FSX Native Douglas A-20 Havoc Released

[wellis]

OK...I pulled an all nighter on the bump maps and I think I have a bump profile that works nicely.

The photos don't really show the effect well, but the flush rivets create a very subtle flat reflection at the bottom of the dimples. The skin itself has a slight stretch in the channels created by the rivet rows.

Have a look at the photos and shoot me some feedback.

I have to thank Mr. Terrell for pointing out the issue of the flush rivets. It is always work to go back and make changes, but at the end of the day, the work needs to be correct.
I certainly don't mind the extra effort, and find the changes to be a major improvement to the model. I'm glad we could resolve things without any major rework of the PK.


Note:

I did not re-size the actual painted rivets in the photos here. I only removed the highlights. This skin is wrong for this model which has no cheek blisters, but is sufficient for the color/light study it was intend for. The parameters for alphas and spec remain the same as I noted in the previous post. I'll make a dedicated metal skin when any modeling changes are finalized.

I think the team can stand down on the PK revamp. If a painter is selecting the multiple layers of rivets for any given section, i.e. wing, they can adjust the mask by one or two pixels to achieve the look they desire.

I recommend no highlights, unless the artist wishes to add a very slight highlight and shadow arc on the extreme edge of the flush rivet. I've tried this at 4K, and the results are a tiny bit muddy. The only way to achieve a clean highlight is to use a hard edged round brush, then mask out all but the "thumbnail moon" crescents for light and dark. The downside of this technique is that there is no way to select all of the rivets and do the operation via a mask. It is do-able...but a ton of work.

Amazing Gordon! And thank you for your effort many times over. It looks great.

[huub vink]

Looks really great Gordon. I had not thought the effect of a bump map was so significant

Cheers,
Huub

[Roger]

Thank you for your labours Gordon. The effect is magnificent

[Duckie]

Gordon, really nice adjust for the rivets. Looking forward to giving them a try. Thank you very much.

Now just rolled Daisy Mae into the paint hanger. Tons yet to do but she is now officially taking paint!

[jankees]

wow Gordon, that looks just brilliant! Outstanding work!
I'd better throw away my bump maps...

[gman5250]

wow Gordon, that looks just brilliant! Outstanding work!
I'd better throw away my bump maps...

Took two tries, but I got it right...I think.

Just keep in mind that any rivet details you want to develop into the texture will get a bit of a hand with the bumps. The rivets themselves are flat, with the exception of the fasteners.
I'm just finishing a video that pans the airplane on the ground. It will show the profiles in motion.

Thanks for the Atta Boy...nice to see you back.

[tankerguy72]

Great work Gman! Impressive!

[gman5250]

The screen shots just don't do the bump maps justice, so I thought a video would show the plane better.

These are the new flush rivet normal maps on the metal skin model. The skin is still a basic color study and adjustments will most likely be needed in spec alpha and other areas, but it is effective for showing the contours.

The entire texture set, including the bumps is done at 4K DXT5 resolution. I initially resized the entire paint kit to 4K before beginning any work. The small detail holds really well using this approach.
I used a trick or two on the alpha channels for the metal skin. I used a slightly different alpha value for the rivets themselves to produce the lighting differences as the point of view moves.

For the painters, the bumps should do most of the heavy lifting in Prepar3D. The light engine will catch the edges of the flush rivets due to the nature of the map itself. I put a perfectly flat spot at the center of each dimple, which will generate a subtle highlight or shadow, depending on movement. On a matte skin the effect will still work in relation to the dull finish, eliminating the need for a bunch of detailing in the baked rivets in the paint work.

The effect won't be as pronounced in FSX. If you have Steve's DX10 fixer the effect should be a bit more lively.

Let me know what you think of the bump profiles.
If we sign off on this set I'll go ahead and publish the new bump sets.

[magoo]

The scale of fasteners looks pretty darned good to me in the video. Over-all, I'd say you've knocked it right out of the park.

Thoroughly outstanding.

[jankees]

I love these rivets, bring them on I say.
And your metal work is not too shabby either, makes me wonder why I even try....

