Jump to content

Fluffy Sharkfin

Resident
  • Posts

    1,115
  • Joined

  • Last visited

  • Days Won

    1

Posts posted by Fluffy Sharkfin


  1. Marlon Wulluf wrote:

    The framed animation I'm attempting to create is similar to a pulsing light, starting luminosity at 5, ending at 20, then reducing back to 5,,,

     

    In which case you can reduce the number of frames required by 50% if you enable PING_PONG mode in the llSetTextureAnim command, which would allow you to utilize twice the pixels per frame.

    Another alternative would be to use SCALE mode to enlarge/reduce the texture in order to create a pulsing effect. 

  2. But, don't alpha layers work in exactly the same way as clothing layers?  In which case having a full perms alpha layer doesn't give you access to the textures used to create it, only the ability to substitute those textures for ones you have in your inventory (in which case you may as well just create a new alpha layer yourself).  I think what you'd need is for creators to include a full perms copy of the texture used to create the alpha layer, not a full perms alpha layer.

    Besides, creating an alpha layer is really not that difficult if you have a copy of photoshop and some SL skin templates.  Just wear the templates as a skin, attach the mesh clothing you're creating an alpha layer for, open the same skin template in photoshop and paint your alpha layer texture.  It's a 5 minute job (if that) and, as many people have pointed out, alpha layers tend to rely a lot on shape settings, so knowing how to create your own is a skill well worth learning.

    • Like 1

  3. Coby Foden wrote:


    Fluffy Sharkfin wrote:

    image6a.jpg

    If you'd rather take a look at it in person feel free, the build is currently located on
    .

    I went to have a look.  The whole thing rezzed almost instantly.  Amazing and impressive showcase of effective mesh build!

     

    Thanks Coby :)

    I'm still experimenting and learning, from what I learned putting it all together I think I could probably lower the LI by around 20% while still adding more detail and reducing the geometry rendered per frame at the same time.

  4. So basically what you're suggesting is that, because you can use sculpted prims and the legacy prim accounting system to cheat on land impact, LL should make it possible for us to do the same with mesh?

    It's obvious that LL aren't going to break existing content (considering the animosity that the introduction of mesh already caused with some residents doing so would have been suicide).

    Sorry, but it seems to me that blindly clinging to the one slim advantage which sculpted prims have over mesh and insisting that it means "mesh is broken" is both childish and misleading to those who are trying to learn how to make the most of the tools currently available to them.

  5. If you're finding that larger mesh objects are too heavily penalized under the new land impact system, have you considered using mesh to make smaller building components instead, then using those to build larger structures?

    I recently made some mesh building components and found that the amount of detail that can be achieved using mesh goes far beyond anything possible with sculpted prims, and with significantly lower land impact.

    Here are a couple of example images of build components with corresponding land impact values shown.

    Image1.jpg

    (single girder 1060 tris, 0.5 LI)

    Image2.jpg

     

    As you can see from the image above the LI stays below 1 per linked mesh object, based purely on polygon for polygon (each girder being 1060 tris and the connector in the center having 2182 tris, which makes a total of 8542 tris for 4 LI) mesh wins out over sculpted prims (4 sculpted prims @2048 tris per prim =   8192 tris).

    Here are a few more example images of other build components.

    Image3.jpg

    Image4.jpg

    image5a.jpg

    image5b.jpg

     

    When you also take into consideration the fact that most of the components have multiple materials to allow selective application of effects like tint, shiny, glow, transparency and texture animation, and UV mapping designed to accomodate the use of multiple seamless textures with varying repeats and offsets from each other all on a single mesh (not to mention that all 28 build components use only 6 shared textures), it becomes clear that mesh gives you the ability to create things far beyond the scope that sculpted prims can offer.

    As for giant sim sized builds, back when we had a 10 meter limit on prim size, it was common to take several of these small blocks and link them together to make larger structures.  So if there's an apparent limit on the efficiency of mesh based on its size, perhaps you should try taking lots of small mesh objects and linking them together to build bigger things, like this...

    image6a.jpg

    image6b.jpg

    If you'd rather take a look at it in person feel free, the build is currently located on my homestead sim.  I'm currently taking a break from adding to it, not because I ran into problems with LI (I've only used around 60% so far) but because the amount of geometry being displayed per frame is getting to the point where older hardware may seriously suffer if the user has their graphics turned up too high, and I'd like to keep the build efficient performance-wise as well as land impact friendly.

