[DahBlount]

If that's the case then no worries.

You're pretty much right about the Iceni, it must have been made from the scrapped remains of a corvette and some cruisers as well as using the turrets of a few destroyers. The NTF is a "rebel" group and will do what they must to survive, even if it means recycling old materials that would normally be disposed of or used to repair other ships (difference between "build a new ship" or "repair old ships to keep them in service"). In the case of the GTVA, option 2 is far more likely as they already have access to industry to create more ships as well as repair and upgrade older ones. With limited materials from ships that could never be repaired, option 1 is the best bet for the NTF, because it means they can utilize equipment combinations that may even out do other older ships using the same tech (example of this in the SOC loop where you escort the Iceni to a jumpnode, the Iceni takes down a cruiser and maybe a corvette if the mission progresses correctly with only about 30-50% hull loss).

My explanation does seem strangely rationalized, it sounded better in my head but trust me it makes at least some sense.

[Mongoose]

From what we know, Bosch oversaw the construction of the Iceni while still a GTVA admiral, so it wasn't something that was just cobbled together by the NTF.  I'd imagine the GTVA initially intended it to be a general high-powered strike frigate, but then Bosch took the reins and directed the project towards his own needs.

[Lc4Hunter]

If that's the case then no worries.

You're pretty much right about the Iceni, it must have been made from the scrapped remains of a corvette and some cruisers as well as using the turrets of a few destroyers. The NTF is a "rebel" group and will do what they must to survive, even if it means recycling old materials that would normally be disposed of or used to repair other ships (difference between "build a new ship" or "repair old ships to keep them in service"). In the case of the GTVA, option 2 is far more likely as they already have access to industry to create more ships as well as repair and upgrade older ones. With limited materials from ships that could never be repaired, option 1 is the best bet for the NTF, because it means they can utilize equipment combinations that may even out do other older ships using the same tech (example of this in the SOC loop where you escort the Iceni to a jumpnode, the Iceni takes down a cruiser and maybe a corvette if the mission progresses correctly with only about 30-50% hull loss).

My explanation does seem strangely rationalized, it sounded better in my head but trust me it makes at least some sense.

I totally agree with that.
As i remember from a discussion the construciton of the Iceni was started some time before the revolution which could be the reason the ship has a uniquer design then we could expect from a rebelgroup. Maybe the ship wasn´t completely finished when the rebellion started and they had to use parts from other ships (engines, weapon-systems,...) to complete the ship.

Like the idea because it makes sense (in my head )

[Luis Dias]

From what we know, Bosch oversaw the construction of the Iceni while still a GTVA admiral, so it wasn't something that was just cobbled together by the NTF.  I'd imagine the GTVA initially intended it to be a general high-powered strike frigate, but then Bosch took the reins and directed the project towards his own needs.

Exactly.

[Lc4Hunter]

Found a way to fix the issue with the faces and uploaded the model again.
I also did some small improvements on the mesh. Currently i don´t really where i can add details and what for details because i don´t want to place some "standard" greebles...

[Hades]

http://filesmelt.com/dl/turrets.rar Here's the Orion turret and the angled Deimos turrets.


As for the mesh, it looks good. What I'd suggest is studying the retail texture patterns and trying to work that detail into the model, putting piping and mechanical greebling on the black parts on the original texture etc.

[Hellzed]

I can't wait to see this beauty emerging from the Boadicea ! (except the Boadicea itself is still pretty ugly ^^ )

[Lc4Hunter]

http://filesmelt.com/dl/turrets.rar Here's the Orion turret and the angled Deimos turrets.


As for the mesh, it looks good. What I'd suggest is studying the retail texture patterns and trying to work that detail into the model, putting piping and mechanical greebling on the black parts on the original texture etc.

The Deimos-Turrets are really pretty after converting them to work ;-)
But i prefer my Orion-turrets :-)
Added all turretpoints now but i think i have to do a special multipart for the SCP because the small one looks terrible ... small

[Nyctaeus]

Hmmm... I'm mapping Your Support Carrier for my project and I've noticed you're modeling details of your models with many geometricaly independent objects. The same method as I see here. My question is, are you going to merge them all into single geometrical object, or rather leave them as they are now?

Also, great work with both Iceni and Demon

[Lc4Hunter]

Hmmm... I'm mapping Your Support Carrier for my project and I've noticed you're modeling details of your models with many geometricaly independent objects. The same method as I see here. My question is, are you going to merge them all into single geometrical object, or rather leave them as they are now?

Also, great work with both Iceni and Demon

Hy
The support carrier? Do you mean the one i´ve rebuild from Sins of a Solar Empire?
And yes, i´m working with a lot of single objects. The reason is: it´s easier to make LODs with that method and in a lot of cases i can save polygons without merging the objects.
The "main" objects of a model do i merge, of course. The mainparts of the Iceni will be merged to prevent rendererrors ingame.

For example: i make a cube and want to add a cylinder on top:


When i merge the objects i have 60 polygons used.
If i just place them in each other, i have 48 polygons - saved 12 polygons for this example.
With complexer models you often have a much more polygons saved.
It´s maybe mot the cleanest method, but for ingame usage its good enough 

[The E]

Four things:
1. Polycount is not important.
2. Models like those may end up causing problems for the renderer and collision detection
3. For every model that is being rendered, FSO needs to invoke its render pipeline once for each submodel and each texture. Having lots of subobjects is a sure way to make inefficient models in this engine.
4. Modelling that way makes pretty sure that you're wasting space on the UV map.

