new amiga 500 doom like !

[spacey]

this is awesome - great work!

[Mathesar]

Originally Posted by KK/Altair

Ok, so what about... a 15:30 episode?

[ Show youtube player ]

ATTENTION: This video is still unlisted, just for you guys to check and maybe find potential problems and such. But if you think all is acceptable (not perfect! ), I'm publishing it.

I also written subtitles, so turn them on if my voice sounds incomprehensible.

Have fun!!

I am speechless! This thing runs so nice and the tricks with the level design blow my mind. It might not be a "full" Doom with multiple floor levels and such but it sure does feel like Doom.

And the speed on an Amiga 500 is just incredible.


Very nice work!

[roondar]

Now this is a long thread so I'm sure I've forgotten part of the discussion, but I wonder if this has been tried without dithering or with different "levels of brightness" for the dither?

Reason for asking is that I wonder how it'd look with slightly less "aggressive" dithering

Edit: Pfft.. Some guy on Facebook is now implying this game won't run this way on a real Amiga 500 because (this is a direct quote): "UAE does not emulate slow CHIP ram accesses, and boatnecks correct". I tried to explain why that's not true, but got nowhere. Gets my blood boiling this kind of stuff - he's basically claiming that the videos are exaggerated and therefore implying that KK/Altair is lying

[d4rk3lf]

Originally Posted by roondar

Edit: Pfft.. Some guy on Facebook is now implying this game won't run this way on a real Amiga 500 because (this is a direct quote): "UAE does not emulate slow CHIP ram accesses, and boatnecks correct". I tried to explain why that's not true, but got nowhere. Gets my blood boiling this kind of stuff - he's basically claiming that the videos are exaggerated and therefore implying that KK/Altair is lying

Well, there will always be people like that, and in my opinion, they doesn't deserve any attention.
I've looked now at video again, and there's 324 likes, and 7 dislikes. Now, who would dislike this, I can't grasp, but usually it's those people that criticize most, are the ones that never did a thing on their own.

And it's not Amiga related, it's everywhere. I guess it's part of human psychology. After all, it's much easier to criticize (negatively) something, then do the work yourself.

[robinsonb5]

Originally Posted by roondar

Now this is a long thread so I'm sure I've forgotten part of the discussion, but I wonder if this has been tried without dithering or with different "levels of brightness" for the dither?

Reason for asking is that I wonder how it'd look with slightly less "aggressive" dithering

I think the dithering would be much improved if it were possible to "invert" the dither for alternating rows, so that the effect was more "chequerboard" than "stripes". The aim of dithering is to "noise-shape" the unavoidable noise caused by reducing the number of colours into high-frequency noise that get smoothed out by the monitor / eye / brain. The current dither does a pretty good job of that horizontally, but vertically, not so much.


I don't know how hard that would be to add, but I imagine if it was easy KK would have already done it!


The other option that was discussed earlier is to use HAM mode. The ECS chipset can be tricked into displaying HAM mode with fixed control planes while using the same amount of Chip RAM bandwidth as 16-colour mode - so a fixed repeating pattern - say, RGBG - is possible. (The same trick doesn't work on AGA, but since on AGA you can just increase the fetch mode to solve the bandwidth problem, that's not a showstopping problem.)

[roondar]

Originally Posted by d4rk3lf

Well, there will always be people like that, and in my opinion, they doesn't deserve any attention.
I've looked now at video again, and there's 324 likes, and 7 dislikes. Now, who would dislike this, I can't grasp, but usually it's those people that criticize most, are the ones that never did a thing on their own.

And it's not Amiga related, it's everywhere. I guess it's part of human psychology. After all, it's much easier to criticize (negatively) something, then do the work yourself.

Yeah, I know.

But I just don't like it when people downplay other's achievements. I mean, it's not only KK/Altair he's downplaying here - he's also claiming that UAE in general is just not accurate. Which is basically dissing 20+ years of development by a whole bunch of people. That doesn't sit right with me.

Oh well... I guess you're right, I should probably ignore it.

Originally Posted by robinsonb5

I think the dithering would be much improved if it were possible to "invert" the dither for alternating rows, so that the effect was more "chequerboard" than "stripes". The aim of dithering is to "noise-shape" the unavoidable noise caused by reducing the number of colours into high-frequency noise that get smoothed out by the monitor / eye / brain. The current dither does a pretty good job of that horizontally, but vertically, not so much.

I don't know how hard that would be to add, but I imagine if it was easy KK would have already done it!

