Wednesday 28 September 2011

Two years old

Yippe..the 18 month project is now 24 months old. Horray. Short post but I felt I couldn't let it slip by unnoticed. Especially with all the Tipex (White-Out) all over my Project Planner.

Thursday 22 September 2011

KBU, FCR mode logic

First screen-shot of a quickly put together KBU (keyboard unit). This is a flat 2D image you click on, taken from the texture used in the cockpit. I wanted to see how good/bad it looked since we don't have time to develop a 3D one.



Clicking on Air Surv will abort current FCR operation
and switch to an active air scan (air surv mode)

The keyboard command list is looking horrible if you want full control over this. The shot cut keys work fine however. I made some tweaks to the heading tape symbols and the Air Surveillance Mode has been adjusted according to recent feedback posted. No sweep indicator is displayed (? is that correct) but the radar is still active. And not NTS or targeting is available. This mode is not available if either crew station has FCR as the selected sight. Either crewman selecting FCR or cycling through HMD/TADS/FCR will cancel this mode.

I made attempts to smooth out the logic for simple joystick control so you can hot key from mode to mode.



FCR Scan and FCR Scan burst will cancel each other out with immediate effect, no waiting for a scan cycle to complete (which is what it did before).

I'll adjust the scale of the FOR box shortly. The dot isn't active yet, will fix that shortly and then I'll correct the cue dots in the HMD,

That will be as much time as I want to spend on the FCR. I think it's not too shabby and does nearly everything it needs to do for now. Completing the command list for the sensors now means I can finish the default control setup for new installations.

For full control on a joystick you need 3 hats minimum.

  • 4 Way Sight Select (HMD, TADS, FCR)
  • 4 Way FCR Mode Select (GTM, ATM, other to come)
  • 4 Way FCR Footprint Set (Zoom, Wide, Medium, Narrow)

There are also keys assigned to FCR Mode Cycle Up/Down to keep the control count down on smaller sticks.

As always if I've got it wrong then please leave some feedback however I'm unlikely to embark on any more major changes for now. Quite pleased with the detail and how it works.

Tuesday 20 September 2011

CH-47D Overhead Panel Complete



Dome light fixture, fist pass on panel illumination

Monday 19 September 2011

Sensor tidy/fixes today...

 Bored with writing XML code and needing to add missing commands to the master command list so I can complete the default/new install player code. Self review day. This is going over my feature list and checking what's implemented against my much reduced feature list. Sometimes there's a gulf between what I remember I did and what's working correctly. Also a good thing to do on a Monday.

I came across a couple of things I missed...
  • FCR now tied to Main Bus Electrical circuit (won't operate unless powered)
  • FCR Burst Mode added
  • Current sweep angles assigned to Avionic.ScanVolume
  • Hot keys for setting scan volume (INS,HOME,DEL,END)
    Wide, Medium, Narrow, Zoom (accurate terms? on stick marked as WMNZ)
  • Changing scan volume resets the active sweep.
  • Commands for FCR and TADS should no longer interfere with each other.
  • PGUP puts commands into TADS operation mode
  • PGDN commands operate FCR

And all FCR/TADS commands are now transmitted to the Avionics class via the message queue. Of which there are many.

*edit*
Air Surveillance Mode logic added

Alternate Sensor Bearing shows where the other crewman's
sensors is pointing on the heading tape.
The Alternate Sensor Bearing (small needle), the Pilots MPDs will update the CPGs active sensor position and vice versa.

Now with added "buttcheeks"
(FCR centreline sized large enough to see)


Air search mode completed.
Finally got the colour correct and track count added top right


New MFD buffer export shows image is offset slightly
*edit*
Missing FOR added, looks a bit better

There's supposed to be a way to rotate or offset the scan volume, in LB2 this rotated the wedge. Currently is locked down the centreline. In burst mode the scan line (opps...drawn in default white in this image), makes a single sweep cycle (left/right). And I also missed the call to the field of regard box at the bottom which I'll fix now.

A chunk of fixes today and more things I can now tick off. Air Search and the offset track should be completed tonight leaving the TADS to go over tomorrow and that will be pretty much all the sensor buttons and functions done.

Lip light

It's early morning and I've not had chance to get much visual work done. Lot of tedious read/writing of XML data but nothing photogenic.

One problem I have testing during the night is doing startups. There are hot-keys for instruments (CTRL I toggles all interior light stations in a flight ready mode) however the lighting panel located by the left hip is near impossible to see.

A SimHQ post pointed me at something unrelated (ground fire acoustic acquisition system)  and something I'd heard mentioned before in passing but didn't twig. The mic mounted lip light.

