Does Jurassic Park use pseudo high-res mode?

Discussion of hardware and software development for Super NES and Super Famicom. See the SNESdev wiki for more information.

Moderator: Moderators

Forum rules
  • For making cartridges of your Super NES games, see Reproduction.
User avatar
Dwedit
Posts: 4924
Joined: Fri Nov 19, 2004 7:35 pm
Contact:

Re: Does Jurassic Park use pseudo high-res mode?

Post by Dwedit »

Once again, the question of Psuedo Hires Mode vs Color Math is decided by whether you want all sprites to participate or not. For color math, only half of the sprite palettes will participate in color math. For pseudo hires, every sprite palette participates in the half transparency effect.
Last edited by Dwedit on Mon Jun 27, 2022 1:46 pm, edited 2 times in total.
Here come the fortune cookies! Here come the fortune cookies! They're wearing paper hats!
User avatar
rainwarrior
Posts: 8732
Joined: Sun Jan 22, 2012 12:03 pm
Location: Canada
Contact:

Re: Does Jurassic Park use pseudo high-res mode?

Post by rainwarrior »

Dwedit wrote: Mon Jun 27, 2022 10:20 amOnce again, the question of Psuedo Hires Mode vs Color Math is decided by whether you want all sprites to participate or not. For color math, only half of the sprite palettes will participate in color math. For pseudo hires, every sprite palette participates in the half transparency effect.
I thought this was just for the main screen?

If you flip your usage of main and sub screen (i.e. the "overlapping" layer is occluding sprites on the main screen and not the sub) then all of them participate.

So, with colour math you have the option to blend only half the sprites if you need that, but you also have the option to blend them all.

(Or have I misunderstood this?)
iNCEPTIONAL

Re: Does Jurassic Park use pseudo high-res mode?

Post by iNCEPTIONAL »

Dwedit wrote: Mon Jun 27, 2022 10:20 am Once again, the question of Psuedo Hires Mode vs Color Math is decided by whether you want all sprites to participate or not. For color math, only half of the sprite palettes will participate in color math. For pseudo hires, every sprite palette participates in the half transparency effect.
In the two game examples I've mentioned in this thread, I don't see any semi-transparent sprites or similar, so I still don't get why they went with the pseudo high-res, if what you say is its main advantage. Unless I'm not quite understanding what you're saying here?
User avatar
rainwarrior
Posts: 8732
Joined: Sun Jan 22, 2012 12:03 pm
Location: Canada
Contact:

Re: Does Jurassic Park use pseudo high-res mode?

Post by rainwarrior »

iNCEPTIONAL wrote: Mon Jun 27, 2022 10:47 am
Dwedit wrote: Mon Jun 27, 2022 10:20 am Once again, the question of Psuedo Hires Mode vs Color Math is decided by whether you want all sprites to participate or not. For color math, only half of the sprite palettes will participate in color math. For pseudo hires, every sprite palette participates in the half transparency effect.
In the two game examples I've mentioned in this thread, I don't see any semi-transparent sprites or similar, so I still don't get why they went with the pseudo high-res, if what you say is its main advantage. Unless I'm not quite understanding what you're saying here?
As far as has been demonstrated the effects in Jurassic Park and Kirby 3 could 100% be done with hardware blending.

However, maybe what Dwedit suggested could indicate a reason for wanting to try this method. Maybe you first try putting your BG3 "transparent" layer on the sub screen and noticed some of your sprites piercing through it, and didn't realize you could get them all to work by putting BG3 on your main screen instead.

Because a 50% blend is half and half, it doesn't normally matter which is main and which is sub, except for the special ability for half the sprites to be exempt from blending on the main screen. Essentially, you can just swap the two screens to make this feature of the sprites optional.

So if you don't realize that you could do it both ways like this, that's something that could cause someone to seek out the pseudo-hires blend method instead.
User avatar
jeffythedragonslayer
Posts: 344
Joined: Thu Dec 09, 2021 12:29 pm

Re: Does Jurassic Park use pseudo high-res mode?

