Text Only | Text with Attachments

QB64.org Forum

Active Forums => QB64 Discussion => Topic started by: venex on January 20, 2021, 04:50:38 pm

Title: Newbie question about 8-bit screen in new QB64
Post by: venex on January 20, 2021, 04:50:38 pm
I have a simple program, which works fine with QB64 SDL in Windows 10:
Code: QB64: [Select]
  2. i = _LOADIMAGE("sample.png")
  4.  

The sample.png is an 8-bit 320x200 image.

The code compiles with the latest QB64 without errors, but when I try to run the executable I'm getting the following runtime error:
Quote
Line: 3 (in main module)
Illegal function call
Continue?

The same thing happens when the first line of the program is
Code: QB64: [Select]
  1. SCREEN _NEWIMAGE(320, 200, 256)
  2.  

but when I change it to
Code: QB64: [Select]
  1. SCREEN _NEWIMAGE(320, 200, 32)
  2.  

it works and displays the image.

Problem is that many tutorials, examples and games I found online use SCREEN 13 and other 8-bit modes. How can I run them with the latest QB64? Many times simply changing to 32-bit mode does not help, because a program manipulates the color palette with _COPYPALETTE, or uses LOCATE to display text, which also seems not to work in 32-bit screen modes.
Title: Re: Newbie question about 8-bit screen in new QB64
Post by: bplus on January 20, 2021, 05:44:46 pm
Are you looking for modern interpretations? QB64 can run SCREEN 13.

If so some small sample?

Maybe this?
http://www.qb64.org/wiki/A_Small_Game_Tutorial
Title: Re: Newbie question about 8-bit screen in new QB64
Post by: venex on January 20, 2021, 06:04:43 pm
Thanks for the tutorial link. However, the problem I have is not with drawing, but with loading images. One example of a game which has this problem is Barbarian 1987 from http://barbarian.1987.free.fr/download/download.htm written in QB64 SDL. You can compile it with QB64 v1.4, but it will crash at runtime, because it has problems loading 8-bit image assets. This does not happen when the game is compiled with QB64 SDL. I would like to know if it’s a bug with _PUTIMAGE or maybe something else.
Title: Re: Newbie question about 8-bit screen in new QB64
Post by: Dav on January 20, 2021, 06:36:03 pm
Hi @venex.  Here's a thread that may shed some light on what's happening:
https://www.qb64.org/forum/index.php?topic=1252.0

- Dav
Title: Re: Newbie question about 8-bit screen in new QB64
Post by: bplus on January 20, 2021, 07:34:39 pm
edit
Title: Re: Newbie question about 8-bit screen in new QB64
Post by: FellippeHeitor on January 20, 2021, 08:20:59 pm
8bit mode image loading has been disabled for a while now. You should always check that loaded image handles are <= -1 first (-1 signals the image failed to load):

Code: QB64: [Select]
  2. i = _LOADIMAGE("sample.png")
  3. IF i <= -1 THEN _PUTIMAGE , i ELSE PRINT "Load failed."
Title: Re: Newbie question about 8-bit screen in new QB64
Post by: SMcNeill on January 20, 2021, 09:03:03 pm
_LOADIMAGE in the GL version (from 1.0 onwards), auto-converts images to 32-bits.  If you want to use it with 256 color screens (8-bit), then you’ll need to first load them as 32-bits, and then convert/dither them back down to 256 colors.  RhoSigma’s libraries can handle the task, as can routines in my SaveImage libraries, along with examples by Petr, here on the forum, along with several others that I’m probably just not thinking of at the moment.

Here’s a link to one such discussion for you: https://www.qb64.org/forum/index.php?topic=2701.msg119104#msg119104
Title: Re: Newbie question about 8-bit screen in new QB64
Post by: Petr on January 21, 2021, 10:24:06 am
Hi. This is function for DIRECT inserting 8 bit image to 8 bit screen. This screen palette is always changed so, as it is used in image!

