Rypley's Hangar
-
I don’t know if it’s useful but there are two museum F-16A cockpits in the Cockpit 360 app for iOS that you can pan around / zoom in on the detail etc.
-
More HUD info!
https://cdn.rochesteravionicarchives.co.uk/img/catalog/ZZ_1358967113_DDBR0012+(O%26A-1b).pdf
Also there is this which gives you a great idea of the width of the HUD glass+frame - 165mm.
https://rochesteravionicarchives.co.uk/collection/head-up-display-hud/f-16a-b-hud-combiner-glass -
@okayasugf said in Rypley's Hangar:
@Rypley What’s the tri count, and are you baking the normals onto the mesh? I would very much recommend that…
Tri count for both Consoles (the only thing I’ve gotten done since the start of the remodelling process, og cockpit with 80% done was around 120k) is roughly 15k, I’ll have to get back to you on exact numbers a little later today.
…And no, I haven’t touched any normals as of now, but figuring blender allows me to bake normals from meshes, I’ll start creating iterations of switches and what not in very high detail, uv unwrap bake the normals, take em and drop them on a final much lower poly version and then instance those object to avoid unnecessary repetitions in textures. (This is a very very WIP idea since I’m waiting for the BML exporter and the supported 3ds max version to test this theory, if not possible, then I’ll just export the base model and instance inside 3ds max instead)
Taking on normals and uv unwrapping id going to definitely be a helluva challenge since I’m not used to PBR, but there’s always a first time aye?
PD: idk the max tri count for bml but hopefully it can handle 200k which is what I expect the whole cockpit to be.
-
@Korbi said in Rypley's Hangar:
@Rypley Looks very promising!
Maybe you could correct the top of the toggleswitches to better match the real one. Also I’d add more details to the DZUS studs. Here are some refference pictures for you. If I can help out with more detailed pictures or overall measurements just tell me.
@Korbi said in Rypley's Hangar:
@Korbi The lightplates are screwed to the backplates and the edges of the lightplates are not directly matching the backplate edges, the lightplates have a slightly smaller footprint on the outline than the backplate.
Will do, maybe a bevel on the DZUS studs will do? hopefully once normals are added I’ll be able to minimize polygons by doing dents and similar on normal maps.
Will also remodel the panel plates based on your suggestions since it’s far easier, for measuring I’m using the xflight.de site that has almost all panel measurements (some i eyeballed them since they’re not on the site) but I’m still lacking the main “tub” measurements and angles (I could’ve swear I’ve had that picture before)
@SOBO-87 said in Rypley's Hangar:
Hey @Rypley Here are some HUD refs that might be useful for you getting it the right size etc.
Seems like the main HUD unit itself is 175mm wide.https://rochesteravionicarchives.co.uk/collection/head-up-display-hud/f-16a-b-hud-pilot-s-display-unit-1 (Page includes some dimensions that might be useful?)
Also from the extended description of that page.
“The Total Field of View of the F-16 A/B PDU is 20deg but the Instantaneous FoV is only 9deg in elevation and 13.38deg in azimuth because the combiner to eye distance is quite large due to the highly angled seat.”
The 20deg will be the total FOV that the HUD can project symbology into, say if you brought your head much closer to the HUD. But from where the pilots head is normal placed he can only see 9 x 13.3 degrees which will be the Binocular IFOV.
“Instantaneous FOV (IFOV) — The union of the two solid angles subtended at each eye by the clear apertures of the HUD optics from a fixed head position within the HUD eyebox. Thus, the instantaneous FOV is comprised of what the left eye sees plus what the right eye sees from a fixed head position within the HUD eyebox.”
I think from the above data you might be able to model where the HUD eyebox should be. I think the average IPD for males is 62mm.
Ok… So it should be more of a XYZ from the main camera thing right? I’m still wrapping my head around FOVs lol, really appreciate the info
@wazza69 said in Rypley's Hangar:
I don’t know if it’s useful but there are two museum F-16A cockpits in the Cockpit 360 app for iOS that you can pan around / zoom in on the detail etc.
It has been recommended before but nevertheless thank you! It’s still useful for basic geometry of some parts.
-
This post is deleted! -
@Rypley said in Rypley's Hangar:
This is a very very WIP idea since I’m waiting for the BML exporter and
@richionizor should I do the honors or you?
-
@I-Hawk is tri-count stated below within a limits for 4.37.u and 4.38?
@Rypley said in Rypley's Hangar:
[…]Tri count for both Consoles (the only thing I’ve gotten done since the start of the remodelling process, og cockpit with 80% done was around 120k) is roughly 15k, I’ll have to get back to you on exact numbers a little later today.
…And no, I haven’t touched any normals as of now, but figuring blender allows me to bake normals from meshes, I’ll start creating iterations of switches and what not in very high detail, uv unwrap bake the normals, take em and drop them on a final much lower poly version and then instance those object to avoid unnecessary repetitions in textures. (This is a very very WIP idea since I’m waiting for the BML exporter and the supported 3ds max version to test this theory, if not possible, then I’ll just export the base model and instance inside 3ds max instead)
Taking on normals and uv unwrapping id going to definitely be a helluva challenge since I’m not used to PBR, but there’s always a first time aye?
