Messages

Aye, that one with the torches will be a bugger.  You might be better making it as a cutaway room, and having the corridor run from the left of the screen to the right.  Just build a set with one wall missing, like they would on a sitcom with an audience.  For most things, imagine a very narrow horizontal strip of land, and design it so that almost all the action takes place there, to limit the amount of character scaling involved.  (Eg, a pavement/sidewalk across the street, a train station platform viewed from another platform, a ledge on a skyscraper from the building opposite, the roof of a train, etc).  Then any vertical movement will be mostly used for transitioning to another screen, not giving the player time to notice the drop in overall pixels as it animates away from the camera.

Just use CGTextures..

A lot a people tend to avoid CGTextures now, CGCookie would put them in their tutorials now, and Blend Swap don't allow their textures.  They apparently have a contradictory license that's subject to change.

EDIT:

Perhaps the grass texture should be scaled down and repeated more often - with no other reference points, it currently makes the house look tiny.  The rock looks more like a pebble, too.  Maybe shaper geometry (with bevels, maybe) and more facets.  All in all, it's improving each image.

I've taken most of the advice given except using better textures, My lack of 2d prowess prevents me from creating my own but if anyone knows a good resource for free non licensed textures or any other information on getting better textures please let me know.
[/URL]

http://gametextures.com/index.php/free-category.html  (check their licensing, as you may have to give credit but there's quite a lot of free ones there).

http://search.creativecommons.org/ use this to find free images, and then GIMP (or PS, etc) to make seamless textures from them.  There's plenty of tutorials, but the basis just involves using offset, a clone brush and a few minutes.


EDIT:
Oh, here, too:
http://opengameart.org/textures/all

Othographic, that's the one.  (Numpad 5 in Blender)

Anti-aliasing is essentially blurring the edges to make the lines less jagged/pixellated.  Blender's default is 8 samples, Mitchell-Netrevalli, pixel width size of 1.  Catmull-Rom's an option there, too.

I wouldn't say I've got a great deal of knowledge.  I'd planned years ago to make backgrounds in Blender, which I would then use to paintover (I wanted background buildings/landmarks/etc to be in the right places).  I couldn't figure out what the heck I was doing at the time, so gave up quickly.  Nowadays, I'm using full 3d (not an adventure, not AGS), but back then I used to have a nightmare trying to get a perspective I liked.  I could draw with perspective, but it never seemed to match up with the characters I was making.  The best I can suggest, would be to use a relatively level but low height for the camera (1.5-ish Blender units, assuming you're building something to scale - BU is 1 meter).  For lighting, I'm only really using emission/diffuse at the moment (whilst experimenting with AO) so I'm not sure what to suggest, other than keeping it very similar thru-out.  Lots of different angles will need the models to be rendered out multiple times to account for it.  If it's daytime, perhaps a fixed angle sun with AO to soften up all the shadows?

I'm not completely sure you understand the term isometric.  I vaguely remember someone berating me for calling it 3D Isometric, when it's nothing to do with being 3d; 3/4 perspective or 2.5D are more accepted terms.  Moreover, most point and click games don't even use it, whereas games like the original Fallout did.  Either way, position the camera at a diagonal corner, with it's tilt at an angle that makes everything on screen 120 degrees, and then turn off the perspective view.  Failing that, the vertical lines should remain vertical, whilst X and Y will be equal angles.  (Having the camera at say -10,-10,10 should work, whilst pointed at 0,0,0).

To be honest, I don't think it's too bad of a choice.  It has a cel-shaded animation on a oil painting background that Disney and the like were using in the 30's and 40's.  The down side, however is that a few of the objects in the background look iffy.  The pillow, especially.  The highlights on it make it look like it's been laminated.  The plant (?) on the right could use some tweaking.  I'd also consider making the neon sign animated; a couple of frames with the highlights dialed up or down slightly would work, I think.