[Black Wolf]

Four things:
1. Polycount is not important.
2. Models like those may end up causing problems for the renderer and collision detection
3. For every model that is being rendered, FSO needs to invoke its render pipeline once for each submodel and each texture. Having lots of subobjects is a sure way to make inefficient models in this engine.
4. Modelling that way makes pretty sure that you're wasting space on the UV map.

Really E? I mean, I know you're a coder and all, and you definitely know the engine better than I do, so if you're sure about all of this stuff then that's that, but some of this sounds very different to the advice I've received in the past, as well as the methods that I know a lot of the really good modellers around here use when going about their business (as well as myself, since I definitely wouldn't put me in that category). As long as the independent models are being merged into a single submodel before pof conversion (as opposed to being booleaned together as appears to be the case in Lc4hunter's image example), in my experience there are few of the consequences you describe. Certainly not in terms of rendering or collision detection in my experience, and not always significantly in UV space if you know what you're doing and are at least basically careful about it (doing things like deleting invisible faces and minimizing the amount of hidden polygon space).

Now, obviously we both know how much help you've been to me in the past with technical engine stuff, so I don't disagree with you lightly, but I really think you need to clarify what you mean here, because it's got some pretty big implications for how people model - it sounds a lot like the "non-manifold meshes are evil" ideas that had been, as far as I knew, entirely debunked.

[Luis Dias]

Yes I've also seen Hades and other great modelers around here being somewhat loose in this general rule with great success.

[Tomo]

It's not about manifold or otherwise.

It's about it being a separate mesh.

Every mesh to be rendered is a separate "block of data" that gets packed up and sent for rendering.
The act of sending for rendering takes is quite expensive to do, so you want to invoke the "Render Mesh N" the smallest number of times you can.

[Shivan Hunter]

No, the post Lc4Hunter made is exactly about whether the mesh should be manifold or not. It is possible to group meshes like these in a single subobject without making them a manifold mesh (I've done it), and I've also heard that manifold meshes are not entirely necessary for collision or rendering (so I'd also like The_E to clarify his post, since I consider him a very reliable source for engine advice).

(the point about wasting UV space is moot since you don't have to map hidden polygons. I always map them to a single point and shove them off to the side)

[blowfish]

I think the point about UV space is worth considering.  Let's say that you have a rectangular extrusion out of a flat face.  If the mesh is manifold, then there is an actual cutout from the flat face.  If it is not manifold, then the flat face continues behind the extrusion.  This saves a few polygons, but the invisible part of the face is also mapped to the texture but (possibly) unused.  If the mesh is manifold, then that cutout in the map can be used for something else, for instance the front of the extrusion (which is the most LOD friendly thing to do).  Though even in the non-manifold case you could map the front of the extrusion to that location (again, which will yield the best result on lower LODs), but this takes some extra work and might require some fiddling to get AO to work properly.

[Shivan Hunter]

True, if there's a significant amount of polygon space on the hidden detail. I guess my point should be made with the caveat that polygons that do have visible UV space should have as little hidden space as possible.

[mjn.mixael]

Guys, there's no 'right way' with this... Stop arguing like there is. There are pros and cons to either method and they both work just as well. Yes, you could save a little UV space. No, you won't kill the engine with little render mesh N calls or whatever Tomo is thinking. Alternatively, you could copy your little details as many times as necessary and save time modelling and UV mapping while also having your polygons match exactly and have an easier time texturing later.

I've done both. Many times. Both work juuuuuuust fine.

[Rga_Noris]

Guys, there's no 'right way' with this... Stop arguing like there is. There are pros and cons to either method and they both work just as well. Yes, you could save a little UV space. No, you won't kill the engine with little render mesh N calls or whatever Tomo is thinking. Alternatively, you could copy your little details as many times as necessary and save time modelling and UV mapping while also having your polygons match exactly and have an easier time texturing later.

I've done both. Many times. Both work juuuuuuust fine.

QFT. The Rakshasa, Sobek, and Arcadia have non-manifold portions. No collision problems. Tip though: Bake an AO map. Lets say you have a tube that extends through a flat surface. That flat surface will have a portion that isn't visible to the user. Baking an AO map will place a precise outline in the form of a shadow, allowing you to actually use that UV space. One final point: going all manifold doesn't add a few polies... it adds a ton of polies. Consider a two triangle flat surface. If you wanted an 8 sided tube coming out of the flat surface, you would need a minimum of 16 tris for the flat surface to accomplish this. Now multiply that scenario out all across your model. Yep, not worth it.

[Lc4Hunter]

Thanks mjn.mixael and Rga_Noris for the statement.
As usual i will try to merge the most post parts to an object, but i don´t really want to merge the whole ship into one single ... thing.
Another reason for haveing multiple objects are smoothing problem. Had a lot of them in the past when merging nearly everything 

Well, but one thing i have to say again: I have not much experience making textures... the only textures i´ve ever made were for some Wing-Commander ships:


And this were easy ones...
I´m not good enough to map a ship like the Iceni and i´m sure nobody has the time and pleasure to explain it 

I have to spend some more time with Photoshop and MudBox...