WIP: Hit and Run

[mouseman]

My New Years resolution is to work at least 1 hour a day 4 days a week on animation. I hope to chronicle the progress here, as well as solicit feedback, etc.

 

To start off, I have a rough list of tasks:

1. Management
   • Project startup
     
   • Managing others
     
   • Reporting bugs to Hash
     

[*]Storyboard

• First revision
  
• Review
  
• Second revision
  

[*] Storyboard animatic

• Scan storyboard drawings
  
• Record scratch soundtrack
  
• Edit drawings and sound
  

[*]Concept Drawings

• Character 1
  
• Character 2
  
• Character 3
  
• Set 1
  
• Set 2
  
• Set 3
  

[*]Model1

• Splining
  
• Texturing
  
• Rigging
  

[*]Model2

• Splining
  
• Texturing
  
• Rigging
  

[*]Model3

• Splining
  
• Texturing
  
• Rigging
  

[*]Sets

• Set 1
  
• Set 2
  
• Set 3
  

[*]Sound recording

• Char 1 audio
  
• Char 2 audio
  
• Char 3 audio
  
• Sound effects
  
• Music
  

[*]Block Pass

• Character placement
  
• Camera placement
  
• Sound
  

[*]Animation

[*]Lip Sync

[*]Cleanup animation

[*]Rendering

[*]Editing

I have almost finished the first draft of the Storyboard. I'm not sure if I will be able to scan it. Can't get the old scanner working on Windows 7 64-bit.

[robcat2075]

Looks like you've put good thought into it. I'll be eager to watch your progress.

[TheSpleen]

me too

[Darkwing]

Well, you're off to a much better start than I ever have, so that's good Good luck on the project

[detbear]

MouseM,

 

Don't give up or let hitches/ hendrances get you down along the way. Work at it. I found that while I was producing 'Boomer', I kept going back and writing constant lists..... kinda like you have there. Over and over again I would re-formulate the stages I had done and the stages I had to complete.

 

But I often found that some later stages of production would often get done earlier ....even before what seemed pre-liminary. I would suggest combining your storyboard digitally from simple thumbnails....This skips a massive "Boarding" process that is often unnecessary.

 

Get two birds with one stone. Work out your "Staging, camera angles, and screenplay" all at the same time and use that as a storyboard.

 

William

[mouseman]

Thanks for the replies, everyone!

 

Don't give up or let hitches/ hendrances get you down along the way. Work at it. I found that while I was producing 'Boomer', I kept going back and writing constant lists..... kinda like you have there. Over and over again I would re-formulate the stages I had done and the stages I had to complete.

 

The list is actually the outline of my "schedule", a spreadsheet that I keep of time spent. I did that also on Eggs, Potatoes & Bacon (see signature below), and found it very useful.

 

I have a huge list of tasks, but I hope the final product will be short ... 1 to 2 minutes long. I want something that could be done within 1 year.

 

But I often found that some later stages of production would often get done earlier ....even before what seemed pre-liminary. I would suggest combining your storyboard digitally from simple thumbnails....This skips a massive "Boarding" process that is often unnecessary.

 

Get two birds with one stone. Work out your "Staging, camera angles, and screenplay" all at the same time and use that as a storyboard.

 

Too late! I created some quick thumbnails for storyboards, which was my first task. And they can't be undone! *laugh* I didn't spend too much time on them ... 2 hours. The drawings are just barely above stick-figure level. I am attaching them.

 

Anyways, if anyone has suggestions, I'd like them. If you have a full critique, I'd especially appreciate that. What do I consider a full critique? Well, going through in detail, first with an overview of the whole thing, and then perhaps a paragraph for each scene. Use the Toastmasters "sandwich" evaluation approach; say what worked well or what you liked; what didn't work well and if possible what you would recommend to do differently; and finally sum up with some additional good points.

 

I'm going to change around the end slightly based on some feedback from a friend. Scene 7 Shot 1 will be split to show 1a and 1b. The new 1b will be a closeup of Rick hearing footsteps behind him, slightly worried. For Shot 2, it should be clear that Dave & Bill leave a wide respectful distance from Rick.

 

I'm a little worried about going over my 1-to-2 minute guideline I mentioned above. Plus there are 3 sets (schoolroom, bus stop, and the area where the repeated actions happen). I guess I won't know for sure until I try to make an animatic with them.