Well, my reason for wanting to see it toned down is basically that I don't really like some of the colours used . It's silly perhaps (and extremely subjective ), but especially the hand looks a bit "off" to me due to the red used. Doesn't change in any way just how impressive the game is, but anything that might help make it look even better seems like a good idea to me

Then again, perhaps I'm the only one that feels this way

[chb]

First let me say, extremely impressive engine and very nice demo video! The frame rate looks amazing for the level of detail! And there's even some shaded floor in front of the door - is it somehow drawn using the wall renderer? I guess you aren't doing a complete floor mapper at that framerate. Anyway, looks great and adds a lot to the atmosphere.

Originally Posted by robinsonb5

I think the dithering would be much improved if it were possible to "invert" the dither for alternating rows, so that the effect was more "chequerboard" than "stripes". The aim of dithering is to "noise-shape" the unavoidable noise caused by reducing the number of colours into high-frequency noise that get smoothed out by the monitor / eye / brain. The current dither does a pretty good job of that horizontally, but vertically, not so much.

I don't know how hard that would be to add, but I imagine if it was easy KK would have already done it!

The problem is the vertical doubling - the vertical stripe pattern is essentially free, while there is no easy way to get a real checkerboard pattern, it would require additional blitter and/or CPU passes. You could of course fake it and shift every second line by one pixel, but that's not a real checkerboard and decreases resolution. I think Britelite experimented with it for his Wolf3D engine and came to the conclusion that it looks worse than the stripe pattern.

But I think KK did focus his effort on the engine at this point, not on the graphics. It looks like everything is machine-converted from the freedoom assets with a fixed palette, which probably explains the strange dithering on the hand etc.. I guess a final version would have a much improved graphical assets.

[robinsonb5]

Originally Posted by chb

The problem is the vertical doubling - the vertical stripe pattern is essentially free, while there is no easy way to get a real checkerboard pattern, it would require additional blitter and/or CPU passes.

Yeah I wasn't sure how the dithering is actually done - is it simply that each texture pixel contains a pair of 16-colour pixels which average to something close to the original colour? If so, it might be worth experimenting with swapping the pixels on alternate rows of the texture - just to break up the vertical lines that often seem to extend the full height of the screen. It'd look fine for near textures, though it might cause obnoxious "shimmering" in far-off textures.

I wonder, too, if pixel swapping could be achieved by modifying the C2P algorithm for alternate lines?

Like I said, if it was easy, I'm sure it would have been done already!

[d4rk3lf]

Originally Posted by roondar

but especially the hand looks a bit "off" to me

In the youtube comments, he explained that he didn't spend time on the hand, and just did it roughly.

Originally Posted by chb

And there's even some shaded floor in front of the door - is it somehow drawn using the wall renderer?

Look at the (making of) video on the first page.
At some point he mentioned floors and how lightning hits it.

Not sure if that's what you're asking for, thought.

[Mathesar]

Originally Posted by chb

First let me say, extremely impressive engine and very nice demo video! The frame rate looks amazing for the level of detail! And there's even some shaded floor in front of the door - is it somehow drawn using the wall renderer? I guess you aren't doing a complete floor mapper at that framerate. Anyway, looks great and adds a lot to the atmosphere.

I like those little things too. It shows that he is not only a programmer but also creative with level design. I guess that comes with being a demo coder . It makes the whole thing extremely Doom-like and atmospheric.
And now for the important question; how many floppy disks will the game be? Can't wait to slip a floppy labelled "Dread" into an A500 and see it boot into this.

[KK/Altair]

Question is how configurable this is? Can I enable/disable floor and ceiling textures if I want to play it with more powerfull amiga?

I have no plans for floor/ceiling textures right now. If you are lucky enough to have more powerful Amiga, you can play some great FPSes already. My primary goal is A500, which lacks a bit in this genre.

It might not be a "full" Doom with multiple floor levels and such but it sure does feel like Doom.

Once I've got this look I was more concerned to get it to play like Doom rather to look like it.

"UAE does not emulate slow CHIP ram accesses, and boatnecks correct"

Then it looks like my A500 does not emulate slow CHIP ram accesses either, because performance values match. But jokes aside, you really have to be careful with UAE settings to get proper results - even the "quick" confgis are valid and that might have been what he* was referring to.

(* at this point I will refrain from using "he/she"; after 9000+ views, my video audience stats show 100.0% male viewers )

Well, there will always be people like that, and in my opinion, they doesn't deserve any attention.

This kind of people don't upset me at all, or anything, so I just respond kindly. They are free to believe what they want until I release a demo. So I'm not afraid of any such comments, because time will tell.

I think the dithering would be much improved if it were possible to "invert" the dither for alternating rows, so that the effect was more "chequerboard" than "stripes".

It DOES get a lot of improvement this way, but I have to do additional Blitter pass to swizzle that and it cuts into the performance. It was my original implementation and I'll probably add this back at some point as an option, but right now I have to set some reasonable performance baselines to keep framerate playable. It's quite a dynamic game, so every saved scanline counts.