You've really captured that "pudgy Indy walking with an invisible cane" feel so many fan games are going for.

That is quite possibly the most hilarious critique I've ever read.

As for the images themselves, can you cobble together a simple line art/block color version for use to see?  Perhaps with a very mild checkerboard pattern on everything.  If we understood exactly how it's working, perhaps we could advise you better.  At the moment, it's like trying to criticise the Mona Lisa based on nothing but a slide puzzle based on it.  We've not really got much as a frame of reference to suggest anything other that "Post 1 is better" or "I prefer post 2".

That said, side view, the far leg seems devoid of shading compared to the first.  Unless the sun shines out of his ass, and every pore, there needs to be some sort of contrast between the distances.  If I look over my glasses (IE, squint), I can't tell which leg is which.

For actually animating something, I'd suggest two things. 1. Find a mirror, the bigger the better.  A large shop window will do.  Walk past it a lot, and take note of how you move.  Then go home, draw it on paper.  Rinse, repeat, repeatedly.  If anyone questions you, just tell them you are testing the reflectivity of the window coating or that you're a security specialist trying to see if the camera opposite picks you up in the mirror image, or tell them you're doing art.  Either way, spend time watching yourself prance, there's no need to be embarrassed.  More than likely they will impale themselves trying to draw a stick figure, you're better than them.

Secondly, walk thru your various animations around your house, concentrating on what feels right.  Once you have a rhythm (even in character), pick a specific point and pretend that's a light.  Walk yourself thru it very slowly imagining exactly where you will be lit up.  Think of 3 colors, and apply 1 to a shadow, then 1 to a base and 1 to a highlight.  Think of it as increasing in light each time.  Do this for pretty much everything you look at.  Ever.  Train your self to see things as a mixture of shadows and highlights.  I'm still having bother myself, as I pick the wrong shades, but I know where they should be applied, and that's helped greatly.  I could copy something or alter it and it might work, but understanding how it should work to begin with has helped no end.

Hands, same issue.

His left hand is odd.  It starts to move forward, then disappears.  His right hand teleports from the back to the front.  I'm not sure if that's because they're not mirroring correctly, or you've just skipped it to get an example out.

As a model/shader, I'd say it looks fairly generic and the colors look somewhat muddy.  (Except the eyes, which seem to bright).

I understand that it's just a test render, but I'd suggest putting more work into it, before porting it to AGS.  That way, you'll have a better idea of how it could be used.  The test render might end up putting you off the idea, if it looks rather hinky the first time around.  Dependent on resolution, you might also want to hoof up the line size on them a bit, too.

I can see it being very handy as a template to paintover, though.

EDIT:

Lighting from that angle may be bothersome, too.  Perhaps a combination of Emission, Subsurface Scattering, and er.. Ambient Occlusion might be a better option?

From what I remember from faffing about with BVHs, they had certain points mapped out, and if there's a keyframe that has a different Location/Rotation/Scale from what they were expecting, it goes haywire.  BVH's can save time on getting a -lot- of animations started, but they then need a lot of tweaking.  The others are right, suggesting making your own animations.  You get a lot more control that way.

As for artifacts and sizing, I'd assume you'd need to experiment with rendering (perhaps with various background colors) and then edit the frames by hand to fix errors.  Perhaps use a huge render size to begin with, then cut out the background with PhotoShop/The GIMP/etc, and scale them down with that.  Should get fairly consistent looking anti-aliasing that way.

I think 'Dramatic' shadows might be a bit tough, as there's very little in the way of unusual objects to cast them.  There's also the problem of having a light source near the camera (blinds, for example) looking awful when a non-shadowed character walks about unaffected by the shadows.

Perhaps you'd be better darkening the entire thing, and picking out reflective highlights.  The way the shadows are working at the moment, it looks like a very bright fluorescent bulb is lighting everything (from the left of the screen).  I'd suggest covering the walls and the floor with the same colored shadows you've already started, and use the existing colors as the highlights.  Doors/lockers will need darkening.  If it's night time, perhaps the lights in the classes would be on, and the hallway dark - this would allow for the light from the door-windows to highlight things.