[roger-wilco-66]

Fantastic work, Gordon. I think this looks extremely realistic. I like the way the immediate area around the flush rivets seems to be under stress from the rivet.

Cheers,
Mark

[warchild]

incredible GMan.. simply incredible... I hope i can look forward to flying this..

[Milton Shupe]

Gordon, Jankees, are you happy with the reflective properties I set up for this test model?

Are there any tweaks that need to be done?

Have you done anything with the Fresnel ramp? The one in the release is a "neutralized" one. You might try one from a default aircraft that has proper fall-off.

Any other changes regarding material properties?

Otherwise, we will call it good for this variant of the "C" model and I will add it to the new models (with painted nose and nose gun configurations) as I release them.

----------------------------------------------------------------

And for all you painters out there, is there any need to do the other "C" variants.
I was thinking that this is the appropriate model without the gun pods for this test and did not see a need beyond that.

I do not want to change the material properties for all the variants because it has a slight specular affect on the flat skins that may be undesirable.

[MrZippy]

Gordon, Jankees, are you happy with the reflective properties I set up for this test model?

Are there any tweaks that need to be done?

Have you done anything with the Fresnel ramp? The one in the release is a "neutralized" one. You might try one from a default aircraft that has proper fall-off.

Any other changes regarding material properties?

Otherwise, we will call it good for this variant of the "C" model and I will add it to the new models (with painted nose and nose gun configurations) as I release them.

----------------------------------------------------------------

And for all you painters out there, is there any need to do the other "C" variants.
I was thinking that this is the appropriate model without the gun pods for this test and did not see a need beyond that.

I do not want to change the material properties for all the variants because it has a slight specular affect on the flat skins that may be undesirable.

Looks like all good stuff! Any chance for the G model??

[gman5250]

Gordon, Jankees, are you happy with the reflective properties I set up for this test model?

Are there any tweaks that need to be done?

Have you done anything with the Fresnel ramp? The one in the release is a "neutralized" one. You might try one from a default aircraft that has proper fall-off.

Any other changes regarding material properties?

Without messing with the materials properties, I think the skin is workable. It is a bit high contrast, as seen in the video, with the first run alpha values I used...around 50% gray..., and I haven't worked with any of my Fresnel ramps yet, but I think your setup will allow for a good range of metal effects.

Taking into account that I am working in P3Dv4 64-bit with all of the DX11 bells and whistles, I'd say your setup is going to be stable for all the intended applications.
For the video, and my work in general, I try to get the skins good with Dynamic Lighting features disabled and HDR lighting activated. If the skins look good there, they tend to look good everywhere else too.

I want to set up a service skin, similar to the F7F-1 "Test" variant I did, with a worn look to check your setup at those settings. That skin will use reduced alpha with built in wear.

At this point it looks good from my chair sir.

[Milton Shupe]

Without messing with the materials properties, I think the skin is workable. It is a bit high contrast, as seen in the video, with the first run alpha values I used...around 50% gray..., and I haven't worked with any of my Fresnel ramps yet, but I think your setup will allow for a good range of metal effects.

I want to set up a service skin, similar to the F7F-1 "Test" variant I did, with a worn look to check your setup at those settings. That skin will use reduced alpha with built in wear.

At this point it looks good from my chair sir.

Thank you for the feedback. I am in no hurry to release this; just curious.

Others may download and experiment with the model.a20rur.zip as they wish (post #141 above). I'll await further feedback once its had some time to be experienced, weathered, and abused. :-)

The other part of the package, the paint-able nose and gun configs, Duckie is working with and that will take some time to see if all is well there.

[Milton Shupe]

Looks like all good stuff! Any chance for the G model??

Let's see if we can get the "C" model settled in and what adjustments may be necessary before we move to the "G".

Not sure if there are "G" variants out there calling for bare metal but I'm sure if there are, painters will surface them. :-)

[MrZippy]

Let's see if we can get the "C" model settled in and what adjustments may be necessary before we move to the "G".

Not sure if there are "G" variants out there calling for bare metal but I'm sure if there are, painters will surface them. :-)

Thanks! Sounds good to me

[gman5250]

The scale of fasteners looks pretty darned good to me in the video. Over-all, I'd say you've knocked it right out of the park.