The other option that was discussed earlier is to use HAM mode. The ECS chipset can be tricked into displaying HAM mode with fixed control planes while using the same amount of Chip RAM bandwidth as 16-colour mode - so a fixed repeating pattern - say, RGBG - is possible. (The same trick doesn't work on AGA, but since on AGA you can just increase the fetch mode to solve the bandwidth problem, that's not a showstopping problem.)

The 4bpp HAM dithering was already tested by someone earlier in this thread on various images, and it indeed could work using the index-green pattern. I'm tempted to test it at some point.

It's silly perhaps (and extremely subjective ), but especially the hand looks a bit "off" to me due to the red used.

I stopped caring about fixing hand dithering the moment I got it on screen. Ok, not completely stopped caring, but it got a really low place on my ToDo list.

And there's even some shaded floor in front of the door - is it somehow drawn using the wall renderer? I guess you aren't doing a complete floor mapper at that framerate.

More or less the engine computes line splits and fills space between them with textures, solid color or sky. I have to fill floors/ceilings anyway, so introducing a split is just like adding 0-height step there. Except that with no step it's much cheaper.

You could of course fake it and shift every second line by one pixel, but that's not a real checkerboard and decreases resolution.

I experimented with it as well. In general, it looks good on the walls, but visually blurs all map geometry edges, so I didn't really like it. But I might implement it as option as well at some point.

It looks like everything is machine-converted from the freedoom assets with a fixed palette, which probably explains the strange dithering on the hand etc..

Exactly. I have played with dithering algorithms, texture choice and palettes, but only up to the point I've got a set of textures looking acceptable. This will need per-map tweaking anyway, so I didn't feel the need to dig further there so far.

Yeah I wasn't sure how the dithering is actually done - is it simply that each texture pixel contains a pair of 16-colour pixels which average to something close to the original colour? If so, it might be worth experimenting with swapping the pixels on alternate rows of the texture - just to break up the vertical lines that often seem to extend the full height of the screen. It'd look fine for near textures, though it might cause obnoxious "shimmering" in far-off textures.

Exactly that. But I didn't experiment with order swapping either.

I wonder, too, if pixel swapping could be achieved by modifying the C2P algorithm for alternate lines?

There are no alternate lines. Each line is simply shown twice using Copper. Also the C2P algorithm never operates at that granularity. Texel colors are packed in such way, that C2P always keeps bit pairs together.

And now for the important question; how many floppy disks will the game be? Can't wait to slip a floppy labelled "Dredd" into an A500 and see it boot into this.

OVER 9000!!!!
Seriously, I'd love people start making content, so maybe eventually we'll reach this number.

And the codename is "Dread".

[roondar]

Originally Posted by KK/Altair

Then it looks like my A500 does not emulate slow CHIP ram accesses either, because performance values match. But jokes aside, you really have to be careful with UAE settings to get proper results - even the "quick" confgis are valid and that might have been what he* was referring to.

(* at this point I will refrain from using "he/she"; after 9000+ views, my video audience stats show 100.0% male viewers )

No, he sadly was not.

He really does seem to believe that UAE can't be accurate for things like this in the first place and that there's no point for UAE to even be accurate in these areas (because in his mind it's not a development tool) and therefore it isn't.

This kind of people don't upset me at all, or anything, so I just respond kindly. They are free to believe what they want until I release a demo. So I'm not afraid of any such comments, because time will tell.

I just dislike it when people downplay the hard work of others. It's not personal, but I still don't like it.

Anway, you're right of course. Best to move on, I said my piece

I stopped caring about fixing hand dithering the moment I got it on screen. Ok, not completely stopped caring, but it got a really low place on my ToDo list.

Fair enough, I can understand that only too well

[KK/Altair]

Ok, since you guys are so good beta-testers of my videos, care to share thoughts about this little recording?

No PC nor Amiga footage - just me sketching the overall map plan.

EDIT: This is just an unlisted preview video, so NO SPREAD.
Once it's done, it will make into the next episode, or get re-recorded. Or something.


[ Show youtube player ]

This was captured in one (first) take, and I somehow think it's so-so, which equals "acceptable" when you take perfectionism out of the equation. I'll definitely process the audio (EQ some bass out, compression/noise gate, volume at -13 LUFS) when it makes into the next episode along with recording of the actual mapping process, but I would like you to share your thoughts about it.
- is it acceptable in general? or should I try some more takes?
- what do you think about the overall plan of the demo map?
- also: how much time should I dedicate to video about rest of the mapping process? make it an in-depth description, or a timelapse?

The response after the first episode was amazing (20k+ views, 600+ subs), so hopefully after a few more I'll get a hang of it.

Thanks!

P.S. I've also checked how well this runs on cycle-exact emulation of off-the-shelf A1200 - constantly pumping out 22-23 FPS and flawless gameplay.