Lip light - 3 bright green LEDs

Activated by a capacitor switch with the upper lip or tongue on the back of the unit.



Writing this blog entry took longer than the implementation. Here's some pics to illustrate my guesstimation of cone angle and colour (I'm told my first colour was too blue-green so here's a more yellow-green that matches the above photo better). The yellow-green mixes with the existing instrument lighting to produce another green. Still a long way off from making "purest green" (Blackadder ref.).

Where are my keys?



With instrument lights, two shades of green

Lighting panel, left hip

So that's lip lights added. Next.

Saturday 17 September 2011

Update from the CH-47D workshop

I'm busy working flat-out on the Apache and game core, getting it ready for serious playtime. This week I  had to rewrite a couple of small classes all eats into the schedule. It's irritating to do because I want to push this into the next phase which is the testing cycle. Meanwhile thousands of miles away (we're a global outfit by virtue of having seats in different continents - and that's all) the Art Department (AD aka Dave) as been busy working in his virtual hanger/cardboard box.

Always a pleasure to see a new batch up update screens first thing on a morning and this morning contained a nice surprise.

The first batches of switches have been added. I think Dave has redefined the role of technical artist since he also went and did the research. And this is the EASY panel. Anyway, here's a breakdown of the main components.


Looking hot. Making the cockpit database is going to to be too :/
Overhead panel
Centre console from pilot side, collective in foreground

Some of the switchology is going to be interesting to do. The crew station intercom of which there are 3 duplicate panels on this console won't have much use in game but could be assigned a switch to station command. We have the crew-chief/ramp gun, left/right door guns and loadmaster as possible player slots.

AN/ALQ-156
Top left on the centre console. Part of BAEs ASE gear (link)

Below this...

AN/ARC-220
HF Aircraft Communications System. 2.0000 to 29.9999 MHz high-frequency range









TSEC/KY-100
Remote Control Unit

Found a diagram of the display board (link).

I don't know what this is for either, yet.





AN/ARC-164
10 Watt UHF AM radio that operates between 225-399.975 MHz and includes independent guard receiver for monitoring (link).

AN/ARC-201
Frequency-hopping, FM communications transceiver. 30 to 87.975 MHz band (link).


The research Dave subjects himself to is for visual accuracy, labels, layout and switch type.


I remembered to add a "reverse axis" options in the quick-start control settings. I got stung by that one before. There's a realism tab for tweaking settings but some more 'eccentric' options will still require a manual edit.


This weekend I'm going over the setup options fixing save/load issues and another pass on the MFDs checking the pages (I found some that link to nowhere, probably scale not set correctly and everything is getting drawn off the page).

You don't see this often, debug landscape texture for
checking distance and movement

At night the menu background switches to a
FLIR mode effect.
Added a 'lip light' so you can see what you're doing in
a dark cockpit (all shadow details on).


Selling out

A t-shirt store will be going up next week so if you want to show a bit of support you can grab one. They are similar to the ones I had made up last year for floor-shows/events but you'll at least get to pick what colour you want. Coming to a blog near you.

The new face of indie game development,
alternate revenue streams, aka merchandise!
All proceeds will go to helping offset development costs plus you get a piece of material with some printing on it. What's not to like? I wear one, now you can too. Plus an exciting new range of cow-wear.

I still have a Microprose F19 Stealth Fighter t-shirt, pink on black. (Why pink? We'll never know) one of the many freebies they would give away in first edition game releases. Hmm, they lost how much?

Wednesday 14 September 2011

More Franck Shader Magic - Cone Step Mapping

A short update tonight. Franck released his conversion of the Cone Step Mapping shader. Now I hadn't heard of this one until he mentioned it. I'd seen POM or Parallax Occlusion Mapping which at first glance this looks very much like it.

This is a flat object with a texture that casts shadows. It uses a pre-computed map and much faster than POM. This is shadows sans geometry. It's pretty amazing actually. I don't have an immediate use for it in out game but I can think of a few areas I wouldn't mind trying it out. Such as the runway border regions and the quilted Chinook interior. Will do some tests in the coming days.

Screenshot from Franck's demo scene

I've been rounding out more mundane GUI code. The artistic concept was to try and capture a 'systems' feel, like that of the fuel system MPD screen with interconnecting lines showing option paths instead of fuel flow. I'm not convinced it works aesthetically but it does the job. Anyway, if you have a hard time with the options screen then heaven help you in the cockpit. I tried to stay away from a full MPD look which was used in GUNSHIP! and I think Longbow 2. Green didn't look so great, orange is our title colour so it had to be orange and something else which is GUI_BLUE. Note the "special tab", you'll only see that with some hardware attached.