Post by jeffythedragonslayer »

iNCEPTIONAL wrote: Mon Jun 27, 2022 8:25 am
calima wrote: Mon Jun 27, 2022 7:27 am If you're that curious about why they used it, ask them. It's not like the production team is dead yet.
Well, I'm not THAT curious, not enough to go hunt down the original dev team and email them about it, knowing fine well they probably won't even respond. lol

It's an interesting thought though.
That typically hasn't been my experience that developers ignore emails, except younger developers who don't value building connections. Since we are having a long conversation on a forum they are probably going to find out that someone wants to know how their game works this badly at some point anyway. :wink:
turboxray
Posts: 348
Joined: Thu Oct 31, 2019 12:56 am

Re: Does Jurassic Park use pseudo high-res mode?

Post by turboxray »

This thread reminds me that I need to pick up a PAL SNES.. for testing out its composite.
bocchi
Posts: 9
Joined: Mon May 16, 2022 7:35 pm

Re: Does Jurassic Park use pseudo high-res mode?

Post by bocchi »

rainwarrior wrote: Mon Jun 27, 2022 10:30 am
Dwedit wrote: Mon Jun 27, 2022 10:20 amOnce again, the question of Psuedo Hires Mode vs Color Math is decided by whether you want all sprites to participate or not. For color math, only half of the sprite palettes will participate in color math. For pseudo hires, every sprite palette participates in the half transparency effect.
I thought this was just for the main screen?

If you flip your usage of main and sub screen (i.e. the "overlapping" layer is occluding sprites on the main screen and not the sub) then all of them participate.

So, with colour math you have the option to blend only half the sprites if you need that, but you also have the option to blend them all.

(Or have I misunderstood this?)
You are correct.
I feel that the wording in fullsnes is misleading, since it indeed is only on the mainscreen side the palette matters. If sprites are on the subscreen, they all can be blended with the mainscreen. Furthermore, with some priority trickery, you can have some be translucent and others not on a sprite by sprite basis.
User avatar
jeffythedragonslayer
Posts: 344
Joined: Thu Dec 09, 2021 12:29 pm

Re: Does Jurassic Park use pseudo high-res mode?

Post by jeffythedragonslayer »

turboxray wrote: Mon Jun 27, 2022 11:29 am This thread reminds me that I need to pick up a PAL SNES.. for testing out its composite.
I would to love where you find one, I want to run PAL tests too...
Pokun
Posts: 2681
Joined: Tue May 28, 2013 5:49 am
Location: Hokkaido, Japan

Re: Does Jurassic Park use pseudo high-res mode?

Post by Pokun »

A PAL SNES is cheap, try eBay. The problem if you live in NTSC land is that your TV probably doesn't support PAL in most cases. If you do get one, make sure to get PAL SNES controllers as well because NTSC controllers doesn't work (the opposite isn't true though).

rainwarrior wrote: Mon Jun 27, 2022 11:02 am Kirby 3 could 100% be done with hardware blending.
But can you use colour math on two BG layers at the same time?


That crude dithering-transparency in Secret of Mana dialogue box backgrounds is also used in Harvest Moon. Some of the later RPG Makers for Windows (RPG Maker VX maybe? I forgot) are also using a dithered patterns for dialogue box backgrounds, and those are made for VGA monitors (or whatever is used by computers nowadays).


I did some non-scientific tests with Seiken Densetsu 2 (Secret of Mana), Bokujou Monogatari (Harvest Moon), Jurrasic Park and Hoshi no Kirby 3 using composite, S-video and RGB cables with my SFC.
I used the Japanese version of all games except Jurrasic Park where I used the USA version.

My setup includes an 1-CHIP-02 model SFC, a custom RGB cable with csync for NTSC SFC/SNES, a custom S-video cable for NTSC SFC/SNES/N64/Gamecube, official Nintendo NTSC composite cable for NTSC SFC/SNES/AV-Famicom/N64/Gamecube and my CRT TV is a Sony PVM-1454QM which means 600 TVL. This is a bit higher resolution than what a normal consumer TV has which usually is a bit below 200 TVL. What this means in practice is that the black scanlines doesn't blend together as much as on a typical consumer level CRT TV, and the black scanlines are actually very visible even on composite with this kind of setup. The fact that the SFC is 1-CHIP means that the video signal is much cleaner and less blurry than on older SFC models.
In other words, this might not be the kind of setup that the game designers designed the games for.


Seiken Densetsu 2
CVBS:
The the diamond pattern background in dialogue windows is kinda visible and doesn't blend together too much, but it's also a bit blurry overall. The dithering pattern has some diagonal stripes that I don't see on RGB. The stripes also vary in colors (purple/blue) which I don't see on RGB either, I guess those are NTSC artifacts.

