Author Topic: Hardware vs Software Images  (Read 4360 times)

0 Members and 1 Guest are viewing this topic.

Offline TerryRitchie

  • Seasoned Forum Regular
  • Posts: 495
  • Semper Fidelis
    • View Profile
Hardware vs Software Images
« on: September 07, 2018, 12:58:25 am »
It's been a Loooong time since I have played around with hardware image capability in QB64, and even then it was in the experimental stage.

Just so I have this straight, once I have committed an image as hardware (mode 33) I can do nothing with that image except for displaying it (_PUTIMAGE, _MAPTRIANGLE), or copying it (_COPYIMAGE).

Even trying to set a hardware image as a source, to say read a POINT() from it will not work, correct?

If I need to make a change to a hardware image, I'll need to have a software image copy of it, make the changes on the software image, then once again convert the image to hardware with _COPYIMAGE to another handle. Is that correct?
In order to understand recursion, one must first understand recursion.

Offline TerryRitchie

  • Seasoned Forum Regular
  • Posts: 495
  • Semper Fidelis
    • View Profile
Re: Hardware vs Software Images
« Reply #1 on: September 07, 2018, 10:29:04 pm »
So I take it no one is using hardware images. :)

I've confirmed through playing around that hardware images are in fact just for viewing.
In order to understand recursion, one must first understand recursion.

Offline Omerta7486

  • Newbie
  • Posts: 33
  • √𝜋²
    • View Profile
Re: Hardware vs Software Images
« Reply #2 on: September 10, 2018, 11:55:13 pm »
I think you may[/i] be able to _putimage from one hardware image to another. Don't quote me on that! Lol. But, no. You can't use any pixel level functions, or any chromatic or graphical functions.
The knowledge that's seeking the favor of another means the murder of self.

Latest version of my game, here  Omertris: Invasion

Offline TerryRitchie

  • Seasoned Forum Regular
  • Posts: 495
  • Semper Fidelis
    • View Profile
Re: Hardware vs Software Images
« Reply #3 on: September 11, 2018, 12:11:23 am »
You are correct. I've tested most of the scenarios out and it would seem the limit is what you state. Thanks for the feedback. :)
In order to understand recursion, one must first understand recursion.

Offline TerryRitchie

  • Seasoned Forum Regular
  • Posts: 495
  • Semper Fidelis
    • View Profile
Re: Hardware vs Software Images
« Reply #4 on: September 14, 2018, 02:19:22 pm »
Another cool thing about hardware images ... they do not interact at all with software images.

Move a hardware image around on a software screen and the software screen is unaffected. No need to CLS  or restore the background at every iteration.

However, because of this behavior it was necessary for me to implement software and hardware sprites into my updated sprite library. Which is cool, because a programmer using the library can now tell the sprite which type it is and change that behavior at any time.
In order to understand recursion, one must first understand recursion.

FellippeHeitor

  • Guest
Re: Hardware vs Software Images
« Reply #5 on: September 14, 2018, 02:49:50 pm »
Hardware images are so cool that you can even place them on SCREEN 0:

Code: QB64: [Select]
  1. 'SCREEN 0 is 640x400 pixels
  2. DIM SHARED circleImage&
  3. temp& = _NEWIMAGE(50, 50, 32)
  4. _DEST temp&
  5. CIRCLE (25, 25), 20
  6. PAINT (25, 25), _RGB32(255, 0, 0), _RGB32(255, 255, 255)
  7. circleImage& = _COPYIMAGE(temp&, 33)
  8.  
  9. DIM SHARED rStep AS INTEGER, cStep AS INTEGER
  10. DIM SHARED xStep AS INTEGER, yStep AS INTEGER
  11.  
  12. r = 1: rStep = 1
  13. c = 1: cStep = 1
  14. x = 320: xStep = 5
  15. y = 200: yStep = 5
  16. m$ = "This is SCREEN 0!"
  17. COLOR 15, 1
  18.  
  19.     CLS
  20.     IF TIMER - lastTextMovement > .2 THEN
  21.         lastTextMovement = TIMER
  22.         moveText
  23.     END IF
  24.     showText
  25.  
  26.     moveImage
  27.     showImage
  28.     _DISPLAY
  29.     _LIMIT 30
  30.  
  31. SUB moveText
  32.     c = c + cStep
  33.     IF c + LEN(m$) > 80 THEN c = c - 1: cStep = cStep * -1: COLOR INT(RND * 14) + 1
  34.     IF c < 1 THEN c = c + 1: cStep = cStep * -1: COLOR INT(RND * 14) + 1
  35.  
  36.     r = r + rStep
  37.     IF r > 25 THEN r = r - 2: rStep = rStep * -1: COLOR , INT(RND * 7) + 1
  38.     IF r < 1 THEN r = r + 2: rStep = rStep * -1: COLOR , INT(RND * 7) + 1
  39.  
  40. SUB showText
  41.     LOCATE r, c
  42.     PRINT m$;
  43.  
  44. SUB moveImage
  45.     x = x + xStep
  46.     IF x + _WIDTH(circleImage&) > 640 THEN x = x - 1: xStep = xStep * -1
  47.     IF x < 1 THEN x = x + 1: xStep = xStep * -1
  48.  
  49.     y = y + yStep
  50.     IF y + _HEIGHT(circleImage&) > 400 THEN y = y - 1: yStep = yStep * -1
  51.     IF y < 1 THEN y = y + 1: yStep = yStep * -1
  52.  
  53. SUB showImage
  54.     _PUTIMAGE (x, y), circleImage&
  55.  