    So, to answer your question, no meshes really aren't "just a hoax with no use", you simply need to experiment a little more and find the optimal way to use them. ;)

    • Like 2

  6. Chosen Few wrote:

    • Don't use the Delete key on your keyboard.  Always use the Delete Edge/Vertex command from the Edit Mesh menu.   When edges are deleted via the Delete key, the vertices remain intact, leaving you with a model full of Ngons, and no actual detail reduction.  The Delete Edge/Vertex command removes the edges and the associated vertices at the same time.

     

    Actually using the Delete key will also remove some vertices, but only those which do not appear at the edges of existing polygons (for example, if you have a polygon with a hole in the center, then select the edges of the hole and use the Mesh>Fill Hole command, then hit the Delete key to remove the edges, the vertices for those edges will also be removed since they aren't part of the edges of the remaining polygon).  

     

    In most cases I'd agree that using the Delete Edge/Vertex command is a better option, however in some instances I've found it can be beneficial to use the Delete key instead, since doing so will remove the selected edges and any vertices which are no longer used to define the edges of the remaining polygons, but will leave the vertices (and corresponding UVs) at polygon edges untouched, you can then use the Split Polygon tool to turn the resulting Ngon into triangles/quads.  While using this method will give you less of a polygon reduction, it can sometimes help to preserve the shape of existing polygons and UV borders.

  7. Each LOD model will need to have it's own UV map, and removing edge loops/vertices will also delete the corresponding UVs.

    It helps to plan ahead when creating the initial model and try to ensure that the UVs you use at the edges of your UV map (or any UVs that you manually adjust to stretch/compress the texture across certain polygons) correspond to vertices that won't be removed when creating the lower LOD models.  If you enable the display of texture border edges in Maya ("Display > Polygon > Texture Border Edges") those edges will appear thicker than non-border edges when you mouseover/select them, which helps to avoid deleting any critical UVs when reducing polycount.  You can also alter the display options in Maya so that all texture border edges appear thicker than regular edges (even when not selected) by going to "Display>Polygon>Custom Polygon Display..." and enabling "Texture borders".

    Also, it can sometimes be easier/more effective to not use the original texture from the higher LOD models, and instead create a smaller version for the lower LODs (since they're only seen from a distance so don't require as much texture detail) and include that in some small unused portion of the texture/uv map.  This allows you to add details to the texture that simulate the geometry that was removed when making the low polygon version, without those details appearing on the texture when viewed at higher LODs.

    • Like 1
  8. If this 'mystery product'  relies on a connection to an external server for it's continued functionality then your purchase would include the item itself, and ongoing access to the online service upon which it relies.  When signing up for an online service it's always a good idea to read the terms of that service, and failure to do so for whatever reason is rarely considered the fault of the provider.

    If you were aware that connection to the service was required for functionality but didn't review the TOS prior to purchase to see whether that service was subject to change without notifications, etc. you really don't have much recourse if you consider the product/service to no longer be satisfactory, other than to simply stop using it.

    As for changing the permissions of the product you purchased, according to this post

     


    RagDoll Lemon wrote:

    Particularly when older versions of products are made disfunctional by the external servers they talk to forcing the buyer to upgrade to a working version only to discover that the updated product's perms have been changed to nocopy, notransfer.



    the permissions of the product itself didn't change, it's just that the newer version which is required for continued use of the online service doesn't have the same permissions as the version you originally purchased.  Once again, most online services are "subject to change", and insisting that a creator selling you a product with one set of perms is some sort of verbal contract guaranteeing you the same perms on future versions of the same or similar products is silly!  What's next, suing LL because you can't connect to SL using viewer 1.1 anymore?

  9. The process you describe for creating objects with multiple materials should work fine (it really is simply a matter of selecting groups of faces and applying different materials to them), my guess is your problem lies in the version of Collada you're using.  For starters, if you're using Maya 2011 you really shouldn't need to export as FBX and then convert to DAE, since that process is just a workaround for those who are stuck with newer versions of the FBX exporter which aren't compatible with SL (i.e. Maya 2012 users).

