SSU Development thread (4.0 to 5.0) [DEVELOPMENT HALTED DUE TIME REQUIREMENTS!]

[DaveS]

Yeah, that would be nice, especially if we could maybe add some scripts to get to a full downloadable ZIP file for a nightly build to download.

That was my thinking actually, to automate the process of creating nightly builds. Maybe we could ask the NASSP team for their scripts.

---------- Post added at 10:54 AM ---------- Previous post was at 10:40 AM ----------

About the mesh corrections:
> the inside of the outboard RSB panel still have mapping issues (it seems like nothing changed);
> the bottom of the tail, just aft of the drag chute door, has a mapping issue where something other than tiles is showing;
> the original tail has the same mapping issue, plus the sides around the bottom aren't correct;
> the body flap base still doesn't connect well to the aft compartment;
> the front and top of the OMS pods looks uneven (not (just) mapping), and the area in front of the RCS pods has wierd lighting when the Sun is from behind.

What about the hidden faces, can I remove then?

The mapping issues stems from this:

The reason for the wrong looking textures is that the orbiter is mapped to suit Wolf's hi-res photo-realistic textures. As those are different to ours, it makes ours incompatible in certain areas. The best solution would be for us to include lower resolution versions of those textures, this of course hinges on Wolf granting us permission to do so.

You can go ahead remove the faces and fix whatever issues you find.

 

[Wolf]

The mesh issues should be addressed now. The reason for the wrong looking textures is that the orbiter is mapped to suit Wolf's hi-res photo-realistic textures. As those are different to ours, it makes ours incompatible in certain areas. The best solution would be for us to include lower resolution versions of those textures, this of course hinges on Wolf granting us permission to do so.

I don't understand: What would be the point in having a lower res version of my textures?
Regarding the standard versus hi-res there might be very slight differences between the textures (I tried to null them at my best) but most of the issues popped up due to the low res textures masking them, so they do not show at low res but become evident at higher resolution and/or if more details are added to certain area. I'll post later an example where you can clearly see there are some issues with the OMS PODs which are hidden if you use a low res texture or a less detailed one.


I am afraid there will inevitably always be some differences between the standard and the hi-res textures.
Problem from my side is I cannot change mine everytime a new updated mesh is released.
So maybe we should pick either one as the "official" texture for SSU and work on it...

 

[Wolf]

Notice the seaming line is cut and the front area LRSI tiles distortions




Same area different view (this is evident even if you use DaveS OMS pod textures)




Also it looks like there are two bumps

 

[Wolf]

Don't want to flood this thread with textures pics, so I will post the lates issues on the related one

 

[GLS]

DaveS, does your program have a "solid" view? Because there are parts that are so uneven that it's scary... :uhh:
E.g., the OMS pods, as noted above by Wolf, and yesterday by me, have creases at the top and forward faces.

Also, there is something wrong with the tails, as blender can't save them, even before I touch them. :facepalm:

The removal of the hidden faces will take sometime, so could I have a day or two alone with the mesh to "do my thing"?

---------- Post added at 12:11 PM ---------- Previous post was at 11:16 AM ----------

The removal of the hidden faces will take sometime, so could I have a day or two alone with the mesh to "do my thing"?

Make it 3 or 4 days... the SSMEs alone will take over a day... are the engines really equal (they should)? If so I could "clean" only one of them then it could be cloned to replace the other 2.

 

[DaveS]

DaveS, does your program have a "solid" view? Because there are parts that are so uneven that it's scary... :uhh:
E.g., the OMS pods, as noted above by Wolf, and yesterday by me, have creases at the top and forward faces.

Also, there is something wrong with the tails, as blender can't save them, even before I touch them. :facepalm:

The removal of the hidden faces will take sometime, so could I have a day or two alone with the mesh to "do my thing"?

---------- Post added at 12:11 PM ---------- Previous post was at 11:16 AM ----------



Make it 3 or 4 days... the SSMEs alone will take over a day... are the engines really equal (they should)? If so I could "clean" only one of them then it could be cloned to replace the other 2.

The SSMEs are pretty old, I think Donamy made them way back. And yes, they're identical. the left engine is just clocked 90°s CCW. Take all the time you need, there's no rush.

 

[GLS]

Early estimates on SSME size reduction:
vertex: 37% :OMG:
triangles: 19%
Even if it is overestimated, that is A LOT, especially when each engine has almost 20K vertexes and another 20K triangles.

