Hi-Resolution Graphics Help Inquiry

[mac53]

Hi Bill...

Pleasure to met you, buddy. Yes, you are correct, the graphics do extract @ 256 color palette 8 bit bitmaps. I use a program called Xwe, for any type of wad editing. Getting these to 24 bit [True Color] @ 300-400 ppi, is a piece of cake. Hell, I have about three or four softwares that will do that for me and one of the came in my Office 2000 package, Microsoft Photo Editor.
This is not what I'm striding for, Bill. One of the first major processes I have to succeed at here, is to convert these graphics from pixels per inch to dots per inch, so I can continue on by softening and smoothing the edges. Trust me when I tell you, this needs to be done! This is the first step I need to accomplish, before the rest of the pain-staking processes begin. If you could provide me any advise as to getting this process completed, then you'll have my complete attention. Then I'll go on asking you to teach me how to use the Adobe PhotoShop Elements 5, I just purchased less that a week ago and have no clue to it's abilities.

I just downloaded PhotoZoom to have a look at it. Christ, I hope I don't have to buy another software! ;-)
----------------------------------------------------------------------------------

Man..!!! I have to edit back in just to say, "What a beautiful job "PhotoZoom" did"!! $149 bucks though! If I would have known this a week ago, Adobe would still be sitting on the shelf back in Staples!

[Soquili]

Pixels per Inch and Dots per Inch are references to two different types of media.

Pixels per Inch is a reference to screen display and Dots per Inch is a reference to printed output. If an image is 200 ppi it will print at 200 dpi on most printers (that can print at that resolution or higher).

For displaying graphics on screen as within Doom II, dots per inch is the same as 200 ppi; no conversion is necessary and even possible.

The program I mentioned resamples the imported image and increases the resolution and maintains the sharpness of the original. Ordinary programs like Photo Editor, Photoshop, Paint Shop Pro, PhotoPaint, etc do not resample so they enlarge the pixels into multiple pixel rectangles. This is called pixelation and PhotoZoom is the best program I know to prevent pixelation when increasing resolution of a raster (bitmapped) image.

The only alternative is to redraw from scratch each image (not a pleasant alternative with 1,381 images) using one of the program you have available.

[Grafixman]

Yes, perhaps a dedicated program for enlarging pictures like PhotoZoom is what you need. I'm not sure about Bill's statement about the other programs' (Photoshop, PSP, PhotoPaint, etc) not resampling when enlarging, though. I thought Bicubic, Bilinear, Nearest Neighbor, which should be standard in most photo editing softwares are resampling algorithms? But from what I gather, there are more advanced resampling methods than these that works better than Bicubic, and PhotoZoom sounds like it's using one of those.
I'm also confused about the importance of converting the ppi to dpi. If you're not going to print them, why should it matter? They're just conventions, and should have no effect on the data contained in the pictures. What should matter is the number of pixels you have in the final high res output, shouldn't it?
Returning to the original question, you said that the original graphics were low resolution 256 colors. By enlarging them, are you striving for accurate enlargement? With the colors and definition intact? Then you'd be getting very large pixellated pictures only, not high definition, which is not present in the original low resolution extracted pictures. I'm not sure what good that would be, since you'd only end up with very large pixel style pics that are going to be resampled again when they are used in the game...
Not sure if there is a simple automated process to convert low res pixel style artwork into high resolution smooth anti-aliased ones... you're asking the software to put data where there is non in the first place. I mean, take for example a tiny sprite. The eyes would probably be only one or two pixels across. How would the software redraw that into a detailed eye in the enlarged version?

[jamesmc]

I think he is confusing color depth 8 bit versus 24 bit with PPI and DPI which deals with image resolution.

Color depth is bits per pixel (bpp.) For instance 1 bit color is black and white. One bit for black and one for white. Which is 1^1. 24 bit is 1^24 which is 24 bits per pixel (16.77 million colors.)

One would want to use higher bpp if they wanted to show more gradation and or subtle differences in hue. This is the reason that 8bit graphics like GIF's cannot show gradient graphics well.

The limiting factor on image resolution and color depth (also called color resolution) is the user's monitor, graphics card and method of delivery... GIF image versus Flash Vector Image.

At least that's my take on the matter and I better quit before I get myself confused, besides I need to eat my breakfast.

[mac53]

Oh yeah... Happy "B" Day, Bill!

Alright guys, take a look at these two images. The first one is a 4x zoom of the original graphic which is in png format @ 400 ppi. The second one is the same pic but ran through PhotoZoom and enlarged to 4x @ 400 dpi. You'll see that in second pic the edges are more rounded than the first one. This is very important if one is trying to do a project such as the one I'm attempting. Gents, this can be done!

[Soquili]

Hi GMan,

Sorry to confuse you, I was trying to keep the terminology level down

Those algorhithms you mention do a sampling, but they mearly extrapolate so they make bigger colour areas from the sampled pixels.

The S-Spline algorithm used in PhotoZoom works differently so it increases the pixels per inch without overly enlarging the colour area represented by a sampled pixel.

With most monitors displaying 96 ppi an increase in an image's ppi will enlarge the image on screen. I'm not sure how the Doom II engine is handling the larger ppi to display it's flats, sprites, etc.

Most monitors produced in the last few years can actually display at higher resolutions and Windows Vista takes advantage of the hardware. You can select 120 ppi display.


Thanks for the birthday greeting Mac53. I was composing my post and missed yours until I hit submit.

Your examples perfectly show what I was explaining about how the older algorithms expand the pixels colour over a larger area. The S-Spline algorithm compensates and does not expand large rectangular areas to repesent a single pixel from the sampled image.

[mac53]

I'm not sure how the Doom II engine is handling the larger ppi to display it's flats, sprites, etc.

The engine requires a scripting file that tells it to reduce the 4x sized textures to their original size. There can be two seperate entries... one is called a "Decorate" and the other, which is much older, is referred to as "Dehacked". Now, without getting too deep into it, [beacuse this is really not the place for it], these entries allow larger and Hi-Res textures to be entered and displayed, not only showing better detail at distances, but increases the the clarity, color and definition while in the close-up contacts.
You boys would have to see this to understand what it is I'm trying to do for the Doom 2 authoring community. It's a seriously tough job to do and I'm looking to crack it.

[Soquili]

Thanks for the info Mac53. I saw references to the Dehacked when I was reading yesterday.

Am I correct in assuming that this Decorate entry also allows you to use png files in place of the bmp files?

If that assumption is correct, do you use 24 bit png files or are you able to use 32 bit (24 bit colour with 8 bit alpha transparency)?