Announcement

Collapse
No announcement yet.

How big should my map be?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • How big should my map be?

    One of the things i struggle with the most when rendering in Max (using VRay offcourse) is to get my maps to look sharp and crisp (even when they r close to the camera).

    For example in the current image i am busy with, i created a texture for the the parking area outside the shopping complex. Very simple map, but i did thought i created it big enough (the jpeg file is 19mb) but still the map looks awfull. I can only see some of the with lines indicating the parking areas and the texture looks stretched.


    Is there away (i guess there is, if i look at all the stunning pictures in this forum) how one calculate what size your texture should be?
    Last edited by WWX; 18-03-2009, 08:49 AM.

  • #2
    Have you UVW maped your object correctly? Also reducing texture blur in coordinates tab might get you better results.

    Comment


    • #3
      I must admit, I am never sure about this either. Sometimes we turn blur down to 0 (which changes it to 0.1 or 0.01 or something anyway?!). Sometimes we shrink the bitmaps that we use to about 1200pixels square to reduce memory overhead. In the past, we have used the bitmap proxy settings, but this started to get a bit complicated with different users using the same files. More often than not, these days, we use the full size bitmaps that we have and don't bother resizing at all. Its slow and clunky in the material editor, but we accept that.

      Another spanner in the works is the image sampler: we use Adaptive QMC in a range of about 1-14 with the clr threshold at about 0.005. With this setup, I often don't bother with an antialiasing filter at all, but I know this will have some bearing on things.

      A prime scenario would be a stone paving bitmap, say 6000x6000 pixels, and an eye level camera that is above a piece of geometry with this bitmap-based texture applied to it. What would be the best setup?
      Kind Regards,
      Richard Birket
      ----------------------------------->
      http://www.blinkimage.com

      ----------------------------------->

      Comment


      • #4
        The theory is that a map should be twice as large as it appears on the screen. For the sake of convenience lets say that your final output res is 1000 x 1000. If you have an object that will fill the full frame then you double the frame res to give the renderer plenty to work with so 2000 x 2000. Say for example you zoom out so that your object fills half the width and height or a 500 x 500 pixel area of the final frame, again double it to arrive at 1000 x 1000. If you end up zooming in on the middle 50% of the object then you need to take this into account too - so the final render size x 2 = 2000 x 2000 and then by 2 again to take into account the zoom in to 50% of it - 4000 x 4000.

        A lot of folks stick an unsharp mask on their final renders in photoshop for that ubercrisp look!

        Comment


        • #5
          Thanks for the tip. The reality is that we just use what we have. But what I am most interested to hear about is what users set for the bitmap blur value - the default 1.0?

          I am sure that I had read that you should set that as low as you can (0.1 or 0,01 or something) and then Vrayt handles all the AA with the Adaptive qmc setup. Is that right?
          Kind Regards,
          Richard Birket
          ----------------------------------->
          http://www.blinkimage.com

          ----------------------------------->

          Comment


          • #6
            Great info so far thanks!

            Something I would like to check is well, so do most people prefer not to choose a AA filter and then run Unsharpen filter post in PS? Why is this option favoured, does it same on rendering time?

            With simple test i have done trying to use this workflow I just seem to have more white spots in my images, which PS won't fix. The white spots is due to glossy reflections and although I thought I was using hih settings, they persists.

            ISUrg - this specific instance is just a rectangle so i have applied a box UVW, so i think it is correctly done and my texture blur is set to 0 (which gives the 0.01 value)

            joconell i am going to try your advice, but man i think i am going to have one massive texture file. The only part of your explanation i didn't understood was this - " If you end up zooming in on the middle 50% of the object then you need to take this into account too - so the final render size x 2 = 2000 x 2000 and then by 2 again to take into account the zoom in to 50% of it - 4000 x 4000. " - first what do you mean with zooming in on the middle 50% of the object? (is this related to fly throughs/fly byes?) I am just doing a still rendering, so how does the zooming related to that?

            Thanks!

            Comment


            • #7
              I'd say most people should avoid using really sharp or really blurry filters - both can be done in post afterwards with a gaussian blur or unsharp mask - things like the video and catmull filter go a bit to far to pull back. No aa filter will save a little time, depending on which one you choose.

              White spots are a tough one alright - probably something like a subtle glow in post would be the best option since it'll occur in real life and it'll soften the problem area.

              Comment


              • #8
                I turn off bitmap blur on nearly all maps except those that will produce moiring effects, like bricks, small tiles etc.
                ____________________________________

                "Sometimes life leaves a hundred dollar bill on your dresser, and you don't realize until later that it's because it fu**ed you."

                Comment

                Working...
                X