EDIT: P.S.2: Yes, I have a cat.

[d4rk3lf]

I love it.
And this (pre-planning), is a very important step, that majority of the people, either skip, either over-complicate (like, trying to sketch a map with ruler and perfect proportions).
Really love the way you think, and work.

Originally Posted by KK/Altair

- is it acceptable in general? or should I try some more takes?
- what do you think about the overall plan of the demo map?
- also: how much time should I dedicate to video about rest of the mapping process? make it an in-depth description, or a timelapse?

Well, the best thing, imho, would be to have 2 types of videos. One type is short - promotional, and the other's is more in depth tutorials, like this. For the first type of video, you could do a timelapse, and just mention that you will leave the link, to more detailed tutorial, to anyone interested. Unfortunately, not everyone will be interested in this, and there will be people that would only like to play a game, and see a short promotional video.

I am just afraid that all these videos don't distract you too much from the creating of the game itself....

Originally Posted by KK/Altair

P.S.2: Yes, I have a cat.

I have one too.

[KK/Altair]

Well, the best thing, imho, would be to have 2 types of videos. One type is short - promotional, and the other's is more in depth tutorials, like this. For the first type of video, you could do a timelapse, and just mention that you will leave the link, to more detailed tutorial, to anyone interested. Unfortunately, not everyone will be interested in this, and there will be people that would only like to play a game, and see a short promotional video.

I'm not sure I have too much mapmaking knowledge left to say at this point. The basics were in the previous video, and details are probably better left until people will be able to actually make maps themselves.

I am just afraid that all these videos don't distract you too much from the creating of the game itself....

That was probably the case at one point, but right now they rather motivate me to keep pushing the project - once you said "Ep. 01", you better say "Ep. 02", don't you?

The thing is, I still work full time, have absorbing kid at home, and more often than not don't have any creative energy left to focus and keep pushing at evenings. So it's more about keeping the progress and doing things when I feel like it, than exhausting myself to get it done ASAP. I just don't want it to feel like second work at this time (I don't even have Patreon or anything).

But having said that, I still want to have fun with it, and all of you to have the same as well.

[utri007]

Has situations changed now? Most of us work home now.

[KK/Altair]

Originally Posted by utri007

Has situations changed now? Most of us work home now.

I worked remotely and completely alone all the time, so nothing changed for me in this matter, except that kindergartens are closed and I spend a bit more time with my kid.

But I'm returning to the project from time to time and working towards the next episode. Slowly, but steadily. So stay tuned.

[d4rk3lf]

You can name the game after your kid name, or main protagonist to have his name.
That could inspire you to work on it even more regularly.

Have you thought about campaign?
It would be cool having some nice story, with some twists, epic ending, and cool antagonist. All this could drive you forward in thinking about engine specific tasks, on specific missions (if you try to add more variety then just finding keys). SF would be the time period, I assume?

[rothers]

If you were to make the gun bigger and place it Duke3D style you could save some rendering time on the right side of the screen.

[KK/Altair]

Originally Posted by d4rk3lf

You can name the game after your kid name, or main protagonist to have his name.
That could inspire you to work on it even more regularly.

I don't think so. Also I don't want to put my 4yo anywhere near that kind of game either (although he already enjoyed Silas Adventure - but that had no firearms and no blood).

Have you thought about campaign?
It would be cool having some nice story, with some twists, epic ending, and cool antagonist. All this could drive you forward in thinking about engine specific tasks, on specific missions (if you try to add more variety then just finding keys).

No - I'm more an engine guy. I like making games and maps, but technical aspects will be my main focus here (including core gameplay). After all, if stories were my main focus, I wouldn't have picked Doom, right?

SF would be the time period, I assume?

I'm slightly more into fantasy/medieval/mythos, so Heretic/Quake blend would be preferred, but FreeDoom simply offered assets that allowed me to jump straight into the problems of coding that thing on poor old A500.

If you were to make the gun bigger and place it Duke3D style you could save some rendering time on the right side of the screen.

I'd love to have the gun on the side, but couldn't find free quality assets that would fit the rest and didn't bother making my own at this point.

But don't expect to win big on performance. The problem is, that when I'm drawing the 3D world, I don't know which animation frame will be used for the gun. I check the trigger input just before blitting the gun at the very end of rendering (even after the C2P) to switch it to first frame of attack if player happens to fire. This saves up to 100ms of fire trigger delay and makes firing much more fun. At minimum assumed 10 FPS the rest of the game might feel laggy, and the firing result (monster pain, splats and pellets) won't draw until the next frame - but the weapon itself fires immediately and that's just enough for it to feel responsive.

Anyway, I'm just finishing edits on Ep 02. Voiceover recording next (I work on my sofa just with subtitles), and it will be ready to air.