AD has been working real hard on the Chinook control panels. He's finding that the texture map is too small to bring out text details which will probably mean breaking it down into more textures. The Chinook is looking like it will be way more detailed and use more textures than the Apache. It could quite possibly be the single worst helicopter we could have chosen in terms of detail and technical difficulty.


*edit*

Using panel drawings as a template mapped to the cockpit consoles. This provides a guide for placement of labels and switches. A most labour intensive task is taking groups of these controls and baking ambient shadows for every single switch onto the panels.

Digital sets like the one with the keyboard below will require an extended version of the TComputer class so it will have a display buffer and can send and receive messages from other parts of the helicopter. Of course it helps to know what the hell it is first :)



Officers Mess

A demo for Ace Combat Assault Horizon appeared today on Playstation Network. I was crestfallen at how awful the Apache was treated. Lovely visuals, but barrel rolls from a hover to evade missiles? I almost cried. As far as my console gaming habits are concerned, this is lined up to be the biggest disappointment of the year just behind Test Drive Unlimited 2.

And finally, I tweeted this earlier but my local airport has some pretty violent cross-winds. We caught the back end of the Hurricane which hit the US east coast and yesterday I hear the unusual sound of turbines screaming instead of the usual reverser roar. http://www.bbc.co.uk/news/uk-14906539

Great fun. Should have been on my flight when the pilot was doing S bends on landing.

Monday 12 September 2011

Dum de dum. Monday afternoon.

Mid afternoon blues setting in. Coffee break and some pics to post, that should kill five minutes.

TADS symbology buffer now has things to render

"Needs more cowbell" (some of you might get that ref)
You might notice the TADS is no longer stuck in that minty flavoured 'night' mode. Turned out that was a by product of me using the wrong buffer (MFDs all have two buffers, one for a video feed and another for symbology but VIDEO has it's own buffer).  I was using the raw video buffer on the TEDAC. Took me a few tries to figure out why non of the symbology was appearing.

Boeing trainer press/web photo link

The TADS/FLIR camera rendering will be based on the shaders I did for the PNVS, currently it's just an edge-enhanced image, there's no heat or light mapping applied so it doesn't look as FLIR like as it should. Irritatingly because the camera is attached to a helicopter, small forces in the physics body (vibrations) transmit themselves to the TADS image. It shakes like a son-of-bitch at high zoom. I need to add camera stabilisation to the simulation! Go figure. The alternative is to detach the camera from the helicopter chassis and apply a bit of smoothing, which is easier.


Darn, still got coffee

AD asked me not to post one of these because one of the collectives has yet to be moved to the correct position (I'm left handed so I don't have a problem with it *grin*). He's all over the Chinook like brownout in a sandstorm. Anyway, because it looks so darn good and it will kill another 5 minutes while I drink my coffee here are some of his progress screenshots.



In engine real-time captures
Oh, were're building a weird CH-47/D variant apparently. I hope we can build the door gunner stations and loadmaster hatch to this detail.

Finally after many hours of memory issues and AO burning pretty much every surface has now been mapped.

Sunday 11 September 2011

FXAA 3.11

Franck Poulain completed his conversion of Timothy Lottes FAXX shader for Leadwerks Engine 2, this brings it up to the same version that I understand is used in Battlefield 3.0.

I've been looking over the firing range with the new full-screen anti-alias shader to see if there are any visual artefacts or performance problems and so far it seems to work great.

Showing solidarity in flag form on this day
Looking over some vertical structures you can see some interference patterns but these are related to cascaded shadow maps.

Apache avionics are not initialised until required
hence the blank displays. One to fix later.
Edges around the canopy look better. The cockpit interior still does not have post processing effects applied but would probably benefit from FXAA, I think the performance improvement would justify a user graphics option to enable cockpit post effects.



Performance hit seems quite small at these windowed resolutions. It's well optimised as it needs to be for consoles which are a few years behind now.

So thanks to Tim at nVidia for publishing the original work and Franck for making the conversion available to the the whole Leadwerks community.

Now if only we can get some pseudo volumetric smoke and the tree LOD transitions smoothed out some more.

Saturday 10 September 2011

Just in ADs CH47 3D cockpit WIP

AD's been working in the hanger on the CH-47 again and just sent some of the first player cockpit textures, it's a big deal from my point of view as the high detail cockpits are a lot of work....and re-work.

Here's a selection of updates from work done today.









*edit*

But wait there's more....new metal texturing technique, no more photo-textures.