render as Jpeg to put online after I was finished editing it and I'd still keep the PNG/Tiff original)
     
    Pendraia likes this.
  9. carmen indorato

    carmen indorato Extraordinary

    As a preflight tech for a grande format digital printing house for 14 years I worked with .PSD files, .TIF files and.jpg files and many other file formats....except .png so I have been a bit queazy working/exporting to that format due to unfamiliarity with it and not sure if it is a format printing houses can easily work with. Every time you convert color tables too I know this for a fact because my job was to see it didn't happen and it happened often (especially noticeable when trying to color match from a clients proof), you get some color changes as well as sometimes image degradations. We got work from across the board of creative facilities and had to do whatever possible to get images magically resized to print "beautifully" on big prints like airport signage and bus and truck wraps etc.. So I am pretty familiar with .Tif files which was the file format of choice in most of the printing industry.
    Because I have gotten a few Poser pieces and other digital art works published in books which require 300 dpi at final size (8x10 or 11x14 and sometimes 12X18) I got into the habit of rendering my work out of Poser at that resolution which is tough if you have an older system. My new Mac can handle it....IF....I can get the file set up in Poser to even render. Which is the problem here, I can't gt the complex images set up.
    So to do multiple path rendering what do you do turn some parts or components of the image off render and export and then go back and turn something else off and render that etc. and when you have all the layered parts you rendered like you show here you re-montage them using the alpha channels as masks?
    Ooof what a nightmare but I suppose it is a bigger nightmare trying to set up and render something that refuses to play nice.
     
    Rae134 likes this.
  10. carmen indorato

    carmen indorato Extraordinary

    Poser will not use .tif textures or place background images in that format. The old Poser system used .bum files for bump maps. It no longer does that forcing old horse like me to re-saddle all my textures and background images and convert them to jpg. What a pain.
    It does export as tif format for the alpha channel which is great. Now i will have to try the .png format to see what it does and compare the two files for color and detail qualities.
    The losses are also quite noticeable in ares of great detail showing as digital artifact ot pixelation and quite ugly or in images where the edges are sharply outlined by contrasting simple backgrounds like a knife blade against a blue sky.
    How accurate are the alpha channels created within Poser when you render and export?
     
  11. Dreamer

    Dreamer Dream Weaver Designs Contributing Artist

    One thing I do for complex scenes is I pose all the figures in a empty scene then save them to the library as a character then load them into the final scene to render it
     
  12. Miss B

    Miss B Drawing Life 1 Pixel at a Time CV-BEE

    Working with images that are for print, and images that are for the Web, are in two whole different ballparks, so what works for one, won't necessarily work for the other. When I'm rendering these days, it's mostly for the Web, so I render in .PNG to bring into PS, and then do a final Maximum export to .JPG to upload here or elsewhere on the Web. If I was doing it for print, I'd go about it differently I'm sure.
     
  13. Alisa

    Alisa HW3D QAV Queen Bee Staff Member QAV-BEE

    Moved this into its own thread - it's something many folks face so good to have a place where everyone can talk about it.
     
  14. Gadget Girl

    Gadget Girl Extraordinary Contributing Artist

    Yes, I've run into that same problem with not being able to pose figures easily as my scenes got more complex. Like Dreamer I often use partial scenes to put together complex scenes.

    I'm not sure this scene looks all that complex, but I was using a large cliff prop in Terra Dome 2, and once I added in Dusk, Harry, and also some snow props, it started to get a bit complex. So here's what I did it make it easier to pose the figures.


    First I set up Terra Dome 2 the way I wanted it, and added in the cliff prop that the figures would be standing on. I spend a lot of time getting the cliff placed correctly and getting the camera set up, because I wanted a specific perspective that you get in the mountains when you are on a mountain yourself.

    Once I had that stuff placed correctly, I saved that scene, and then saved out just the cliff prop as a partial scene. That kept it's positioning and stuff for me since it wasn't just at 0, 0, 0.

    Then I opened the partial scene of the cliff, and worked on getting Harry posed. I actually at this point added in the snow prop they are standing in as well. Once that was done I save out Harry and the snow prop (since that was now effectively my ground) as a partial scene. In that partial scene, I added in Dusk, since he needed to be posed with Harry, the snow, and especially the reins. When that was done I reserved one more time as a partial scene with just Dusk and what he was wearing.

    Finally I went back to the scene with Terra Dome and the cliff. I loaded in some other snow props (the falling snow). Then I loaded the partial scene that was Harry and the snow on the ground prop. And then the partial scene of Dusk. Although there was still some tweaking to do at that point, most of the major posing was done, and all the characters were correctly positioned in the overall scene, so it wasn't too painful even though things were slow.
     