    I recommend using the FBX 2011.3.1 plugin (which can be found here), so far I haven't had any problems exporting DAE files directly from Maya (multiple materials, skin weights, joint offsets, etc. all seem to work just fine).

    • Like 1
  10. Rather than using Duplicate (CTRL+D) try Edit > Duplicate Special, open the options and change the scale value of the axis you wish to mirror the object on from 1.0 to -1.0,  That should give you a mirrored copy of the object with correctly mirrored normals. 

    If you're working with something you intend to continue editing set the Geometry Type to Instance rather than Copy since that way any further changes you make to the geometry of the original will also occur in the mirrored instance as well, just remember to delete the object history on the instance before exporting.

  11. As an alternative to generating more particles to try and close the gaps you can also try increasing the size of each particle thereby reducing the number you need to make a complete chain.  Unfortunately particle systems have been prone to the "sputter" effect for some time now, so to some extent you're probably just going to have to live with it in busy/built-up areas.

     

     

  12. You could also try using PSYS_SRC_ACCEL to add a little "drag" to the particle string by applying a slight amount of force to each particle on the Z axis (to simulate gravitational pull).  It's worth remembering though that the more the particle stream curves the shorter and more numerous the particles will need to be in order to create the illusion of a straight line rather than a "crazy stitching" effect, which in visually laggy areas will probably result in "breaks" in the string (since particle systems seem to be the first thing to suffer when SL gets laggy).


  13. Yohan Roux wrote:

    It seems many of you didn't like my view of content creators that sell, ok why not give them away,
    all I have heard here is how people are being hit by the lindens going into competition with them
    , well the lindens went into competition with me, forced me to fight them, that is life here, one must swallow that and move on not cry about it, market place has killed the malls, I know how they can be revived but it is impossible to share because of the competition that exist between creators, this clearly shows that your view as to why creators are here is wrong, this thread shows all traders here are here for profit.

    In which case you clearly haven't bothered to read every post?!

    As for your complaints about the competition that exists between creators, I find them a tad ironic considering your post in the LSL scripting forum where you told someone asking for help with a combat system "as you are a competitor I should not help you"  (since the main function of that forum is as a place for residents to seek help with scripting problems, I can't help but wonder why you'd post there if all you were willing to offer was cryptic suggestions?)

    Anyway, regarding the original topic, you may want to consider the difference between a "trader" and a "creator".  Pretty much all traders are creators, but that doesn't mean that all creators are traders, and to ignore the opinions of those that create without seeking to profit (or deny their existence entirely) simply because they don't agree with your point of view isn't going to strengthen your position or make your arguments seem more valid, it simply makes you appear rigid and uncompromising.

  14. I must admit I'm still a little dubious about the parametric deformer idea.  Not that I'm suggesting Qarl isn't capable, given the tremendous contributions he's made to SL over the years it's pretty obvious that he's more than capable.  It's just that, even with the additions/changes being discussed since people saw the initial version (i.e. using multiple vertices to calculate deformation rather than just the closest vertex), it's still not going to make rigged mesh "perfect".  I'm not saying that it isn't a step in the right direction, but the improvement in quality and usability of mesh products is as much reliant on creators learning to use it properly as it is any code that LL or a 3rd party can implement.

    Incidentally one particular problem I foresee relates to the point you made about mesh clothing deforming with avatar shape (or at least the parameters that define joint positions rather than the "fat" sliders that only alter the avatar mesh), since for every "right" way to make a certain bodyshape there's usually several "wrong" ways to create an almost identical looking shape by altering completely different sliders (for instance, "stubby torso/long legs" vs "stretched torso/stubby legs").  If using a parametric deformer, it would seem logical that those two almost identical looking shapes are going to have quite different results on the deformation of rigged mesh clothing.


  15. Irene Muni wrote:


    Yohan Roux wrote:

    lol, get real, those that create and sell are here for the money, end off,

    If you think
    that,
    you know not
    fully
    the reality of
    SL.

    Sadly I think a lot of people share that misguided belief, mostly thanks to LL spending years pushing the "Join SL and get rich!" myth.  It's not uncommon for people to assume that everyone elses motives for doing something are similar to their own, however it is a little narrow-minded IMO.

×
×
  • Create New...