 

(BTW, no, I don't remember where I got the template from.) ETA: I found the template!

hitrun_storyboard.zip

[thefreshestever]

keep on going... honestly for me the story wasn´t that catchy when i was reading the storyboard, but in the end it´s all about the realization, often details tell the story... i don´t think you really need scene 7, the story is already told at the end of scene 6. if dave is falling really hard and rick laughs at it, everything is said....

 

about your "to-do-list": you forgot lighting, that will also take a while

[mouseman]

keep on going... honestly for me the story wasn´t that catchy when i was reading the storyboard, but in the end it´s all about the realization, often details tell the story...

Yeah, there's always the "it was great in my head" factor. I saw an interview with Ben Affleck, and he said it was very hard to know whether a movie would be good based on the script. There are so many other factors that contribute to a work being a success. It's hard to come up with a 1-2 minute story that is compelling, so most of it has to be in the acting. And the basic storyboard is not at the acting level of detail. (Maybe it should be? I was thinking the next level of storyboarding would have some of that detail.)

 

i don´t think you really need scene 7, the story is already told at the end of scene 6. if dave is falling really hard and rick laughs at it, everything is said....

I see what you're saying. I was hoping to get a strong sense of "things have changed". Maybe extend scene 6 slightly, with a view of Bill in shock as Rick passes, and nervously (and deferentially) saying "hi"? Or Bill in shock, then laughing uncontrollably, and patting Rick on the back as he passes. Maybe showing that Bill just liked seeing conflicts, and didn't necessarily have a friendship with Dave? What do you think?

 

I think there is some similarity between this story and Warner Brothers' Tree for Two which has a role reversal at the end. (Sorry for the US-centric 6 decade old reference.) Although I'm not doing a pure role reversal, just breaking the role between the antagonist and the protagonist; they are equals.

 

about your "to-do-list": you forgot lighting, that will also take a while

Yes, indeed! Good addition!

[thefreshestever]

yeah, i like the idea of showing bill as a kid that just like conflicts, no matter what... that would give the story a nice touch...

[mouseman]

I started working on a more detailed storyboard. I'm worried at the same time that I'm going in both too much and too little detail.

 

One thing I've thought of is that I should write down the PURPOSE of each scene. I think this is a great idea, and will help make the story a more concrete and, well, a little more purposeful! For example, the first scene, I've written down "Quickly introduce to the main character. Establish he is a student, pays attention even when he's bored, and looks slightly geeky."

 

The other thing I've been thinking about is writing internal dialogue. While watching some really good movies, I've notice there are many scenes where I could write dialogue for everything the character is thinking, even though they are not saying anything during those times. I don't know whether these choices come from the actor, the director, or script writer. But I'm thinking that maybe some of that should happen up front at script (or storyboard) time.

 

I'm a little worried that I'm over-planning for a short.

[Rodney]

I'm a little worried that I'm over-planning for a short.

 

That would take a lot of doing. I think you are on track.

I believe that 80% of your budgeted time spent on a project should be spent in the planning.

Of course during that planning you are looking for ways to keep the other 20% of the production on budget, on schedule and fully focused on executing the plan. (Deviations from the plan are expensive in money and manhours but not as expensive as just wandering around)

 

IMO the more planning the better.

 

 

Note: Nothing says you cannot execute some parts of the plan prior to going into full (final) production if those particulars are sufficiently locked down. Usually these classify as tests of your plan.

[mouseman]

Taking a short break away from the storyboard for a bit.

 

Just getting started on a Type C school bus model. I didn't see any in the contributor's cue, and the bus on the extras DVD was too cartooney.

 

I couldn't find blue-prints or photos suitable for rotoscoping, so I'm just doing it based on specifications and "by eye".

 

Here's some of the research I did:

• Data Sheet for Blue Bird Vision bus
  
• School bus dimensions
  
• 11R22.5 tires ~42 inches in diameter, 11 inches wide
  

Apparently Type A and B are smaller buses, typically using the frame of a smaller vehicle such as a van. Type C uses a truck frame (or custom built), usually with the engine in front. Type D is closer to a transit style bus, with a flat front and usually with the engine in the rear. There is even a School Bus Yellow color.

 

Here's what I have so far.

[mouseman]

A little more done. I have a little more pixel pushing to do on the hood area to get the right shape before doing a CFA. I'm not sure what the best way to do the grille and the headlights would be. Make room for them in the splineage, or do boolean cutters (which I've never done before)? ETA: Or perhaps use decals.