I used the technique a lot in my ever under-development Spaceship game too.

Offline Pete

  • Forum Resident
  • Posts: 2361
  • Cuz I sez so, varmint!
    • View Profile
Re: Hardware vs Software Images
« Reply #6 on: September 14, 2018, 02:56:06 pm »
What's the big deal? I always have a ball in SCREEN 0!

Pete :D
Want to learn how to write code on cave walls? https://www.tapatalk.com/groups/qbasic/qbasic-f1/

FellippeHeitor

  • Guest
Re: Hardware vs Software Images
« Reply #7 on: September 14, 2018, 04:23:43 pm »

Offline Gets

  • Newbie
  • Posts: 28
    • View Profile
Re: Hardware vs Software Images
« Reply #8 on: September 14, 2018, 06:23:33 pm »
I use hardware images, but like you pointed out, you mainly trade an increase in speed for a uniform drop in functionality. You also have to separate software and hardware graphic routines since they're displayed completely separately and there are some compatibility issues with hardware images that require running everything in software mode for some people. Can't just draw a few lines on top of a sprites, I have to store line data with the sprite, load it when rendering the sprite, and then draw them right before the screen is displayed. I'm assuming the issues are related to the old version of OpenGL QB64 uses, but I don't know how to detect the problem ahead of time or recreate it.

Offline TerryRitchie

  • Seasoned Forum Regular
  • Posts: 495
  • Semper Fidelis
    • View Profile
Re: Hardware vs Software Images
« Reply #9 on: September 14, 2018, 11:22:56 pm »
Can't just draw a few lines on top of a sprites, I have to store line data with the sprite, load it when rendering the sprite, and then draw them right before the screen is displayed. I'm assuming the issues are related to the old version of OpenGL QB64 uses, but I don't know how to detect the problem ahead of time or recreate it.

I'm not quite following what you mean here. Could you please elaborate more on this? I have not run into any issues yet because I'm still in the beginning stages of the library. Any heads up on issues I may encounter I am very interested in. Thanks :)

Edit: after reading a few times I think I see what you mean. Drawing a LINE(), or anything software related, will be drawn under the hardware image. Yes, I see that as a potential issue to overcome.

Hmmm... I suppose a second software screen would need to be maintained in the background that contains only those graphics you wish to overlay onto a hardware sprite. The portion of the screen where the sprite resides would need to be lifted and laid over top of the software sprite, then converted to a hardware sprite using mode 33. Yikes, I'll need to look into this.
« Last Edit: September 14, 2018, 11:38:17 pm by TerryRitchie »
In order to understand recursion, one must first understand recursion.

FellippeHeitor

  • Guest
Re: Hardware vs Software Images
« Reply #10 on: September 14, 2018, 11:50:42 pm »
You can always change _DISPLAYORDER.

Offline TerryRitchie

  • Seasoned Forum Regular
  • Posts: 495
  • Semper Fidelis
    • View Profile
Re: Hardware vs Software Images
« Reply #11 on: September 14, 2018, 11:59:17 pm »
You can always change _DISPLAYORDER.

Will that work?

_DISPLAYORDER _HARDWARE, _SOFTWARE

The screen will always be software. If I need to refresh an entire background, any hardware images drawn will be completely covered by the software screen's background. Maybe?

Ok, I'll need to play around with this. Thanks for the idea Fellippe.
In order to understand recursion, one must first understand recursion.

FellippeHeitor

  • Guest
Re: Hardware vs Software Images
« Reply #12 on: September 15, 2018, 12:02:42 am »
Unless you CLS with a transparent color, then it's a transparent software screen with hardware images beneath.

Code: QB64: [Select]
  1. CLS , _RGBA32(0, 0, 0, 0)
  2. 'or
  3. CLS , _RGB32(0, 0) 'valid for the latest dev build
  4. 'or
  5. CLS , 0

Offline Gets

  • Newbie
  • Posts: 28
    • View Profile
Re: Hardware vs Software Images
« Reply #13 on: September 15, 2018, 01:55:04 am »
The main issues were related to the fact that hardware images won't work at all for some people though(don't know how common these issues are), and the solution was to manually keep all hardware and software graphic work separate,mimicking what _Displayorder would normally do. That way everything would function the same even if you couldn't use hardware images.

 Right now my programs are set up so that I only need to change one variable to  switch between a hardware and software version. When I first discovered it was an issue at all, it was with a program that was already completed and depended on _Displayorder to keep the graphics separated for me, so going back through the program and separating everything manually was a hassle.