Thoroughly outstanding.

Thank you sir,

Funny thing, when I was a younger man, I could hit a baseball 400ft...all day long. I still have my Louisville Slugger all these years later.

Painters:
An FYI on the rivets. I reduced the rivet head size by two pixels @ 4K resolution and got the rivet scale seen in the video.
Doing the same thing @ 2K will double the reduction, so our skinners should keep that in mind when setting up their textures.

[gman5250]

I love these rivets, bring them on I say.
And your metal work is not too shabby either, makes me wonder why I even try....

I'm too busy updating projects to do a lot of skinning.

Your art is superlative, and I'm very keen to see what you come up with on your metal skins/alpha/spec !

[jankees]

Let's see if we can get the "C" model settled in and what adjustments may be necessary before we move to the "G".

Not sure if there are "G" variants out there calling for bare metal but I'm sure if there are, painters will surface them. :-)

Sorry Milton, I got distracted by real life and the C-54, but I think the current settings are definitely workable, as Gordon has shown. I still need to revisit my paint, maybe later this evening.
As for bare metal A-20's, the first DB-7 delivered to the French Air Force was bare metal, as were a few early USAAF ones. But I've also found G-models: one nice hack aircraft of the 90th BG with excellent nose art that was on my to-do list, and one post war civil Havoc of which I'm still looking for the registration...both are without guns in the nose though.
Here's a few pics:

[BendyFlyer]

Great work on the skins going on here really impressed.

On another issue (thanks to the dramatic pics of Havoc in action in PNG during WW2), where it is clear to me the very unlucky pilot of that Boston lost it and crashed, I had a look at the documentation (that came with this one) but no mention of VMCA (single engine control speed) there is a safety speed of 160 knots mentioned and I noted the stall speed at high angle of bank can range from 150 through to 170 or 180 (at 60 degrees which is a steep turn indeed). So with a cruise speed of about 230 that is only a margin of about 70 knots or less. My guess is that the safety speed covers the VMCA speed, although in general usage it means a speed to provide for a safe climb or margin above the stall speed. So if you lose an engine in this one your speed is going to degrade very quickly to the danger zone. Not sure how well FSX replicates VMCA going to take this one out and about and find out. Why is this significant? most even experienced flyers do not appreciate the loss of control aspect for a twin or multi-engine aeroplane and how high these speeds can be vis a vis the normal stall speed, it is all about rudder effectiveness to stop yaw and roll after and engine fails, for example the Mosquito had a VMCA of 180 knots but a stall speed of less than 100. A big 4 engined aircraft can be uncontrollable if you lost two on one side or not, depended, the RAAF found out the hard way some years back that a B707 becomes uncontrollable with two out one one side.

Let you know what I find out for those interested. Always fun to explore the flight regimes of these beasties.

[Milton Shupe]

BF, this may help.

My reading says 160MPH or 139 kias.

[wellis]

Here are a couple shots of "Hell 'N Pelican II", 42-86786 as the machine looked (more or less) in 1944 based on the profile in the book "Rampage of the Roarin' 20's" by Clarignbould, Hickey, and Levy. That profile depicts a livery similar to that of the example "Little Joe" located at the National Museum of the United States Air Force in Ohio. Both the nose art and the overall scheme in the 1944 version are quite a bit different than the beautifully restored 42-86786 at Amberley Aviation Heritage Centre near Ipswich, Queensland. Should have this one and the restored version up sometime tomorrow. :-)

[Milton Shupe]

Here are a couple shots of "Hell 'N Pelican II", 42-86786 as the machine looked (more or less) in 1944 based on the profile in the book "Rampage of the Roarin' 20's" by Clarignbould, Hickey, and Levy. That profile depicts a livery similar to that of the example "Little Joe" located at the National Museum of the United States Air Force in Ohio. Both the nose art and the overall scheme in the 1944 version are quite a bit different than the beautifully restored 42-86786 at Amberley Aviation Heritage Centre near Ipswich, Queensland. Should have this one and the restored version up sometime tomorrow. :-)

Ah, beautiful work Sir!