Release: IAR80 v1.1 [2011-12-12 update]

[i4dnf]

i guess the livery select should change the texture refference in the effect... i didn't test that part yet... as the texture mapping isn't final yet. the size of the texture concerns me a bit too.. that's why i've been considering the idea of splitting it into several parts (fuselage, wings, control surfaces, gear). there isn't any constrain on texture size.. i did it that big because at lower resolutions the rivets (and fine details in general) would be too big scalewise at the same time keeping the bluriness low .
thanks for the info on yasim. now i get something like this:

Code: Select all

Solution results:       Iterations: 1058
 Drag Coefficient: 5.472822
       Lift Ratio: 99.595795
       Cruise AoA: -3.626069
   Tail Incidence: 4.599699
Approach Elevator: -0.791674
               CG: x:-2.952, y:-0.000, z:-0.045

  Inertia tensor : 8288.054, -0.000, 921.072
        [kg*m^2]   -0.000, 10506.476, 0.000
     Origo at CG   921.072, 0.000, 18355.416

the cg is a little aft .. but trying to bring it forward gets me in trouble with the solver right now.. still got a lot to tweak about the fdm.
my problem is that i get either too much elevator authority.. or too little.. can't seem to find that middle range
also.. can some one give a guestimate on the weight of a german 3bladed metalic prop cca. 1940 ('cause so far i couldn't find any data on that).

[DFaber]

You can tune the elevator lift to balance the CG shift. Reducing the Elevator lift will also bring the Approach elevator to a reasonable value (something between 0.9 and 0.99).

The Cruise AOA means the nose points 3.6 degrees downwards while in cruise flight. Maybe a bit high, unless you have evidence that it is accurate (the fw190 was reported to have a nose down attitude). High nose down aoa usually happen if you have a wing incidence and a non zero camber value in the wing.

Regarding textures, maybe it's possible to leave the texture handling to the material animation and only do the bumpmapping via shader?

Greetings

[i4dnf]

the incidence at cruise appears correct.. since it's a cambered wing, with 2 degrees of incidence. but that might also be the problem.. since i'm not quite shure if the camber is right .
that brings me to another question, i know the undersurface dihedral, but i suppose yasim considers th dihedral of the "middle" part of the wing, right ?



about the textures, yes that might be possible.. but it's way out of my league.. since it would require combining the default diffuse shader with the bumpshader (probably in two passes.. and i don't know how to do that or what impact it would have on performance).
but passing the texture to the effect would be much simpler now, since effects can inherit from each other... (and could achieve the same goal... even for those who don't want to enable the shader). i'm just thinking aloud here.. as i said i haven't yet looked into livery-select.

[DFaber]

that brings me to another question, i know the undersurface dihedral, but i suppose yasim considers th dihedral of the "middle" part of the wing, right ?

Right. But I doubt that a degree more or less makes a big difference here.

Greetings

[i4dnf]

regarding the livery-select.. i just realised it's very simple to solve, i was inheriting from the default.eff already, so it was just a matter of comenting out the texture override. this should work, in theory :

