Flight models, theory & decisions.

Now that I have again temporarily put the Sopwith Snipe to one side for a few days I am shifting focus. With my ambition to create payware/commercial aircraft one cannot use the Flight Model Data Workbook or Airwrench for any commercial projects as their user agreements preclude such. (One can still use these for freeware releases and I still intend to produce a few along the way.)

I will use other tools such as Aircraft Airfile Manager and MS Excel using my own formulae. Of course something as comprehensive as the FMDWB is beyond my capabilities at this time. However I am collating information and equations from a variety of sources, many on-line, and over time this should evolve into my own workbook. I will of course, be concentrating on aircraft of the 1916-1938 period though I have two 40's/50's jets earmarked for possible commercial release much later on.

I am tempted to take time out to focus on the theory side of things but I still hope to create many new meshes. Return to my usual vocation looks likely too and this may help with the theory side of things. I'll have plenty of time to read, commuting to and from work, between the island & Brisbane.

Aircraft will be ready when they are ready. ;)

cheers

Rob

Information sought Sopwith Snipe

You know the saying 'so near... yet so far....'. That's the way the Snipe is shaping out.

I have encountered difficulty obtaining information related to cockpit details and the fuel system.

I have good detail from a variety of sources of the left side of the cockpit and main instrument panel however little of the right hand side of the cockpit. I have managed to fill in many gaps but some details remain elusive. If anyone has any decent photographs of the right hand internal details of the cockpit I'd be very pleased to hear from them.

The other area I seek assistance is clarification of the fuel system. I have posted a question on this at the Aerodrome but have yet to receive a response. I'll leave things be for a few days and see how I go.

I discovered some possible sources of information on the weekend. Expensive (on my budget anyway LOL) and possibly not obtainable from Australia given the vendors' purchase options. Also possible overkill for 3D model reference material.

http://www.ww1aerojournal.org/mml1.html
http://www.replicraft.us.fm/

I am considering a number of options. We're not beaten yet. Though the Snipe may be shelved and be completed as payware at a later date. We'll see...... :)

Rob

Brief Update

From the virtual aircraft creation perspective, most of this week has been spent working on Sopwith Snipe related issues. I have been working on uv mapping the virtual cockpit.

Now I have paused to research a few cockpit issues and figure out how I'm going to handle cockpit instrument lighting. This is a little bit different from normal FS aircraft and I also use 3D parts rather than the FS gauge system. I also have to figure out how the fuel system works. I am also still researching and experimenting with the flight model.

Little work has been carried out on the Sopwith Triplane this week. I still intend to pursue that as a freeware project and freeware projects have low priority compared to projects ear marked as future commercial/payware projects (apart from the Snipe). I have also worked a little bit this week on the Me P.1112 for CFS3, but again this has low priority.

This weekend some more work on the Snipe and I'll start refreshing the Dr.I's mesh.

Rob

Biff Download Stats as at 9 March 2007

I was very surprised when I checked and talied the Biff's download stats the other day.

As at March 9 2007 - 30,040 downloads of all versions over 3 platforms, FSX, FS2004 & CFS3. These are from Avsim, Simviation, Sim-Outhouse & Flightsim.com

Total FSX (all packages) - 17,103 (From 2 Feb 2007)
Total FS9 (all packages) - 6,329 (From 18 Feb 2007)
Total CFS3 - 6,608 (From 30 March 2006)

It'll be very interesting to see how the FSX version ends up a few months from now.

All in all quite pleasing.

Rob

Slight Change of Plans

Now that I am back to aircraft creation after an enforced week's absence due to a Windows crash and other software related problems, I have again had to change priorities and focus on the Sopwith Snipe. It's fortunate that I managed to get the situation with Max 8.0 sorted out because if I had not, the Snipe and all my current WIP would have to have been shelved until MS update their SDK to make use of Max 9.0. The Snipe would have been an eventual payware release if it did have to be shelved.

Anyway Max 8.0 appears to be running OK. I'll just have to continue work on the Triplane as and when I can. I had originally intended resume work on the Snipe this coming Sunday anyway, so I might as well start now.

I enjoyed the change in activities during the week and revisited some of my earlier artistic pursuits from years gone by. When I'll find the time to pursue them properly I don't know?

cheers

Rob

Back in the design game

Now back up and the running with Max 8.0, many thanks to my Autodesk Reseller. :)

Future Projects .... Decisions....decisions LOL.

I've made some tentative plans as the best way to move forward in the coming months.

Firstly I'll stick to WW1 aircraft for now. I have some interwar aircraft planned but will leave then for the time being as with the Defiant & Luft '46 aircraft. I'm still after decent reference material on the Defiant BTW.