S-video:
The diamond pattern is clearly visible and does not blend together at all. The diagonal stripes and the NTSC color artifacts are also gone.
Actually I see very little difference from RGB, I'm surprised S-video looks this sharp since I normally play using the RGB cable.


Bokujou Monogatari
CVBS:
This game uses a chess-pattern background in dialogue windows which is quite visible in the edges. They do blend together in the middle though and I see similar NTSC artifacts and diagonal stripe patterns on them as in SD2.

S-video:
Just like with SD2, the chess-pattern dithering is clearly visible and does not blend together. There are no diagonal stripes nor any NTSC artifacts. It looks pretty much the same as with RGB.


Jurassic Park (USA)
CVBS:
The dithering is less visible in this game, I guess thanks to the H-pseudo-512 mode. But it's still night and day compared to the hardware transparency, and the dithering is actually still kind of visible, although I can't distinguish a grill-pattern. NTSC artifacts and those diagonal stripes are also present.
I noticed that dithering is used a lot in this game overall, not only for transparency but also for things like shadows and darker areas in the grass and the bushes, in a similar way that hatching, cross-hatching and screentone techniques are used when doing dip-pen ink art. I think the artist was just very used to dithering techniques in general which may explain why they choose this for the semi-transparency.
It's a noteworthy example of a game that both uses dithering and the H-pseudo-512 mode.

S-video:
Again this looks a lot like RGB. The grill-pattern dithering is fully visible and the NTSC artifacts and diagonal stripes are gone, it looks pretty much the same as on that youtube video as far as I can tell.


Hoshi no Kirby 3
CVBS:
The first stage doesn't have much transparency so I tested on the second stage which has those pastell things in the parallax foreground made semi-transparent via dithering. Again the dithering is kinda visible although the pattern used can't really be distinguished. It has diagonal stripes and NTSC color artifacting as usual.
The graphics in this game is pretty cool and uses three BG layers: 1 parallax background, 1 midground where Kirby walks on and 1 foreground which uses the mentioned dithering. There is no HUD and the game instead uses a fully opaque status bar at the bottom of the screen, freeing up BG 3 for the foreground I guess.


S-video:
As usual this looks much like RGB. The grill-pattern dithering is now clearly visible and it looks pretty much like on emulators. Diagonal stripes and NTSC artifacting are both gone as usual.



Conclusion
It appears that the composite cable is required for the dithering in Jurassic Park and Kirby 3 to work at all, and S-video is surprisingly similar to RGB. Harvest Moon and Secret of Mana has a cruder dithering that is kinda visible even on composite. I guess things might be different on a non-1-CHIP SFC/SNES and a cheaper CRT TV. I might redo the test with my old PAL SNES (non-1-CHIP) and a cheap PAL CRT some time, though it won't have S-video, only composite and RGB (and I don't have PAL SNES S-video and RGB cables to test with anyway).

Dithered semi-transparency is still quite crude and looks much less like transparency than the colour math that is used in for example Fushigi no Dungeon 2: Fuurai no Shiren in dialogue window backgrounds and in the on-screen mini-map (which is overlaid the action on the middle of the screen). Dithering would probably not work very well at all in this game for the mini-map which uses quite fine lines. I would guess that it would just look like a stripy mess and be more annoying to the player than helpful. It seems to me that Fushigi no Dungeon 2 is a good example of a game that uses the SFC hardware in a way that wouldn't work well on the Megadrive or other systems without hardware transparency (like the Saturn).
Edit: I checked the first Fushigi no Dungeon game and it uses the same mini-map and dialogue boxes so that's another example.
User avatar
rainwarrior
Posts: 8732
Joined: Sun Jan 22, 2012 12:03 pm
Location: Canada
Contact:

Re: Does Jurassic Park use pseudo high-res mode?

Post by rainwarrior »

Pokun wrote: Mon Jun 27, 2022 4:05 pm
rainwarrior wrote: Mon Jun 27, 2022 11:02 am Kirby 3 could 100% be done with hardware blending.
But can you use colour math on two BG layers at the same time?
I'm not sure I understand what the suggestion is? You can't use BG layer colour-math with pseudo-hires, because the main and sub-screen are both spoken for via the hires option. That's why COLDATA is being suggested as something you could do with it. It's colour math without requiring its own layer. (Kirby 3 does not use COLDATA, as far as I can tell.)
Pokun wrote: Mon Jun 27, 2022 4:05 pmS-video: Actually I see very little difference from RGB, I'm surprised S-video looks this sharp since I normally play using the RGB cable.
Yes, I think in theory S-Video is very close to RGB in capability. Most detail comes from contrasts of luminance, and that is accomodated very well by the S-Video signal. The one thing that RGB can do better is sharp contrasts of chrominance instead, but this need is far less prominent than sharp luminance.