Code: Select all

<?xml version='1.0' encoding='utf-8'?>
<PropertyList>
  <name>Effects/paint</name>
  <inherits-from>Effects/model-default</inherits-from>
  <parameters>
    <texture n="2">
      <image>Aircraft/IAR80/Models/Effects/textures/iar80_bump.png</image>
      <wrap-s>repeat</wrap-s>
      <wrap-t>repeat</wrap-t>
    </texture>
    <texture n="3">
      <image>Aircraft/IAR80/Models/Effects/textures/iar80_spec.png</image>
      <wrap-s>repeat</wrap-s>
      <wrap-t>repeat</wrap-t>
    </texture>
  </parameters>
  <technique n="10">
    <pass>
      <texture-unit>
   <unit>0</unit>
   <image>
     <use>texture[0]/image</use>
   </image>
   <filter>
     <use>texture[0]/filter</use>
   </filter>
   <wrap-s>
     <use>texture[0]/wrap-s</use>
   </wrap-s>
   <wrap-t>
     <use>texture[0]/wrap-t</use>
   </wrap-t>
      </texture-unit>
      <texture-unit>
   <unit>1</unit>
   <image>
     <use>texture[2]/image</use>
   </image>
   <filter>
     <use>texture[2]/filter</use>
   </filter>
   <wrap-s>
     <use>texture[2]/wrap-s</use>
   </wrap-s>
   <wrap-t>
     <use>texture[2]/wrap-t</use>
   </wrap-t>
      </texture-unit>
      <texture-unit>
   <unit>2</unit>
   <image>
     <use>texture[3]/image</use>
   </image>
   <filter>
     <use>texture[3]/filter</use>
   </filter>
   <wrap-s>
     <use>texture[3]/wrap-s</use>
   </wrap-s>
   <wrap-t>
     <use>texture[3]/wrap-t</use>
   </wrap-t>
      </texture-unit>
      <program>
   <vertex-shader n="0">Aircraft/IAR80/Models/Effects/Shaders/paint.vert</vertex-shader>
   <fragment-shader n="0">Aircraft/IAR80/Models/Effects/Shaders/paint.frag</fragment-shader>
      </program>
      <uniform>
   <name>tex_color</name>
     <type>sampler-2d</type>
   <value type="int">0</value>
      </uniform>
      <uniform>
   <name>tex_normal</name>
     <type>sampler-2d</type>
   <value type="int">1</value>
      </uniform>
      <uniform>
   <name>tex_specular</name>
     <type>sampler-2d</type>
   <value type="int">2</value>
      </uniform>
    </pass>
  </technique>
</PropertyList>


(sadly the board attachment quota has been reached so i ca'nt attach the shaders too )

[HHS]

regarding the livery-select.. i just realised it's very simple to solve, i was inheriting from the default.eff already, so it was just a matter of comenting out the texture override. this should work, in theory :

Edit: it is working really nice! Thank you!

[i4dnf]

hmm, just tested with the ec.. and it works. maybe you forgot to change some paths in some config.. (posted at the same time as your edit )



i've changed the shaders, the normal/bump map handling that is.. to use a real (the blueish one) normalmap not a bump map.. also i've comented out some ifdefs that probably take some options specific to their code... (the specular map part has an ifdef)

[HHS]

hmm, just tested with the ec.. and it works. maybe you forgot to change some paths in some config..



i've changed the shaders, the normal/bump map handling that is.. to use a real normalmap not a bump map.. also i've comented out some ifdefs that probably take some options specific to their code...

I just edited my posting... Thanks anyway!

[HHS]

hmm, just tested with the ec.. and it works. maybe you forgot to change some paths in some config..


i've changed the shaders, the normal/bump map handling that is.. to use a real normalmap not a bump map.. also i've comented out some ifdefs that probably take some options specific to their code...

I just edited my posting... Thanks anyway!

If you want I can take the shader on my homepage unless there is no other page where you can upload it.

[fredb]

Someone (me) can commit a generic bump map effect into CVS. Don't hesitate to send me your shader and effect via PM

[i4dnf]

If you want I can take the shader on my homepage unless there is no other page where you can upload it.

Someone (me) can commit a generic bump map effect into CVS. Don't hesitate to send me your shader and effect via PM

wow, thank you both, that would be great. the effect probably needs some cleanup and some instructions.
sent pm to fredb.

[fredb]

Do you want to be compatible with v2.0.0 or with CVS ? Because CVS has now provisions to compute the tangents for us.
I will look at it when I return home.

[i4dnf]

i guess for now it's better to be compatible with v2.0.0. (if it won't be too much of a hassle to convert it later)

[Tuxklok]

Because CVS has now provisions to compute the tangents for us.

That makes me very happy

[Tuxklok]

So fred...are the tangents and what not available automagically in some form like 'attribute vec3 tangent;' ala ogre..or does one need to do something in the effect file? Looking at the source made it seem like the latter but I'm not quite certain. Or perhaps it's just not quite yet ready for public tinkering?

cheers