PD: idk the max tri count for bml but hopefully it can handle 200k which is what I expect the whole cockpit to be.
-
@Xeno said in Rypley's Hangar:
@I-Hawk is tri-count stated below within a limits for 4.37.u and 4.38?
@Rypley said in Rypley's Hangar:
[…]Tri count for both Consoles (the only thing I’ve gotten done since the start of the remodelling process, og cockpit with 80% done was around 120k) is roughly 15k, I’ll have to get back to you on exact numbers a little later today.
…And no, I haven’t touched any normals as of now, but figuring blender allows me to bake normals from meshes, I’ll start creating iterations of switches and what not in very high detail, uv unwrap bake the normals, take em and drop them on a final much lower poly version and then instance those object to avoid unnecessary repetitions in textures. (This is a very very WIP idea since I’m waiting for the BML exporter and the supported 3ds max version to test this theory, if not possible, then I’ll just export the base model and instance inside 3ds max instead)
Taking on normals and uv unwrapping id going to definitely be a helluva challenge since I’m not used to PBR, but there’s always a first time aye?
PD: idk the max tri count for bml but hopefully it can handle 200k which is what I expect the whole cockpit to be.
For a cockpit 200K is fine. 300-400K is also fine… as long as the cockpit is highly optimized regarding tris count, and also (not less important, even more) draw calls, it’ll be fine.
-
A good consideration to have details on low polygon is to use the High-Low workflow.
For example a hard edge like a crew head can look like a chamfered smooth one with this workflow. and many other small details. -
@I-Hawk Perfect! Cockpit will have faces on all ends (even on objects behind the camera) to avoid for any shadow glitch, although some of those will be very low poly since its just to avoid glitched shadow casts. on draw calls, I suppose that’s just how to mesh is going to display in sim right?
@Arty said in Rypley's Hangar:
A good consideration to have details on low polygon is to use the High-Low workflow.
For example a hard edge like a crew head can look like a chamfered smooth one with this workflow. and many other small details.Thanks, do you perhaps have more info on that subject?
-
Hello all! Right now I’m still battling with UVs and making the first instance of most of the objects that are going to be multiplied like switches, knobs and etc.
Right now I’m still thinking how many textures I’ll end up with, but I plan to have:
1 for the whole tub/glareshield (including the canopy frame and back parts)
1 for the all the console panels and gauges
1 for the light up labels and gauges
1 for the whole canopy assembly (glass, reflection, side walls, motors and etc.
1 for the ACES II seat
1 for all the switches, knobs, guards and any other movable part like the armrests.
and finally 1 for the SSC and TQSDo let me know if there’s a limit on how many textures can be appended into BMLs or into the engine itself.
-
Love it!
-
Who would’ve thought normals also induced a “smooth” effect on surfaces? Definitely not me lol,
I’m sure that almost everyone here knows about the High-Low poly workflow (As Arty suggested) but one way to make the process better in blender is to ‘divide’ the topology process in 3 parts.
The first being the creation of the hi-poly model that we want to create the normal map from,
Then create a middle-poly model that has roughly the same characteristics as the hi-poly model from which we will UV unwrap to bake the normal map,
And finally the lo-poly model that we will also UV unwrap and then contort the UV islands to those of the middle-poly model to match the normal map.Why do this? Sometimes blender makes errors when baking normal maps with a very low poly model (the one we actually want to implement) and end up with jagged edges or undulated lines that should otherwise be fully straight. Should you do this for every model? I’ve come to find that no, you don’t need to, but you can fallback the such a method when the usual workflow yields you incorrect normal maps.
Uncolored models are the high poly models and the colored ones are low poly
This the low poly model with the same starting point so the baked normals turn out fine
This is the final low poly model that we now can conform to the previously baked normal map.Hopefully, this can help those who are like me: learning on the go or are having a go at blender and don’t know about PBR or UV unwrapping. Please let me know if I’m wrong on anything written above!
-
just some small tips or info.
beign on DX for normals and in tangent color space is the thing for normals and BMS. blender is on OpenGL by default iirc.
high poly model on the high-low poly workflow doesn’t need uvw unwrap or mapping at all at least with substance painter (hope it’s the same for blender).
Using substance painter , again without actual knowledge on blender texturing, is quick, it helps with some anomalies or errors to be discovered in the transition and baking the basic maps from PT (PT=substance painter) also helps alot to find such hidden issues or have a great result.