Code: QB64: [Select]
  1. 'LOADIMAGE256 experimental
  2.  
  3. '1] Load image as 32 bit image
  4. '2] find how much colors image contains. If 256 and less, continue. If more than 256, return -1 as error message. (can be solved with  RhoSigma's ditherring which is a part of SaveImage)
  5. '3] create 8 bit image and color palette
  6.  
  7. '4] BIG ASK IS: If is this image inserted to 8 bit screen, copy image palette to this screen? OR try to do the same palette for both images?
  8.  
  9.  
  10. SCREEN _NEWIMAGE(1700, 800, 256)
  11. img8 = LOADIMAGE256("anim1.png")
  13. _PUTIMAGE (0, 0), img8
  14.  
  15.  
  16.  
  17.  
  18. FUNCTION LOADIMAGE256 (img$)
  19.     DEFLNG A-Z
  20.     image = _LOADIMAGE(img$, 32)
  21.     DIM m AS _MEM, clr8(255) AS _UNSIGNED LONG, Clr32 AS _UNSIGNED LONG, test AS LONG, s AS LONG
  22.     FOR s = 0 TO 255
  23.         clr8(s) = 99999
  24.     NEXT s
  25.     m = _MEMIMAGE(image)
  26.     DO UNTIL p& = m.SIZE
  27.         _MEMGET m, m.OFFSET + p&, Clr32~&
  28.         test = 0
  29.         'this block prevent for writing the same color more than 1x to palette array
  30.         DO UNTIL test = 255
  31.             IF clr8(test) = Clr32~& THEN GOTO NextColor
  32.             IF clr8(test) = 99999 THEN EXIT DO
  33.             test = test + 1
  34.         LOOP
  35.         'if is empty place in palette, save this color as next palette color
  36.         IF test > 255 THEN PRINT "Image contains more than 256 colors, can not be directly copyed as 8 bit image": END
  37.         clr8(test) = Clr32
  38.         'color is saved as palette for 8 bit image
  39.         NextColor: p& = p& + 4
  40.     LOOP
  41.     image8 = _NEWIMAGE(_WIDTH(image), _HEIGHT(image), 256)
  42.     'set palette
  44.     N = _MEMIMAGE(image8)
  45.     FOR palett = 0 TO 255
  46.         _PALETTECOLOR palett, clr8(palett), image8
  47.     NEXT
  48.     'create 8 bit mask (set colors 0 to 255 to 8 bit image)
  49.     FOR C = 255 TO 0 STEP -1
  50.         clr~& = clr8(C)
  51.         R& = 0
  52.         R8& = 0
  53.         DO UNTIL R& = m.SIZE
  54.             _MEMGET m, m.OFFSET + R&, Clr32
  55.             IF Clr32 = clr~& THEN _MEMPUT N, N.OFFSET + R8&, C
  56.             R& = R& + 4
  57.             R8& = R8& + 1
  58.         LOOP
  59.     NEXT C
  60.     LOADIMAGE256 = _COPYIMAGE(image8, 256)
  61.     _MEMFREE m
  62.     _MEMFREE N
  63.     _FREEIMAGE image
  64.     _FREEIMAGE image8
  66.  

Of course, you can work with multiple 8-bit images with different palettes at the same time. You just place them on a 32-bit screen. You can make changes using 8-bit colors on 8-bit images in the background. To do this, you need knowledge of the _SOURCE, _DEST and _PALETTECOLOR statements.
Title: Re: Newbie question about 8-bit screen in new QB64
Post by: venex on January 21, 2021, 12:58:55 pm
Thank you very much guys for your help!

I read the materials you linked and also had a quick look at the QB64 source code of internal/c/parts/video/image/src.c and I think I understand a little bit better now how it works.

@Petr your code rocks! I changed my initial program from:
Code: QB64: [Select]
  2. i = _LOADIMAGE("sample.png")
  4.  

into:
Code: QB64: [Select]
  2. i = LOADIMAGE256("sample.png")
  5.  

and now everything works like a charm :) Thank you!
Text Only | Text with Attachments