I will also do one other freeware aircraft in addition to the Snipe, mainly for the experience and feedback. It has been an agonising (lol) two days trying to make up my mind and reviewing dozens of possibilities. I have decided to move forward with a Sopwith Triplane which will be freeware. The reasons being that I have heaps of reference material for the Tripe; I have already started one, way back in my OFF days LOL (just the cowl and front fuselage); and I can re-use many parts from the Snipe in the Tripe too (instruments & guns). The triplane has replaced the Sopwith Baby as my final freeware project before moving on to other things. I've had difficulty obtaining good reference material of the Baby interior to date and it will probably cost a few dollars to obtain such. The Sopwith Triplane is planned for FS X only, at this date.

I have tentatively planned my initial payware releases...... some time this year.... can't say when. I can say they will be 2 well known WW1 fighter aircraft, one British and one German. I'm not sure if should name them at this stage....... or should I? ;) I think it will be safe to name these even if I have competition. However I'll wait and see if news of this percolates around the web. These will be followed by a couple of late WW1 German fighter fighter aircraft that were just being introduced when the war ended... not Fokkers (though a late war Fokker will also follow - hint rotary powered. Also another famous Fokker too.). A French aircraft has also been well researched and I am more than happy with the mesh, and will follow these late war Germans. A second French aircraft, a two seater, is also in the works and I will be laying down the mesh shortly. Going into 2008, the Bristol Fighter will be revisited with a Bristol Fighter pro pack. These are all tentative plans subject to change. Much depends on the demands of real life.

I hope to tackle some of the less well known aircraft of WW1 also but much depends on how all of the above goes.

FS X is the planned platform for these at this stage. I still hope to finish the Snipe for CFS3, that will most likely final CFS3 release if I do.

At this stage I do not plan to build 2D instrument panels for FSX aircraft. They will be virtual cockpit only featuring 3D gauges (i.e.) smooth, as in the FSX Biff.

I am starting the Sopwith Triplane immediately and bring that up to the stage I am at with the Snipe i.e. mapping the VC then switch back to the Snipe, alternating between the two.

This should keep me quiet for a while ;)

cheers

Rob

FS9 Biff - 2 to 4 Blade Prop

A recent post at Sim-outhouse asks the question can one change the texture of a 2 blade prop version of the FS2004 Bristol fighter to a 4 blade prop version. Yes one can, but it's a little bit more complicated than just swapping the texture around. The 4 blade version has its own mdl and as the flight dynamics differ slightly from the 2 blade version, exists in a seperate folder.

To add a 4 blade prop version of any variant of the Bristol Fighter, requires that part 4 of the FS2004 Bristol Fighter package be installed. (f2bfs9p4.zip)

In this instance the user wants to see a 4 blade version of the Dutch BR401.

1. Copy the texture folder texture.BR401 from ......\Flight Simulator 9\Aircraft\AeroplaneArt_Bristol_F2B_FS9 to .......\Flight Simulator 9\Aircraft\AeroplaneArt_Bristol_F2B_FS9_4BP
2. Copy the texture Prop_brisfit.bmp from one of the other 4 blade aircraft to the ...\AeroplaneArt_Bristol_F2B_FS9_4BP\texture.BR401 and overwrite the existing texture in that folder.
3. Edit the file Aircraft.cfg in ..\AeroplaneArt_Bristol_F2B_FS9_4BP\ and add the following entry after the [fltsim.1] section.
   

[fltsim.2]
title =Bristol F2B BR401 4P
sim = bristol_f2b_4p_fs9
model =
panel =
sound =
texture =BR401
panel_alias =
sound_alias =
kb_checklists =Bristol_F2B_check
kb_reference =Bristol_F2B_ref
atc_id =
atc_id_color =
atc_parking_types =
atc_parking_code =
atc_airline=
atc_flight_number=
ui_manufacturer = Bristol Aeroplane Company Ltd
ui_type = F2B Mk.1 'Biff'
ui_variation = "BR401 - Luchtvaart Adfeling 4 blade Prop"
description = "Your own text goes here."
visual_damage = 0

The above entry will point to the correct visual model and textures.

On successful completion of the above you should then see the 4 blade version in the aircraft selection menu and in sim.



PS: It's also possible to change other aircraft using similar methods. e.g. D8084 was later repainted and I think had a 2 blade prop. So that can be easily changed using these methods (A restored Bristol F2B, now with Peter Jackson in New Zealand, was completed in the later markings of D8084)

A quick note on Bristol F2B props (some of this mentioned in my SOH reply). Initially Bristol F2B's could have left the factory with either a 2 or 4 blade prop. They could additionally be changed in the field. Tests were carried out by Bristol to see if there was much difference between the two. It was revealed that the 2 blade version performed better and was more manouverable. However the 2 blade version was prone to overheating. The RFC couldn't make up their minds either and instructions were issued at one stage to standardise to 2 blade in the field, however many 4 blade props remained in service during WW1. After WW1, 2 blade props appeared to be standard on the Bristol.