[mouseman]

A little more again. There is one 5-point patch that I can't make into a 5-point patch, unfortunately. I've circled it on the screen capture. Any ideas why? I'll also include the model.

 

I'll be away for a week on vacation without my awesome laptop (which is part of the point of getting away), so I won't have further updates for a while.

School_Bus_Type_C___problem.zip

[itsjustme]

A little more again. There is one 5-point patch that I can't make into a 5-point patch, unfortunately. I've circled it on the screen capture. Any ideas why? I'll also include the model.

 

I'll be away for a week on vacation without my awesome laptop (which is part of the point of getting away), so I won't have further updates for a while.

 

It worked for me...here's an updated model.

 

Hope that helps.

School_Bus_Type_C___problem_update.zip

[mouseman]

It worked for me...here's an updated model.

Hope that helps.

Thanks! I wonder what is different between you and me that you can do it and I can't. I tried selecting individual points, group selecting, lasso, selecting the same points but on different splines, and so on. Is there some configuration? I have lots of other 5-point patches that worked fine. Why not this one ... for me, at least? Could there be a configuration setting, or is it something I'm doing wrong? It's happened to me a few times in the past. This is a pretty basic and fundamental part of modeling; getting stuck like this really kills productivity, and it's not a reasonable long term solution to send out models to the forum every time I have this issue pop up!

 

Thanks again, and thanks in advance for any ideas anyone has of what is going on.

[nino banano]

It sounds a lot of work there, the planning is right for me...are you considering some time to realize this project? ...have a good luck and keep posting the advances,

[NancyGormezano]

It worked for me...here's an updated model.

Hope that helps.

I tried selecting individual points, group selecting, lasso, selecting the same points but on different splines, and so on. Is there some configuration? I have lots of other 5-point patches that worked fine. Why not this one ... for me, at least?

 

I haven't downloaded your model, but usually when I run into 5 pointers "not taking": I select the 5 cps, then hit "." twice (to deselect, reselect). Green Donut should show up. When that doesn't work, I select the 5 cps, hit "hide" so only those cps show, then reselect (with lasso, etc). Donut should show up.

 

However, what I am noticing on your model is a reluctance to use hooks for ending splines. I usually reserve 5 pointers for those places where I want to "extrude" an appendage, ie, change direction of the patches, or the 5 cps are not co-planar. You have situations where the 5 cps appear to be co-planar, and in those situations it might be better to use hooks.

[mouseman]

I haven't downloaded your model, but usually when I run into 5 pointers "not taking": I select the 5 cps, then hit "." twice (to deselect, reselect). Green Donut should show up. When that doesn't work, I select the 5 cps, hit "hide" so only those cps show, then reselect (with lasso, etc). Donut should show up.

Good pointers! In this case, that didn't do the trick, but that is definitely something I will use!

 

However, what I am noticing on your model is a reluctance to use hooks for ending splines. I usually reserve 5 pointers for those places where I want to "extrude" an appendage, ie, change direction of the patches, or the 5 cps are not co-planar. You have situations where the 5 cps appear to be co-planar, and in those situations it might be better to use hooks.

There are places I tried to use hooks, and the geometry went crazy - all over the place. So the next solution (short of creating a three triangles or a triangle and a 4-point patch) was 5-point patch. I think hooks work better where there is a relatively flat or regular area where you are trying to reduce the spline count, i.e. in a cylinder as you get away from the main body and don't need higher spline/patch count for smoother joints (example in TECHREF.pdf pg 54 Hooks). The places that I considered using them on this model were either in a tight area which resulted in the crazy geometry or was on a curve and resulted in too flat a surface where there should have been a curve.

 

Perhaps (or likely) I'm not understanding 100% which places you are meaning. There are 2 5-point patches in the images I posted ... one is the one I mentioned, the other is on the other side of the spline towards the hood area and the windshield.

 

One thing I realized is that the spline that goes from the base of the windshield to down by the mudflap is a single spline. Technically there were not 5 separate splines with those 5 CPs. I broke it, made it a separate spline, then hooked it back together, and I was able to create a 5-point patch (using the hiding trick you mentioned above). The description in the TECHREF "Make Five Point Patch" pg. 266 doesn't specify this limitation, though. That solution works (and maybe will help out in the future), but if you have suggestions for better splineage, I'd love to see/read them.

 

Perhaps I would have been better off with this kind of splineage (see images), that is, if the hook would have attached anywhere near the right place. Or is there something else?