The negative on the high-low poly workflow, is more time on the model and very good organization in order to be able to make changes if needed afterwards so not to search in a mess and mostly important to have a clean project as all this work will pass to BMS eventually and someone else in the future might have to deal with your model, so it must be clear in order to find his way in that.
using the High-Low poly workflow gives a great potential for more detailed models while you save triangles. But it introduces an important factor which is Texel-density. This is even without the term known to artists but with this workflow it comes even more important. So you might save tris in the model but in the UVW real estate you need more pixels, so your uv-islands (parts of the uvw of the 3D part) you must make them bigger , or increase the resolution or brake the texture from 1 to many. Here kicks in TexelDensity. This can assist if used correctly to determine Texel Density for your model parts and separate them to texel density groups and have an acceptable for u result without for every part going back and forth to fix things and maybe at the end to have to redo everything from scratch and save time.
Another important thing has to do with non English creators that use their native language to name parts in the 3D model.
So please use English and easy to understand names, example not “the upper thing” or “leva” or “changer” or whatever native word besides English.
If u ask well what we do with multipart objects? easy “piston-main” “piston-main-1” “piston-main-2”. I know it sounds harsh but besides future guys will also help you on a very complex model. -
@qawa any words on this?
I mean for the F-16A pit not for the workflow. -
You have teached me how to use PT, and I still not using Blender… so what you recommend to Rypley is the best! … our BMS community will have a perfect F16A, so I can abandon my idea of converting a half-built, incomplete cockpit, and I would need permission from the 3D modeler to distribute it… I can’t reach him anymore via email… At the moment I’m dedicate my time to a new project for 4.38.
The work Rypley is doing is wonderful, I open this thread every time to see his progress !
Rypley… if you need material photos, manuals, or anything else feel free yourself to send me a PM !PS: I finished editing the post now !!! lol
-
@Arty said in Rypley's Hangar:
just some small tips or info.
beign on DX for normals and in tangent color space is the thing for normals and BMS. blender is on OpenGL by default iirc.
high poly model on the high-low poly workflow doesn’t need uvw unwrap or mapping at all at least with substance painter (hope it’s the same for blender).
Using substance painter , again without actual knowledge on blender texturing, is quick, it helps with some anomalies or errors to be discovered in the transition and baking the basic maps from PT (PT=substance painter) also helps alot to find such hidden issues or have a great result.
The negative on the high-low poly workflow, is more time on the model and very good organization in order to be able to make changes if needed afterwards so not to search in a mess and mostly important to have a clean project as all this work will pass to BMS eventually and someone else in the future might have to deal with your model, so it must be clear in order to find his way in that.
using the High-Low poly workflow gives a great potential for more detailed models while you save triangles. But it introduces an important factor which is Texel-density. This is even without the term known to artists but with this workflow it comes even more important. So you might save tris in the model but in the UVW real estate you need more pixels, so your uv-islands (parts of the uvw of the 3D part) you must make them bigger , or increase the resolution or brake the texture from 1 to many. Here kicks in TexelDensity. This can assist if used correctly to determine Texel Density for your model parts and separate them to texel density groups and have an acceptable for u result without for every part going back and forth to fix things and maybe at the end to have to redo everything from scratch and save time.
Another important thing has to do with non English creators that use their native language to name parts in the 3D model.
So please use English and easy to understand names, example not “the upper thing” or “leva” or “changer” or whatever native word besides English.
If u ask well what we do with multipart objects? easy “piston-main” “piston-main-1” “piston-main-2”. I know it sounds harsh but besides future guys will also help you on a very complex model.Thanks for the recommendations Arty, I do appreciate them, Will do on the last tip so my posts become clearer
PD: yes, you do need to UV unwrap on blender to bake normals, tedious but it helps to already have the means to do a rough sketch of the model’s silhouette in a small file so i get to know ‘what is what’ when I’m on PT or PS.But I’ll try the workflow with PT as the normal baker to see how it goes, maybe I can even paint the models and import them to blender with textures already, I’ll have a go at it.
@qawa said in Rypley's Hangar:
You have teached me how to use PT, and I still not using Blender… so what you recommend to Rypley is the best! … our BMS community will have a perfect F16A, so I can abandon my idea of converting a half-built, incomplete cockpit, and I would need permission from the 3D modeler to distribute it… I can’t reach him anymore via email… At the moment I’m dedicate my time to a new project for 4.38.
The work Rypley is doing is wonderful, I open this thread every time to see his progress !
Rypley… if you need material photos, manuals, or anything else feel free yourself to send me a PM !PS: I finished editing the post now !!! lol
Thanks!! Will definitely PM soon
-
This project is in progress but the day-to-day work affects a lot… Good things will come soon for the BMS community
w
-
Greetings to everyone! I am planning to announce an update regarding the F-16A No MLU Cockpit. Together with my friend Ripley, we have collaborated on launching a beta version 1.0 of the cockpit in the coming weeks. However, a crucial requirement for the aircraft is the design of special avionics.
The FCNP (FIRE CONTROL NAVIGATION PANEL) is an older avionics system that served a similar function until it was replaced by the DED. I am familiar with its operation and manual, but I lack the expertise to design a comparable system. I hope that explanation clarifies things.
I am open to anyone who can assist me…
-
Awesome! Looking promising