---------- Post added at 01:42 PM ---------- Previous post was at 01:40 PM ----------

The SSMEs are pretty old, I think Donamy made them way back. And yes, they're identical. the left engine is just clocked 90°s CCW. Take all the time you need, there's no rush.

Old or new, it doesn't matter. It just that the pipes have "vanes" inside, which we can't see and even if we could, the real engine didn't have them... :shrug:

 

[Donamy]

Sometimes, smoothing and optimizing meshes, create artifacts. It's a AC3D thing.

 

[DaveS]

New OMS pod, are these any better than the old ones?

https://www.dropbox.com/s/0gyx1l4xodvibbg/SSU_NEW_OMS_POD.jpg?dl=0
https://www.dropbox.com/s/seyzyffr5jpu9bt/SSU_NEW_OMS_POD2.jpg?dl=0

 

[GLS]

Sometimes, smoothing and optimizing meshes, create artifacts. It's a AC3D thing.

Well, it's a very bad thing as I'm still not finished with the SSME, and we already have vertex count decrease of 25%.

New OMS pod, are these any better than the old ones?

https://www.dropbox.com/s/0gyx1l4xodvibbg/SSU_NEW_OMS_POD.jpg?dl=0
https://www.dropbox.com/s/seyzyffr5jpu9bt/SSU_NEW_OMS_POD2.jpg?dl=0

If they fit well and have smooth surfaces, then yes.

 

[Donamy]

My mesh shows 9070 vertices per engine.

 

[GLS]

My mesh shows 9070 vertices per engine.

That is not what is in the trunk. The current count is 18388 vertices and 18736 triangles.

So I finally finished deleting the hidden faces in the SSME, and the result is 13154 vertices and 15933 triangles, so ~28% less vertices and ~15% less triangles.
Multiply that by 3, and for the complete model we get a reduction, just with the work in the SSMEs, of ~8% for the vertices and ~5% for the triangles. Giving that the graphics are the long pole in the tent, I think this is a pretty good exercice.

 

[Donamy]

The mesh I have was given to me by DaveS, a few days ago.

Did you weld the vertices ?

 

[GLS]

The mesh I have was given to me by DaveS, a few days ago.

Then your program isn't counting well... :uhh:
Here is a comparison of sizes for recent versions:

SSU 4.2:   GEOM 21905 18736
trunk:     GEOM 18388 18736
post-diet: GEOM 13154 15933

Did you weld the vertices ?

All I'm doing is selecting the faces that are not visible (e.g. cylinder covers that are inside other parts) and deleting them. Right there the triangle count drops, and if their defining vertices aren't used for the remaining triangles, they go as well.

 

[DaveS]

GLS: Doanmy is counting very accurately. 9070 vertices and 18736 for one SSME. The reason why it's higher in your version is that when you export to msh, the script unwelds all the vertices (yes, it's very much a pain to reweld them when you use the import script) leading to a much higher count.

There's your explanation, none of the msh importer scripts that exists today keeps the welds.

 

[GLS]

GLS: Doanmy is counting very accurately. 9070 vertices and 18736 for one SSME. The reason why it's higher in your version is that when you export to msh, the script unwelds all the vertices (yes, it's very much a pain to reweld them when you use the import script) leading to a much higher count.

There's your explanation, none of the msh importer scripts that exists today keeps the welds.

With blender, that uses a .msh import/export script by vlad, the count that it shows is the same as what is in the mesh file... and in the end that's what Orbiter loads... :shrug:

From my experience with blender, I think the vertices only get separated when they are shared by 2 (or more) triangles with different UV mappings, and that is something that cannot be avoided (and it happens during UV mapping). Anyway, in the final product they all count.

I just tested on an original SSME the blender operation to remove double vertices, and the count drops to 9070, but when saving it goes up to 10031... still better than the +18K you two get.

 

[Donamy]

I can optimize the mesh to a lot less, but you get less detail.