Polycounts Revisited

I notice a new forum at Military Meshes devoted to low poly modeling. The topic of polycount limits comes up frequently at design sites such as FsDeveloper.com and Freeflightdesign.com forums. So what are the polycount limits when it comes to creating virtual aircraft for the MSFS & CFS series?

Firstly I would urge the reader to purchase a good book on creating content for computer games. My bible of the last 2 years has been 'Game Art: Creation, Direction and Careers' by Riccard Linde. This book put me right in a number of areas and dispelled a few myths about model creation that I had encountered across a few web forums. Additionaly I would also recommend 'Creating Art of the Game' by Matthew Omernick. Perhaps not as comprehensive as Linde's book but useful and enjoyable none the less.

What are the technical limitations of creating aircraft for MS simulators?

For CFS3 I think there is a limit of 64k indices per object in gmax. An aircraft model scene can contain many objects. It's unlikely that a well designed model would ever reach such limits. Models of 200,000k triangles are technically possible but not practical as one would encounter graphic card memory and bandwidth limitations. Counts of 70,000+ triangles are feasable and possible in the top level LOD and the aircraft will remain functional in the context of a combat sim such as CFS3, provided, and this is important...., provided the remaining LOD's are modeled in accordance with CFS3 SDK guidelines and examples. More sensible counts are in the 30,000-50,000 region and ideally in the 20,000-30,000 region. It's useful to be aware of these limits. A reasonable external model could be modelled using less than 30,000 triangles, but details such as engines, wires, rigging, struts, guns, crew, fixed undercarriage, floats, and exterior cockpit details can quickly exceed the triangle count of the main airframe if not careful.

Additionaly in CFS3 there is only a slight difference performance wise between an aircraft with a polycount of 5,000 and that with a count exceeding 60,000. The lighter model peaks at higher fps in CFS3 however there is only an approximate difference of about 3-4fps when it comes to average fps in the external view. When it comes to virtual cockpits I have found that triangle counts of about 80,000 are feasable, though counts much lower than that are preferable. Part count, animations and textures are important in determining performance in the virtual cockpit. The triangle count in the CFS3 Bristol Fighter was higher than I would have liked but that was due to it being a two man cockpit. Feedback on the perf of the CFS3 Bristol Fighter has been positive, with comments that it is light on the frames.

FS 2004 is the most limiting flight sim to model for. The makemdl compiler from the FS9 SDK has a limit of 64k vertices per module. i.e. 64k in the interior model and I would guess the same in the exterior LOD 400 module. The FS9 compiler also optimises and welds vertices closer than 4mm, though there are possible solutions, involving more work. I find that FS9 does a poor job on small round parts regardless of the resolution.

FSX on the other hand is absolutely beautiful to model for. At this time I know of no technical limitations of the export tools with vertex counts. I would say that modeling practices used for CFS3 could easily be applied to FSX with the added benefit of FSX materials system. With the overall perf issues in FSX one should always be cautious about going too far.

Feedback on perf of the FSX Biff has been positive... so I appear to be on the right track. After all it's all about pleasing the users.

These days with modern graphics cards polycount is not as critical as it used to be. Factors such as design of the mesh, textures, uv maps, part counts and materials all come in consideration. Indeed I would view part count and texture bandwidth as being more critical than polycount.

In a future article I hope to touch on some mesh creation techniques useful in the creation of virtual aircraft.

In the meantime I'm off to have a break for a couple of weeks.

cheers

Rob

Footnote: While there is considerable lattitude as to polycount in the current FS Sims.... always design for the smoothest and most detailed mesh, however aim for the lowest poly count possible to acheive that result. - Rob

Download Stats

This morning I decided to check the download stats on the various versions of the Bristol F2B and received a few surprises.

FS X version - first released on 2 Feb 2007 - Total downloads of Part 1 - 5,108
FS 2004 version - 1st released on 18 Feb 2007- Total downloads of Part 1 - 1,835
CFS3 Version - First released 30 March 2006 - Total downloads (11 mths)- 6,103

I was really surprised at the CFS3 version numbers. However the CFS3 version stats are for 11 months. It will be interesting to see how the FS versions fare after a similar period of time. Almost half the downloads of the CFS3 version are from Simviation with Sim-outhouse a distant second. Also the best month for downloads of the CFS3 version at Simviation is the current month. Thus the FS versions have brought renewed interest in the CFS3 version.

I am really pleased that the FSX version has exceeded 5,000 in less than 3 weeks though downloads are definitely slowing down, and in the case of Avsim to a trickle.

These figures are very useful to me as an indicator of interest in WW1 Aviation and also show that interest in CFS3 is certainly not dead and will help plan future projects. It would be difficult to translate the above to payware success, as many variables would be involved and probably less than 10% of those who download a freeware aircraft would pay for it. In the case of CFS3 probably very few indeed. Of course a modern 'tubeliner' would attract 10 times the downloads but would require a significant team to produce.

Food for thought :)

Rob