 

I made a 3rd choice that just used more splines. This seems to give a reasonable result after playing with the spline handles a bit, and gets rid of the 5-point patch in question. It's image #3. (I see I have an inconsistency that I will have to fix which is on the other 5-point patch that is in there, but I should be packing now for vacation.) Is this the best solution?

 

I'll probably be skimming the forums while I'm on vacation, but not animating.

[itsjustme]

It worked for me...here's an updated model.

Hope that helps.

Thanks! I wonder what is different between you and me that you can do it and I can't.

 

I'm not sure what the difference would be. I do have a v16 loaded that isn't the current release, but I doubt that's the reason.

[mouseman]

I'm fishing for some advice here. I'm wondering what the best approach is for splining the side of the bus. I have a large area, with somewhat complicated inside pieces (the windows), a complicated bottom due to the wheel well, and then a currently simple top.

 

I have some ideas for dealing with the bottom portion. Are they any good?

 

 

For the top ... I'm at a loss. I'd rather not add a bunch of geometry on the roof, but I'm not sure there is any other way. Another option might be to have the top of the bus and the bottom of the bus be two physically separate pieces of geometry; however that feels yucky, too. Any suggestions?

 

I'm attaching a template if you want to "draw" on the wireframe as I did above.

 

 

Thanks in advance. Any ideas would be appreciated.

[robcat2075]

I'm fishing for some advice here....

 

I'm confused about the "windows", those look like they will be filled in patches to me.

[robcat2075]

have the top of the bus and the bottom of the bus be two physically separate pieces of geometry; however that feels yucky, too.

 

On a real bus they'd be separate pieces anyway, no?

[HomeSlice]

Instead of stitching all your windows into the body of the buss, make a large open rectangle where all the windows go. Then create your window assemblies and just place them into the open rectangle without attaching them to the rest of the body.

 

The roof of the bus probably doesn't have to be attached to the body. It probably has a small rain gutter that runs along the sides. Use that to hide the fact that it is a separate piece of geometry.

bus.zip

[mouseman]

Separate pieces for the win! Thank you both so much! I'm sorry if it sounded like a dumb question, but I'd rather ask a dumb question than spend a lot of time going down a wrong path.

 

With separate pieces, I can avoid the awkward diagonal lines I had been thinking of and go with e.g. the more appropriate and elegant way shown in HomeSlice's example.

 

This will let me keep clean geometry on the roof (low patch count) but still have the detail needed. Getting the right curvature on the roof requires fiddling with the handles, so I'm just as glad to keep that with a minimal set of patches.

[robcat2075]

Getting the right curvature on the roof requires fiddling with the handles, so I'm just as glad to keep that with a minimal set of patches.

 

Make one cross-section that's curved right, then you can just extrude that forward as many or as few times as you want to make most of the rest of the roof.

[mouseman]

Getting the right curvature on the roof requires fiddling with the handles, so I'm just as glad to keep that with a minimal set of patches.

Make one cross-section that's curved right, then you can just extrude that forward as many or as few times as you want to make most of the rest of the roof.

Yes, I think that works well when you have a plan for the model, but less well when you have to go back and stitch in new sections. My problem is that I frequently don't know what I don't know, and have to deal with issues when I come across them. I'm sure that will improve in time, though. It's funny, I feel a bit better at modeling people than mechanical things.

[robcat2075]

Getting the right curvature on the roof requires fiddling with the handles, so I'm just as glad to keep that with a minimal set of patches.

Make one cross-section that's curved right, then you can just extrude that forward as many or as few times as you want to make most of the rest of the roof.

Yes, I think that works well when you have a plan for the model, but less well when you have to go back and stitch in new sections. My problem is that I frequently don't know what I don't know, and have to deal with issues when I come across them. I'm sure that will improve in time, though. It's funny, I feel a bit better at modeling people than mechanical things.

 

Yup, part of spline modeling is "seeing it" before you make it. But backtracking and redoing things occasionally isn't fatal.

[mouseman]

205 patches so far. Making slow but steady progress, thanks to the help from Homeslice and Robcat.

 

I'm working on the steps to figure out what their dimensions should be. That will take a little experimentation and calculation. I've already figured out that the steps were too high to start with. This would all be much simpler if I had an actual bus to take measurements.

[Shelton]

The bus that is used in the bus stop is mine. I found a great deal of reference material online at the bus manufacturers web sites. Some of them had multiple views and some even had drawings of the buses.