 

[GLS]

I'll will either get a medal for this, or be shot...
So, finished deleting the hidden faces of the Orbiter mesh. Here are the numbers:
> before
181308v
165785t

> now (still with 2 old SSMEs)
166093v (-8.4%)
155336t (-6.3%)

The left and right SSMEs have to be replaced with the "lighter" center SSME.
Most of the deleted triangles where "cylinder caps". The PLB sill, both tails, and the already mentioned SSMEs, were totally insane! I was relatively conservative, so more could be removed without any changes to what is visible. Didn't touch the handrails, as they have a hole... :shrug:
There should be no (new) problems, but if something is wrong, I've kept the order of the groups, so isolating things and reverting them should be easy.

Now the hard part: I found a number of issues, doubts, whatever, that I'd like to be addressed.
(the forum needs a spoiler control... :uhh

* ET umbilical doors rollers are hovering above doors
* ET umbilical doors drive arms/links/rods/whatever don't connect properly to where they should
* payload bay liner extends forward too much and punches out on the sides of the crew compartment
* holes in aft rad hoses
* hole on one side PLB handrails
* merge BODYFLAP and BODYFLAP_SIDES groups?
* aft compartment aft face doesn't connect to bottom face
* body flap base has too many faces in some areas, and is missing some in others (or they aren't in the correct place)
* side hatch window frame (or glass) not round
* side hatch window glass not parallel to hatch
* holes between the OMS pods and the PLBDs
* gear wheels stick above gear wheel well (tires too big? wrong gear position? wrong well size?)
* holes between main gear well and bottom of vehicle
* nose gear doors are not correct (should be of different sizes)
* nose gear doors have holes between sides and bottom
* PANTILT_CAM_A has a misplaced vertex, shared with other part (somehow I can't edit this, so it could/should be replaced with another camera)
* PLB camera supports don't extend all the way into the bottom blanket, and all the way to the PTU, and blankets are hovering above base
* PLB camera B base is attaching to bulkhead thru handrail
* FWD_HOOKS, MID_FWD_HOOKS, MID_AFT_HOOKS and AFT_HOOKS have double faces, a hole and a curved edge gone wrong?
* STBD_PLBD_INT has a missing drive cylinder on aft hook?
* on STBD_PLBD_INT, defect on cylinder cover, just aft of 3rd hook counting from aft?
* on STBD_PLBD_INT, the hook supports extend too much outboard are visible behind the structure
* aft PLB winch has part of a handrail inside
* 4 holes in handle of BayPFR
* are PLBD forward bulkhead latches correct? if so, many faces could be deleted
* PLBD forward bulkhead latches don't "connect" to the door as the ones in the aft bulkhead do (or the aft ones connect too much?)
* are fwd rads really closer to the PLBDs than the others? (starboard one collides with PLBD ribs...)
* aft RCS cones and holes like the FRCS?
* aft port MPM hits PLBD drive arm
* RMS base hits sidewall when tilted out
* PLBD animations need correct coordinates
* from an aft angle, it is possible to see thru the PLBD aft end
* fwd margin of tail is disconnected from main tail (visible from behind)
* RSB panels should be flush together when closed
* tails aft bottom doesn't connect to aft compartment
* tails has holes on sides of bottom aft tip
* tails are visibly not symmetrical
* eye-lid for OMS engines, to avoid visible "ring" inside?
* midbody vent doors are not lined up with cavity
* in FUSELAGE, unused triangles between chines and crew compartment?
* in FUSELAGE, unused triangles inside boxes at wing tips?
* holes between wing tops and tips
* holes between the side and bottom of the elevons
* holes at the aft face of the wings, between the elevons
* should tubes at bottom of the elevons all line up?
* holes in tubes of ELEVON_HINGE_BLOCKS
* almost all of the PLB_TCS group is hidden behind other parts, is it (all) needed? the only visible part collides with other group in bay 12/13
* kinks in upper fuselage, behind windows 1/6
* holes between PLBDs and upper fuselage
* kinks above and below FRCS down nozzles
* holes between FRCS sides and lower fuselage
* some kinks and holes around the nose cap, at the 5 and 7 o'clock positions when viewed from the front
* aft winch handle is closer to mechanism than forward one
* bases of cones on side of winches are different between fwd and aft bulkheads, is this correct?
* PLB windows don't fit the frames
* PLBD handrails have 2 triangles "cutting" each support, is this correct?
* PLBD handrails have hidden box behind each support, is this needed?
* PLBD hinges are visible inside the PLB
* in PLB_SILL_EQUIPMENT, the PLBD drive supports have boxes at the top that probably aren't needed as they are hidden. they also don't have faces outboard, while similar structures have... they probably aren't needed
* new hose roller supports collide with bridgerails, and rollers with rads
* some of the holes for the PLBD drive bars are not in correct place and collide with vertical supports
* PLB_TCS_LINER face between bay 12/13 and curved PLB, just aft of the last curved frame, has a triangle that sticks above that frame

