i like the mob torch

Mmmmm, yea. He's a huge one all right. I might try to see what i can do with him sometime--but probably not soon. I've got enough to do .

Cheers,
Unislash

Apologies for off topic in my own thread, but this made me laugh (although not really worth starting a new thread for).  This is a screen shot from "Smith's Basement" today - apart from cropping image, totally unmodified!  Cleo, the Kobold Shaman ...

Well, I've spent some more time on these mobs and come to some conclusions.

(1) I can't really draw (I knew that already though)

(2) Although the figures I painted look good, I can't get away from the fact that when they are photographed and then shrunk down to 30ish pixels square they just don't look so good.  Shrinking to 50 pixels makes them look better but then they are really a bit too big and also they don't really fit with the "style" of Daimonin.  I put the mob on a map and he really looked quite out of place

(3) Animating them is a pain (see also point 1!)

(4) I'm stuck to my collection of figures - no room for other graphics.

So ... I decided to look further at Blender (prompted by Anchakor's post HERE).

From earlier comments made about the Stoneglow guard not being fully rotated / animated I decided to have a go at creating a guard based on the existing guard1.131.png image:



Now, the model I created in Blender was my first one (after having done one of the tuturials) so its a bit rough round the edges ... however, I was able to get the following from it:



Obviously, he is still a little bit too big by a few pixels, and his pose is not very exciting, but those can be quickly adjusted.

I'm not sure though - he still doesn't have the Daimonin "style" - the real guard image is much "crisper" in my opinion.

But what do you all think?

Well (1) applies even more to me, so I won't criticise the actual image.

But otherwise, woot! Very nice first go. Was the rotation done with Anchak's script or was that just inspiration?

Our original art was made by hand, pixel by pixel, with a limited palette. Most started as a single frame for games many years ago. The crispness you see is more an effect the coarseness of the hand selected colors. A modern rendering engine is capable of smoothing the shading far greater than an artist can do pixel by pixel. I like the rendered look, but if you want to break up that smoothness you could add a little detail to the armor or change the texture of its surface.

As to size, he fits on a floor tile, the trick is to see when he walks through different doorways does his sword or shield look like it is damaging the door frame.  

I definitely like this rendered look for the guards.

Good Job!

Looks pretty good. I'd just take tehill's advice and add some detail to the armor... maybe give him a chest, you know

Also, the bottom of the shield seems to be curved in too far, imo.

Cheers,
Unislash

@Smacky - Well, obviously the model itself is a full 3D model ... what Anch's script does is to position the camera around the model at each of the 8 compass points in turn and then spit out a rendering (with all the render settings and image names nicely done so they should fit almost straight into Daimonin.  So really, there are NO rotations to do ...

@tehill, Uni - Thanks for the feedback.  Looks like I might be going in the right direction - I'm encouraged to do some more tweaking to improve the "shape" of the model (chest, shield, feet are too narrow, etc.).

Anyway, it occurred to me I can still use GIMP on the resultant images ... so I applied the "Sharpen" filter to each image:

....

New ........... Original



Is this better or worse do you think?

Original is better; the new one has icky blue stuff.

Cheers,
Unislash

Hmm ... that's strange - didn't notice that.  That is something to do with the animated Gif ... the original png frames don't show those strange blue spots (or the black shield) ... I'll try to recreate the animated Gif later so you can do a proper comparison.

Right - I'm curious about something else now ...

All the documentation for creating mob images talks about sets of 3 animation frames for each action - 3 for resting, 3 for walking, 3 for attacking.  I assume that 3 is the recommended number of frames based around the amount of effort required to create all those individual images and nothing else?

I ask because with the Blender approach, once the model has been created and rigged (Blender term for applying an internal skeleton), it is only a 1 minute job to then "pose" the model exactly as desired.

With not too much more work, I could take my Guard model above, and create 5 each for resting, walking and attacking and then Anch's script would chuck out the 120 still frames with no further work required.

So, are there any server capacity issues (or similar) with doing this?  Anything that means 3 frames is the max?

Just curious ...

The server code uses a fixed number of facings (9 or 25) for each of rest, walk, attack. That makes it 1 or 3 frames in 8 directions plus a dummy frame.

Fine - 3 frames max then ... makes it simple!

So, not much time to do stuff today, but I managed to fix the icky blue stuff ... so here is the original and Gimp'd versions.  Let me know which you prefer, and I'll then spend some time fixing the model itself.

Original:


Latest:


(I think looking at them again, I prefer the original - but its up to you lot!)

Eh, it's a toss up for me. But, as far as the animation frames bit, i'd not go over 4 animation frames per dir/stance. It's the arching that gets hard when you up that number...

Cheers,
Unislash

images have the filename suffix XYZ where X is stance or state, Y is direction, Z is frame. Each can be a digit from 1-9, but see below.

The facings is not the number of frames of animation, but the number of stances * number of directions + 1.

There are 8 directions, so the number of stances is fixed at 1 for non-mobs (1*8+1=9) or 3 for mobs (3*8+1=25): idle, walk, attack. Stance 0 is a special case (I think it is only used for resting players ATM) so ignore that.

Frames can technically be up to 36 in number. That is, 0-9a-z. But numbering should start at 1, and letters really shouldn't be used, so that brings us down to 9 frames. On top of that more frames = more memory, so the 'standard' is 3. For example, most fully animated mobs have 3 frames, but some (ice and lava golems, bears) have 6. But stick to 3.

So stances (3) * directions (8) * frames (3) = 72 images per mob. (Players also have 2 (should be 3) resting images so players = 74 (75) images).