The banner (with bizarre font placement) confuses me, and the poster shouldn't have that much of a shadow if any.

Anyhow, I find it's generally easier to start with a dark scene, adding layers of lights and highlights than it is the other way around.

I ended up trying to position all the normals by hand, but ended up giving up as it wasn't working out properly.

There should be an option somewhere to remap all the selected normals automatically.

Loomix's FaceWorks is another option to consider for turning pictures into models, I found it to be pretty simple when I tried it.

As for 'creating unique characters', I can see it being handy to cobble together a bunch of gifs to use as textures, for passer-by type characters.  However, I think your time would be better spent making simplistic base models, and then making the textures yourself.  Once you have a texture done (at a larger size, with more definition), it would be fairly simple to duplicate and edit the textures, making small changes (eye color, shading to enhance angles, etc).  This way, you're also not limited to the rather limiting style most generic programs offer; they all seem to look like early attempts at CG in kids shows, and have very little character.

For blender specifically (and others in general) I'd suggest checking out David Ward's tutorials on both youtube and cgcookie.  I've found his videos very easy to learn from, and they'll result in a much more unique look than the current style.

EDIT:

A couple of other quick suggestions.

If you're using the model to pre-render out a bunch of 2D sprites, then I'd suggest going to town on the texture sizes. 1024x1024 or 2048x2048 will produce much better results than what a small gif will do.  The eyes have clarity, yet the rest of the model is far too blurred in comparison.  A more defined texture will add a ton more detail, but will only be slower for the original rendering.  Performance of the 2D sprite won't be affected.

Secondly, have a look at 3d.sk (has nudity), as they've got a vast collection of reference shots of front/side of head, and are perfect for projecting on to models to make textures.  The hair is usually tied back for close ups, too, so you can make the hair as an entirely separate and inter-changeable entity, and as such add lots more variety to multiple models with a minimum of effort.

I've not used the program, but from what it sounds like, you're drawing a wide and shallow "W" shaped line for the mouth?  The original gif has a smile with teeth, and the texture has something similar looking.  The model has a closed mouth.  I'd suggest altering the texture to close the mouth, and then reapply that to the model.

As for the usefulness of the model, I've no idea.  What did you want to use it for?

2. Bets.

I've thrown up on most of the colored lines there, too.

EDIT:
That YHA place seems quite suitable, though.  Unless people have other plans set in stone, I'm certainly not opposed to that place.

Now I just need to know about bookings.  Should I wait for someone to book something?  Should I book a single place/ticket now and try to change to a dorm later?  What should I do about the crowd funding thing?  I'd at least want the free tea and coffee with the VIP card, though I'd prefer them in separate mugs if possible.  If £20 gets me a VIP card, and £25 gets me a VIP card and an advert, I'm totally tempted to randomly pimp myself.  Are there any options for a singular VIP card and more pimpage?  The current options seem to go up in VIP cards per donation, and I really doubt I could afford that many strippers on the day..

You'd let me on a bus?  Even sober's a bad idea.  Disco-pants can attest to that.

keep that in mind if you plan on drinking

I drunk-walked home across London, to the outskirts, from the wrong side of the river.  I woke up in a park.  30 minutes is fairly decent.

EDIT:
I'd happily share a room if others are wanting this place at a cheaper rate and are willing to put up with snoring.  (It's loud enough that you'll hear it regardless of room).

I vaguely know the University, but I don't think there was a central London campus when I'd checked.  £100's a fairly decent price, so put me down as a definite maybe.  I'll look into hotels and travels shortly.  Anything left over, budget-wise, I'll try and throw at the event itself.  250 business cards saying something like "Professional Do-Nothing" or "Freelance Asshole" is too good an opportunity to miss.