 

[DaveS]

Status as of this post:



* payload bay liner extends forward too much and punches out on the sides of the crew compartment [DONE]
* holes in aft rad hoses [TO BE COMPLETED]
* hole on one side PLB handrails [DONE]
* merge BODYFLAP and BODYFLAP_SIDES groups? [DONE]
* aft compartment aft face doesn't connect to bottom face [DONE]
* body flap base has too many faces in some areas, and is missing some in others (or they aren't in the correct place) [OLD HINGE DELETED, NEW ONE IN WORK]
* side hatch window frame (or glass) not round [TO BE COMPLETED]
* side hatch window glass not parallel to hatch [TO BE COMPLETED]
* holes between the OMS pods and the PLBDs [DONE, NOT AN ISSUE WITH NEW OMS PODS]
* gear wheels stick above gear wheel well (tires too big? wrong gear position? wrong well size?) [FIXED]
* holes between main gear well and bottom of vehicle [FIXED]
* nose gear doors are not correct (should be of different sizes) [FIXED]
* nose gear doors have holes between sides and bottom [FIXED]
* PANTILT_CAM_A has a misplaced vertex, shared with other part [FIXED]
* PLB camera supports don't extend all the way into the bottom blanket, and all the way to the PTU, and blankets are hovering above base [FIXED]
* PLB camera B base is attaching to bulkhead thru handrail [FIXED]
* FWD_HOOKS, MID_FWD_HOOKS, MID_AFT_HOOKS and AFT_HOOKS have double faces, a hole and a curved edge gone wrong? [TO BE INVESTIGATED]
* STBD_PLBD_INT has a missing drive cylinder on aft hook? [TO BE INVESTIGATED]
* on STBD_PLBD_INT, defect on cylinder cover, just aft of 3rd hook counting from aft? [TO BE INVESTIGATED]
* on STBD_PLBD_INT, the hook supports extend too much outboard are visible behind the structure [TO BE INVESTIGATED]
* aft PLB winch has part of a handrail inside [FIXED]
* 4 holes in handle of BayPFR [TO BE INVESTIGATED]
* are PLBD forward bulkhead latches correct? if so, many faces could be deleted [TO BE INVESTIGATED]
* PLBD forward bulkhead latches don't "connect" to the door as the ones in the aft bulkhead do (or the aft ones connect too much?) [TO BE INVESTIGATED]
* are fwd rads really closer to the PLBDs than the others? (starboard one collides with PLBD ribs...) [HERE I THINK THAT THE FACT THERE'S NO RIBS FOR THE AFT PANELS THROW THINGS OFF. COLLISION BETWEEN PANEL 1 STARBOARD DOOR HAS BEEN FIXED]
* aft RCS cones and holes like the FRCS? [MAYBE, DEPENDS ON HOW WELL AC3D COOPERATES]
* aft port MPM hits PLBD drive arm [POSITIONS OF MPMS COULD NEED TWEAKING]
* RMS base hits sidewall when tilted out [POSITION COULD NEED TWEAKING]
* PLBD animations need correct coordinates [TO BE COMPLETED]
* from an aft angle, it is possible to see thru the PLBD aft end [FIXED WITH NEW OMS PODS]
* fwd margin of tail is disconnected from main tail (visible from behind) [FIXED]
* RSB panels should be flush together when closed [FIXED]
* tails aft bottom doesn't connect to aft compartment [FIXED]
* tails has holes on sides of bottom aft tip [FIXED]
* tails are visibly not symmetrical [FIXED]
* eye-lid for OMS engines, to avoid visible "ring" inside? [TO BE COMPLETED]
* midbody vent doors are not lined up with cavity [FIXED]
* in FUSELAGE, unused triangles between chines and crew compartment? [TO BE COMPLETED]
* in FUSELAGE, unused triangles inside boxes at wing tips? [TO BE COMPLETED]
* holes between wing tops and tips [TO BE COMPLETED]
* holes between the side and bottom of the elevons [FIXED]
* holes at the aft face of the wings, between the elevons [FIXED]
* should tubes at bottom of the elevons all line up? [YES, ELEVON HINGES, FIXED]
* holes in tubes of ELEVON_HINGE_BLOCKS [FIXED]
* almost all of the PLB_TCS group is hidden behind other parts, is it (all) needed? the only visible part collides with other group in bay 12/13 [LINER COVERS ALL PARTS OF THE BAY]
* kinks in upper fuselage, behind windows 1/6 [CAN'T FIND IT, A SCREENSHOT WOULD HELP]
* holes between PLBDs and upper fuselage [NO ISSUES HERE]
* kinks above and below FRCS down nozzles [FIXED]
* holes between FRCS sides and lower fuselage [FIXED]
* some kinks and holes around the nose cap, at the 5 and 7 o'clock positions when viewed from the front [FIXED]
* aft winch handle is closer to mechanism than forward one [TO BE FIXED]
* bases of cones on side of winches are different between fwd and aft bulkheads, is this correct? [TO BE INVESTIGATED]
* PLB windows don't fit the frames [TO BE INVESTIGATED]
* PLBD handrails have 2 triangles "cutting" each support, is this correct? [TO BE FIXED]
* PLBD handrails have hidden box behind each support, is this needed? [TO BE FIXED]
* PLBD hinges are visible inside the PLB [THIS IS CORRECT: https://www.dropbox.com/s/4r66n71rbat4sm1/s125e007236.jpg?dl=0]
* in PLB_SILL_EQUIPMENT, the PLBD drive supports have boxes at the top that probably aren't needed as they are hidden. they also don't have faces outboard, while similar structures have... they probably aren't needed [TO BE FIXED]
* new hose roller supports collide with bridgerails, and rollers with rads [TO BE FIXED]
* some of the holes for the PLBD drive bars are not in correct place and collide with vertical supports [TO BE FIXED]
* PLB_TCS_LINER face between bay 12/13 and curved PLB, just aft of the last curved frame, has a triangle that sticks above that frame [FIXED]