    Pendraia, Rae134 and RAMWolff like this.
  15. carmen indorato

    carmen indorato Extraordinary

    I have tried that but it only works if the set is simple w/o leaves and stuff in both background and foreground the figures have to interact with or if flat ground. The scene was complex with ground cover and a hill the kirin was coming over etc.
     
  16. carmen indorato

    carmen indorato Extraordinary

    Wow. really complex but will think on the strategy. thanks for the formula!
     
  17. seachnasaigh

    seachnasaigh Busy Bee

    This won't solve your problem right now, but the next time you buy a computer, do not buy a web surfing desktop. Do not buy a gaming computer. Don't even *think* of buying anything 32bit.
    Buy a used commercial grade workstation. Look for workstations which have two Xeon processors on the motherboard, preferably hex-core and HyperThreaded.
    Search eBay, Newegg, Amazon, et al for "2x X5650", or two X5660, two X5670, X5680, or X5690. A pair of these will give you 24 render threads.

    And get one with plenty of memory, at least 32GB. That large amount of RAM is what will give you the capacity to do large scenes.

    Urania is my weakest workstation. She has two HyperThreaded hex-core X5660 Xeons (24 render threads at 2.8GHz) and 48GB of RAM; she cost me $1,260. Urania is a 2008 model, and yet she will render faster than a new $10,000 gaming computer, and handle large scenes that would bring the gaming computer to its knees.

    For 3D rendering, you want as many processor cores as you can get (to speed rendering), and massive RAM (for large/complex scene capacity). The way to get that is to buy used enterprise gear - a professional workstation. New workstations are frighteningly expensive, but used ones are quite affordable.
     
    Last edited: Aug 23, 2016
    Pendraia, Banditcameraman and Rae134 like this.
  18. Rae134

    Rae134 Renowned CV-BEE Contributing Artist

    Yes, that's what I do.
    But if its moving around the scene that your computer is having trouble with, Pen's way sounds like it might be a the way to go.

    I'm a compositor by trade and have been in the printing industry for the last 27 years (besides some bar work for extra cash, its the only job I've ever had). Since I started, my trade has basically died (ahh I remember well the days when "Cut & Paste" really did mean that literally), I more of a finished artist but people don't know what that means (and its pretty much not heard of anymore either) so I tell them I'm a Graphic Designer (which I am, but I still feel uneasy saying it for some reason :p) so I know where you're coming from Carmen. :)
     
  19. Miss B

    Miss B Drawing Life 1 Pixel at a Time CV-BEE

    Boy do I remember cut & paste. I did it almost on a daily basis when I started working for my long-time boss back in 1970. Then again, we didn't have computers back then, so it was the only way to fix a "few" things, rather than retyping a very large document. You're right, though, they haven't done that since the advent of memory typewriters, and then computers. The necessity to cut & paste went out the window, and no one has looked back.
     
  20. carmen indorato

    carmen indorato Extraordinary

    Interesting dear. I stopped using the term: artist when I stopped doing art of any consequence. Poser has not afforded me the ability to be an "ARTIST" as I have come to understand the term or as I was one time what now seems like a lifetime ago. I get too caught up in the darn technology (WHICH I FRELLING HATE!) and seem to have lost connection with my mmmmmuse! Not sure if I abandoned her or if she abandoned me but suffice it to say we stopped "speaking". When one no longer does what one has done pretty much all ones life one tends to stop being himself or herself. I defined my "SELF" with the work I did as an artist, my profession, my passion. Now I no longer know who or what I am. I do some images because when you get used to expurgating your inner self via pictures it is hard to stop. But "ARTIST" Me? No. I can't draw an imge for the life of me and I USED TO DRAW 24/7 it seemed. i used to always have a camera glued to y face and have the crooked nose to prove it but I can no longer take a decent picture because I have forgotten how to use the camera a function that had become to be second nature and another way i defined myself.
    When one falls it is difficult to rise again. Several years ago, I fell...HARD> Financial, professional emotional and finally the Coupe de Gras...three deaths in my immediate family within 8 months broke me. Never really got up.
    Goya wrote:
    “When the brain is injured by an accident,
    or the mind disordered by dreams or sickness,
    the fancy is overrun with wild dismal ideas
    and terrified by a thousand hideous monsters
    of its own framing”.

    A brain is a terribly thing to waste and mine is terribly wasted. :(
     

Share This Page