32/24Bit Color Depth!
Moderator: GZDoom Developers
- David Ferstat
- Posts: 1113
- Joined: Wed Jul 16, 2003 8:53 am
- Location: Perth, Western Australia
- Contact:
All right, I give in. Where are you guys all getting your cutesy avatars from?
And in relation to the topic, I dont see any problem with using ZDGL for 32 bit colour. If you're really against GL, change the filtering mode to none and see how it looks exactly like ZDoom but with correct perspective. Considering how the eventual arrival of polymost will also mean true perspective in regular Zdoom anyway, there is not any reason to ostracise ZDGL.
And in relation to the topic, I dont see any problem with using ZDGL for 32 bit colour. If you're really against GL, change the filtering mode to none and see how it looks exactly like ZDoom but with correct perspective. Considering how the eventual arrival of polymost will also mean true perspective in regular Zdoom anyway, there is not any reason to ostracise ZDGL.
- Graf Zahl
- Lead GZDoom+Raze Developer
- Posts: 49234
- Joined: Sat Jul 19, 2003 10:19 am
- Location: Germany
charris wrote:The Mac might have other optimizations.
PS: I'm going to say this and duck, what about a move into OpenGL instead of a software based renderer?
That would be an option if:
-ZDoomGL performed as well as software ZDoom
-it had less rendering glitches
-it could handle self referencing sectors and other rendering tricks
but unfortunately right now it just produces errors with many complex maps or they are so slow that they become unplayable.
Just enter the first cave of E1M6 in P:AR. Even on my P4 3.2 Ghz with a Geforce 5900XT the frame rate drops to 10 fps and there's a shitload of not very pretty rendering bugs. With PrBoom's GL renderer I get more than 30 fps and it displays without any problems at all.
So let's see what Randy is doing with Polymost. If this goes well OpenGL support may improve significantly.
@Chilvence: http://www.vgcats.com ... just grab a comic, and crop away 
@Graf Zahl: Would switching to GL still allow us to use Line_Horizon? That really annoyed me when I found out ZDoomGL didn't do it.

@Graf Zahl: Would switching to GL still allow us to use Line_Horizon? That really annoyed me when I found out ZDoomGL didn't do it.
- Tormentor667
- Posts: 13556
- Joined: Wed Jul 16, 2003 3:52 am
- Preferred Pronouns: He/Him
- Operating System Version (Optional): Windows 11
- Graphics Processor: nVidia (Modern GZDoom)
- Location: Germany
- Contact:
Well, when I started this topic, I had something else in mind: Keep the DoomPalette for textures, things and everything else, but add the possibility of 32Bit ingame, so you can apply different colors for sector_setcolor, sector_setfade and other stuff like that. Even the light levels itself would look by far better from 0 - 192 if we hade 32Bit! Think about cyan or purple faded rooms, bright blue lights in rooms or things like that
This would enhance doom, it really would and there wouldn't be a need to change something about the doompalette itself right now!

- Graf Zahl
- Lead GZDoom+Raze Developer
- Posts: 49234
- Joined: Sat Jul 19, 2003 10:19 am
- Location: Germany
That depends on the implementation. An exact replica is not possible because it'd require infinite vertex coordinates but you can come close if you think about it the right way.Shinjanji wrote: @Graf Zahl: Would switching to GL still allow us to use Line_Horizon? That really annoyed me when I found out ZDoomGL didn't do it.
- HobbsTiger1
- Posts: 1235
- Joined: Fri Jan 07, 2005 7:29 pm
- Location: #DMClub
- Contact:
15bit would be R5G5B5. This format doesnt really exist AFAIK, because 15bits is a very odd number. Simular 16bit formats that im sure exist are: A1R5G5B5 and R5G6B5. With A1R5G5B5, the one bit is given to the alpha channel, which can indicate "solid" or "transparent". With R5G6B5 the extra bit is used for more precise green colors, because the human eye is more sensitive to green.
- Chris
- Posts: 2978
- Joined: Thu Jul 17, 2003 12:07 am
- Graphics Processor: ATI/AMD with Vulkan/Metal Support
15-bit color is 16-bits, but the last bit is not used (this makes it so each color component has the same number of bits.. 5 each; though some image formats can actually use the last bit as a transpancy mask (1-bit alpha)). 24-bit color is 24-bit color. It's 3 bytes wide and thus misaligned on all CPUs (since there are no 24-bit CPUs, AFAIK). 32-bit color has 24-bits of color and 8-bits of junk, to properly align to 32-bit for more efficient performance. Images usually use the last 8 bits as an alpha channal.