 

Steve

[mouseman]

Hi, Shelton,

 

I did a bit more digging based on your message. I didn't find any nice diagrams that could be used as rotoscopes, unfortunately. Maybe transit bus companies are better with that kind of information? Most of the school bus pictures were 3/4 views, or close-ups of particular areas. Most photos are not going to be useful for rotoscopes, anyways, since the bus itself is so long; you'd need a far-away shot with telephoto to get a more orthogonal-like view of it.

 

Here's what I'm mainly using for reference:

 

Blue Bird Vision product page.

Of particular use is the image that shows "Best In Class Visibility" and "Vision's Angled Hood"; I am using that for a side-view rotoscope of the front section. Of note is that it shows the height of the stairs! That also gave me the floor height within the bus (which happens to be just above the rear tire height). I was amazed at how close I was with eyeballing it previously.

 

Full view of the driver's side.

Not useful for a rotoscope, but this shows what the side without the door looks like.

 

Blue Bird Vision data sheet.

I really like the view of the driver's area; that will be a tremendous help. The "Tire Size" info (11R22.5 (G)) led me to accurate measurements for the tires.

 

Rear view with emergency door

From the Wikipedia School Bus page. This particular picture looks like a bus in the Netherlands. Although not exactly what I'm looking for, it's the closest thing to a rotoscopable image of the rear of the bus.

 

Safety fail view of rear emergency door

About 1/2 way down the page. (Notice that the handle is in the down position, so it is probably broken; thus the lock to prevent the door from opening on its own while the bus is in operation instead of getting the door handle fixed.)

 

A 3/4 rear view

From AirportShuttleExpress. I modeled the rear bumpers based on this view.

 

A view of the Stairs

Not useful for measurements, but good reference for what they look like.

 

Edit additional links:

1950s to 1980s

1990s to today

[mouseman]

I almost wish groups could have 2 textures; one for the outward-facing normal, and one for the inward-facing normal. For example, I could specify yellow for the outer surface of the bus, and dingy white as the inner surface. This would require separate lighting for each side of the surface; for example, light hitting the outer surface of the bus should not illuminate the inner surface.

 

This would save having to create an outer shell and an inner shell of the same surface. What do you think, crazy idea?

[Shelton]

excellent way of measurements. I looked back and that some of the diagrams were found on places like blueprints.com and the like. Sorry. What I did find were two manufacturers web site had front side and back shots of the buses. This what I used rather than the diagrams. You can email me if you are still having issues at

 

steve

[Rodney]

I almost wish groups could have 2 textures; one for the outward-facing normal, and one for the inward-facing normal. For example, I could specify yellow for the outer surface of the bus, and dingy white as the inner surface. This would require separate lighting for each side of the surface; for example, light hitting the outer surface of the bus should not illuminate the inner surface.

 

This would save having to create an outer shell and an inner shell of the same surface. What do you think, crazy idea?

 

Put it in as a feature request in A:M Reports.

One never know when an idea will stick.

 

You remind me of a time I thought I had stumbled across a multitexture solution for coloring two sides of the same patch.

I believe Robert Holmen approached this by way of rendering twice and there is a topic for that.

 

My memory said that I did this with materials but when I attempted it again... no deal.

Perhaps it was different colored lights I was shining onto opposite sides of the patch.

 

As a feature, the normal facing solution work well if the textures showed from the bottom and top of the stack.

I'm not sure of the programming challenges there.

[itsjustme]

I almost wish groups could have 2 textures; one for the outward-facing normal, and one for the inward-facing normal. For example, I could specify yellow for the outer surface of the bus, and dingy white as the inner surface. This would require separate lighting for each side of the surface; for example, light hitting the outer surface of the bus should not illuminate the inner surface.

 

This would save having to create an outer shell and an inner shell of the same surface. What do you think, crazy idea?

 

The only problem I see is that then the surface would have no thickness...so it wouldn't look realistic.

[mouseman]

excellent way of measurements. I looked back and that some of the diagrams were found on places like blueprints.com and the like. Sorry. What I did find were two manufacturers web site had front side and back shots of the buses. This what I used rather than the diagrams. You can email me if you are still having issues at

With the reference links I posted, it seems to be working out pretty well, so I think I'm good.

[mouseman]

The only problem I see is that then the surface would have no thickness...so it wouldn't look realistic.

