First Screen Shot of MECHA

[kevin]

   While you can optimize your images via colour reduction, if you loose colours and or colour accuracy (24/32bit to 16bit for example), then that invalidates the argument.    Since the resulting image becomes an equivalent of, rather than equal to the original.

[Alex]

in recent times, alot of pixel artists have been raving about Pro Motion, so you might want to check that out. I think it is pretty cheap if i remember correctly. http://www.cosmigo.com/promotion/?Home+%28Pro+Motion%29

[Kman1011]

First I'll just say i had no idea I would start a Grafix debate.  :o

Secondly... MS Paint does have its place. I do have Corel Photo Paint which I used on that second Character because it can blend colours giving smoother images(And a whoe lot more). However Photo Paint can't give me that crucial rgb(0,0,0)( to get transparent background)colour for some reason(or I can't get it to). So i use MS paint for certain reasons. I use
AutoCad for objects that are less complex because I just plain suck at drawing GraFX. As far as odd shapes go (like human and animal forms) I have to use a bitmap editor .

Thanx for the advice Draco.
Here is what I am doing...

1)
I am taking out the shadowing so give some continuity to the tiles, otherwise you get what U see. The light coming from the topleft. This may help some. 

2)
The 'Half opaque' shadow was designed like that because there is overhead lighting which will be taken out the final version. The shadows are on a map level which is just black pixels in a checkerboard pattern.
They will remain in the final version. It's only thing giving it depth. I also plan the darken the background as well.

3) The title bar was a problem. I was going to keep it off to the side even. The thing is, I have to fight for screen space because the character has to see his surroundings. Especially what's below him because he jumps down shafts and he needs to know where he's landing next. The key is showing most of the gamescreen as possible and still show game stats that stand out.

As far as the optimization of images, I have to agree with Kevin. I would only use it as an option in the  game itself for users that have older Grafx cards that are memory poor. The edited screenshot you sent back, I would consider, SUB-PAR.

Although It's nice to know I have options. I'll check out some of these bitmap editors. Thanx ;)

[Draco9898]

First I'll just say i had no idea I would start a Grafix debate.  :o

Secondly... MS Paint does have its place. I do have Corel Photo Paint which I used on that second Character because it can blend colours giving smoother images(And a whoe lot more). However Photo Paint can't give me that crucial rgb(0,0,0)( to get transparent background)colour for some reason(or I can't get it to). So i use MS paint for certain reasons. I use
AutoCad for objects that are less complex because I just plain suck at drawing GraFX. As far as odd shapes go (like human and animal forms) I have to use a bitmap editor .

You can choose to use any color in the top-left corner of your bitmaps for transparency, ya know?
SetImageTranspaceyColour or something of the like.

Anyways-
Ok, well I would just rip the title bar out completely, I'm sure users can remember what they're playing unless they're the gold-fish type (and quite frankly, this feels insulting or even on the verge of demeaning inside most games. I can remember the title of the game, ya know...I would ONLY suggest this be done inside games for young people, but even then...The big idea here is your trying to make people feel like they're NOT PLAYING a game and are inside your magical meta-verse. LEt's say you sit down to play starcraft, do you even see the mention of the title of the game outside the title-screen? No.)...Once you've done that, just move various icons to the left bottom corner or something, but make sure they're not completely flush or you create an effect which feels like the entire screen-space is getting blocked-out and claustrophobic-ish...errr...yah, not a real-word, but whatever. If your STILL fighting for screen-space, use the mario64 / banjo-kazooie trick, just hide the icons until something related to them changes (Lost a life, got hit, gained a life, gained HP...W/e, Etc. Welcome to new-age game design, here folks...)...
Basic 2d design suggestions here- Moving on-

I'd also like to touch on the anatomy and shading of your characters, but that's an entire book unto itself...yada yada, I'm not too happy about these on a per-pixel level...Just me probably, I'm anal like this.

And yes, the fact that everything is pre-rendered creates a huge problem in terms of trying for an optimized pallete, I apologize :), this is just coming from a person who works on a per-pixel level... Just hit me over the head with a pan or something, I'm sure I'd create a comical effect...*Gwwwam!*

Have a nice day.

[Kman1011]

I have to say Draco, you just gave an awsome idea that would work (And MAYBE even make the game faster)

When you said:

use the mario64 / banjo-kazooie trick, just hide the icons until something related to them changes (Lost a life, got hit, gained a life, gained HP...W/e, Etc.

That will work on the bottom where items get selected. I was wondering how I could open that up. Yes i agree with maby taking out the title bar but the life & Air Have to stay as well as the score. Originally there was going to be no title bar. The title bar was a means to get the character status to stand out.

I wanted to use Alfa Images& sprites but that made is WAAAAAAAAAAAAAAAAAAAAAY to SLOW :(

And yes, the fact that everything is pre-rendered creates a huge problem in terms of trying for an optimized pallete, I apologize , this is just coming from a person who works on a per-pixel level.

Pixel by Pixel. Wouldn't that take forever?? I'm not that patient, Although I do touch up's once in a while.