 

[GLS]

* aft RCS cones and holes like the FRCS? [MAYBE, DEPENDS ON HOW WELL AC3D COOPERATES]

I can try this one in blender if you want.

* aft port MPM hits PLBD drive arm [POSITIONS OF MPMS COULD NEED TWEAKING]
* RMS base hits sidewall when tilted out [POSITION COULD NEED TWEAKING]

Just checked the big diagrams and there are X0 positions for the MPMs (679.5, 911.05, 1189 and 1256.5), and there's also a diagram that shows a collision between the MPM and the PLB edge... :uhh:

* from an aft angle, it is possible to see thru the PLBD aft end [FIXED WITH NEW OMS PODS]

This is a door issue, not OMS pods, as it shows up when the doors are open.

* almost all of the PLB_TCS group is hidden behind other parts, is it (all) needed? the only visible part collides with other group in bay 12/13 [LINER COVERS ALL PARTS OF THE BAY]

So the PLB_TCS group is useless?

* kinks in upper fuselage, behind windows 1/6 [CAN'T FIND IT, A SCREENSHOT WOULD HELP]

This type of view (called Solid in blender) shows the forward fuselage is kinda messy.

* holes between PLBDs and upper fuselage [NO ISSUES HERE]

Took a closer look and there are 2 triangles in the forward bulkhead that stick above the OML, right in that flat part of the PLBDs, just aft of the kink in the image above.

* PLBD hinges are visible inside the PLB [THIS IS CORRECT: https://www.dropbox.com/s/4r66n71rbat4sm1/s125e007236.jpg?dl=0]

I was referring to this:

It's small and +/- out of view, but it another thing that could be deleted (only the offending triangles).
Also visible in the image are the "things" in the handrail supports, and a newly discovered hole in the PLB aft corners (both sides).

* in PLB_SILL_EQUIPMENT, the PLBD drive supports have boxes at the top that probably aren't needed as they are hidden. they also don't have faces outboard, while similar structures have... they probably aren't needed [TO BE FIXED]

Just looked at some photos and the "structures" in the sill, only the ones that are NOT in pairs and have the details at the bottom, go to that "ceiling" and in the model they go thru, and as they stick (correctly) a bit inboard, their tops should at the height of the "ceiling".