I agree it wouldn't always be of use due to the reason you mention. If you aren't looking edge-on, it probably wouldn't be noticeable. I think in a case like the bus, where for example the windows will all have frames to hide the seams, it wouldn't be noticeable.

[robcat2075]

To model an inside of a shell structure you could use Steffen's Puzh replacement plugin to quickly get a slightly smaller version.

 

It's out there. I don't have the link.

[mouseman]

Current progress: 371 patches, 8 5-point patches. I got the floor, stairs, and a lot of the rear section of the bus.

Windows and seats are coming up. Probably the wheels after that, then flashers, grille and lights, working towards being able to do a copy-flip-attach. After the CFA, I'll do the entrance doors, driver's seat area and window, emergency door handle.

 

ETA: Oh, yeah, and the inner skin for the bus. Hopefully I'll find the Puzh plugin when I go looking.

ETA: Steffen has created the PushCPs plugin which has similar functionality to the Puzh plugin.

[mouseman]

Current Progress: 3382 patches; 8 5-point patches

 

Created windows, seats, and flashers. Expanded width from 7.5' to 8' per spec sheet.

 

 

TODO:

• Interior (use PushCPs plugin)
  
• Wheels (and maybe more under-carriage)
  
• Grille
  
• Headlights
  
• Copy-flip-attach
  
• Entrance doors
  
• Driver's seat area and window
  
• Emergency door windows, handle
  
• Texturing
  

[mouseman]

Current Progress: 6869 patches, 16 5-point patches

 

Created interior. Did the copy-flip-attach (probably a little prematurely).

 

 

TODO:

• Wheels (and maybe more under-carriage)
  
• Grille
  
• Headlights
  
• Entrance doors
  
• Driver's seat area and window
  
• Emergency door windows, handle
  
• Texturing
  
• Rigging
  

[mouseman]

Current Progress: 8188 patches, 16 5-point patches

 

Created wheels. Boy do I wish I didn't have other work to do tonight.

 

 

TODO:

• maybe more under-carriage
  
• Grille
  
• Headlights
  
• Entrance doors
  
• Driver's seat area and window
  
• Emergency door windows, handle
  
• Texturing
  
• Rigging
  

[mouseman]

Finally getting back to doing some modeling this weekend.

 

Current Progress: 8607 patches, 16 5-point patches

 

Worked on rigging (based on the one created by mtpeak et al), but not in a final format yet.

Some work on the driver's side.

Added the driver's seat and a steering wheel; I will probably have to reposition it when I get the rest of the console in there.

 

 

TODO:

• Maybe more under-carriage
  
• Grille
  
• Headlights
  
• Entrance doors
  
• Driver's console and window
  
• Emergency door windows, handle
  
• Texturing
  
• Rigging [ in progress ]
  
• Add edge to the fender
  

[robcat2075]

Good progress. I'd suggest putting an edge on the fender so it's not just a infinitely thin surface.

[mouseman]

Not as much progress lately as I'd like, but some today. Was working on the driver's control panel and dashboard; still a ways to go before it's presentable.

 

Current Progress: 8640 patches, 16 5-point patches

[mouseman]

Registered up for another year last night.

[TheSpleen]

looks great!

[mouseman]

Lesson: If you have transparency, use TGAs for decals, not PNGs. This is with PNG; one (with the white background) is set to "color", the other is set to "transparency" and causes part of the geometry not to display.

I have a sample project that hopefully I will remember to submit this weekend and make sure it's not a duplicate.

BusFrontTitle.zip

[robcat2075]

Different programs seem to interpret and create PNG transparency differently so i avoid using them with A:M.

 

However, when i test your PRJ I don't get any missing geometry. Is this what you intended?...

 

[mouseman]

Different programs seem to interpret and create PNG transparency differently so i avoid using them with A:M.

 

However, when i test your PRJ I don't get any missing geometry. Is this what you intended?...

 

Hmm, interesting. That's a good sign, I guess, although I don't know why it happens on my machine. I have 64-bit Windows 7 with Nvdia GeForce GTX 460M.

 

The other thing is that the upper-left PNG is intended to display the letters in black, and then transparency allows the yellow to display. The lower right is the same thing except it's a TGA, and it works as I hoped it would.

 

I included an HTML file that shows that the background shows through around the PNG, so browsers (at least Firefox) recognize the transparency as I would expect.

 

I would prefer to use PNG for size sake, but at least TGA works.