[Draco9898]

I wanted to use Alfa Images& sprites but that made is WAAAAAAAAAAAAAAAAAAAAAY to SLOW :(

Ya, if you could just hang-on until the newer version(s) of playbasic come out, I'm sure you'd be quite pleased using 3dimages/sprites instead of regular FXimages/sprites(free transparency...wowsers...)

Pixel by Pixel. Wouldn't that take forever?? I'm not that patient, Although I do touch up's once in a while.

This is what makes a pixel artist a pixel artist. The fact your willing to devote so much time and love into your graphics. Making them shine like a whistle. Making them pixel-perfect. This is what true graphics are about, IMO. This is true of just about anything really, this is what makes work stand-out, the devil is all in the details, as they say.

[Kman1011]

I've been hearing about the new version 8)

Sounds like it might come with some nice bells & whitsles...SAYYYY....3D stuff?? :)

[kevin]

    Sigh, You can do translucent overlays already.  All you need is two maps (video + fx version),  two cameras and sprites set to use FX use FX  images.

    The screen is split into two cameras. The gfx items are batched to the scene, then we draw camera 1. Camera 1 is attached to the video frame buffer (as those draws are a direct blit).   For camera 2, we recapture items (this time we draw the FX maps) and draw cam #2 . Since the camera #2 is drawing to an fx buffer, we can then overlay, blend whatever you want on top of it.  Then draw it to the screen.

    There are various examples that show this very approach.    Ie.  Rubgy   Which quite happily runs on my old duron 800 machine. 

[Kman1011]

I did a comparison between .PNG and .Bmp and found no difference in quality at all but found a significant difference in size. The .PNG was like half the size of the .bmp file. Obviously the clear choice is .PNG when using GFX in your games.


Quoting myself:

However Photo Paint can't give me that crucial rgb(0,0,0)( to get transparent background)colour for some reason(or I can't get it to).

Really now that I think about it, What I meant was, Corel Photo-Paint does not give a solid colour(meaning All then pixels are not neccessarily the same color). It's very inconsistant. MS paint gives me the basic color model to work with and a fill will guarantee all the pixel colors will be the same. CPP tends to blend for some reason.

Anyway... Mecha is being redone GFXwise and functionwise. Looking at a late summer demo vers.

Kevin...When I get time I'll try the method you suggested above. Sounds easy

[kevin]

   PNG and BMP both include 32bit support.  So they shouldn't appear any different.  You'd only get colour loss if you'd remapped the PNGS down to  16bit or 8bit palette even.  Not that those formats can't represent the images correctly, just that colour loss becomes more likely.

   One thing that is interesting though, is that by pre-compressing your images this can actually be less efficient ( in terms of overall file size),  when the files are packed together for release.   

[Ian Price]

Kevin is indeed correct.

Using one image - saved as .PNG and 24bit .BMP I RAR'd them and got some interesting results -

Size before .RAR
.PNG = 2.95Kb
.BMP = 104kb

Size after .RAR
.PNG = 3.02kb
.BMP = 1.32kb

This was just one image and the .PNG file actually increased in size. I then tested the difference between files in the game I'm currently working on - two identical versions, one with .PNG images, one with .BMP images.

Total file size before .RAR
.PNG = 292kb
.BMP = 1.73Mb

Total file sizes after .RAR
.PNG - 192kb
.BMP - 177kb

An interesting comparison :)

[Kman1011]

I should listen my wife. She always says, size matters. :D

Seriously, its' a interesting comparison and compression seems to make the file larger which is strange. When it comes to storing images I would think .PNG would comeout on top. Storing as an FX image wouldn't take up more room would it?

[kevin]

I should listen my wife. She always says, size matters.

You're a  brave man ! :)

Seriously, its' a interesting comparison and compression seems to make the file larger which is strange. When it comes to storing images I would think .PNG would comeout on top.

  Depends upon the compression method being employed.   Therefore, if the image packer produces a compressed stream that doesn't have many repeatable sequences in it, this can make recompressing the image less efficient.   As the second pass compressor will be unable to find as many repeatable strings,  so it has to store each series of bytes as they are.   So unpacked sequences in the output becomes larger.  Since it includes original packed data (the PNG) with secondary packers depack tokens interleaved into the repack data.   So it grows.

Storing as an FX image wouldn't take up more room would it?

That's a different discussion really.   FX refers to the 'in memory' foot print.  There's no difference in size between Video Image and FX image, it's just where they are being stored.   

    In your case, if you have lots of gfx data (and i assume you do)  you might want to give users the option of the using either Video or FX images.   While purely video will be faster for most players,  this assumes the user has enough free video memory in their video card to load them.   If they don't, it'll fall over.   

   If you have two copies of the media you're doubling the over head, but splitting it between video and system memory.  ie. half and half.   There are few alternative approaches also,  like storing everything as FX image then using some type of video cache or  multi pass sequence to build the screen.  Which use less memory (than storing 2 copies) but isn't as straight forward  (or obvious) to set up.  Nothing ever is :)