In an earlier thread about this there was a link to an imgur post (and reddit source info) with some side-by-side comparisons of composite, S-Video and RGB. In those examples I felt the best place to look was the Mega Man X buildings in the background, where the coloured vertical stripes in the buildings (e.g. yellow vs. blue) were noticeably less sharp on S-Video, but it was quite a subtle difference, IMO.

I think there are probably some older TV inputs that don't do a great job with S-Video, but I think a lot of the time it's nearly as good as RGB.

On my setup with Jurassic Park, on my CRT TV the pop-up overlay blend looks perfectly smooth. On my capture device I can see some prominent vertical grain to the blending effect. I think composite decoding will vary in effect quite a bit.

My composite capture:
jp_composite_capture.jpg
I've seen other people's JP capture have NTSC miscoloring artifacts (esp. around the "motion sensor" ellipses at the bottom of the screen), or uneven vertical grains that look like "jailbars". There's a lot of ways the composite process can treat this a little differently.
Pokun
Posts: 2681
Joined: Tue May 28, 2013 5:49 am
Location: Hokkaido, Japan

Re: Does Jurassic Park use pseudo high-res mode?

Post by Pokun »

rainwarrior wrote: Mon Jun 27, 2022 4:40 pm
Pokun wrote: Mon Jun 27, 2022 4:05 pm
rainwarrior wrote: Mon Jun 27, 2022 11:02 am Kirby 3 could 100% be done with hardware blending.
But can you use colour math on two BG layers at the same time?
I'm not sure I understand what the suggestion is? You can't use BG layer colour-math with pseudo-hires, because the main and sub-screen are both spoken for via the hires option. That's why COLDATA is being suggested as something you could do with it. It's colour math without requiring its own layer. (Kirby 3 does not use COLDATA, as far as I can tell.)
I'm basically asking if you can use colour math effects on more than one BG layer at the same time and without them affecting each other. Kirby 3 may have transparent water on the midground (which is BG 1 or 2 I guess) and also a transparent foreground with falling petals, bushes, rocks and other stuff (which I guess is BG 3) at the same time. I don't know if this is the case, nor have I checked if it actually uses two transparent BG layers at the same time.
I don't fully understand how COLDATA, COLOR Window, Fixed Color Addition/Subtraction, "colour math" (isn't this just another weird name for the fixed color add/sub thing?) and all that works. Nintendo's documents explains things very briefly, all over the place and with very weird English.

rainwarrior wrote: Mon Jun 27, 2022 4:40 pm
Pokun wrote: Mon Jun 27, 2022 4:05 pmS-video: Actually I see very little difference from RGB, I'm surprised S-video looks this sharp since I normally play using the RGB cable.
Yes, I think in theory S-Video is very close to RGB in capability...
......
I've seen other people's JP capture have NTSC miscoloring artifacts (esp. around the "motion sensor" ellipses at the bottom of the screen), or uneven vertical grains that look like "jailbars". There's a lot of ways the composite process can treat this a little differently.
I see, my PVM has something called an "NTSC comb filter" and a ton of settings in the service menu. Maybe that's part of it too.
It shall be interesting how it looks like on a cheaper PAL TV with my non-1-CHIP PAL SNES. The H-Pseudo-512 is supposedly not really working correctly on non-1-CHIP models due to their fringy video signals.

The conclusion I've made is that S-video is comparable to RGB but also, like RGB, no good for effects that are made for composite video in mind in most cases.
Examples of these that I've found includes:

Code: Select all

  -Dithering semi-transparency effects.
  -Certain systems like Apple II, Atari 8-bit and IBM-PC with a CGA video card
   that relies on NTSC artifacts to produce more colors than the video
   hardware normally can do.
  -MSX with a light-pen has a "cursor" that isn't displayed on the RGB output
   (might work on S-video though).
  -Early 3D consoles like PS1 and Nintendo 64 may rely on composite for fog
   effects. These may just look like dots instead of fog on sharper video
   signals.
  -PC-Engine disables color burst in its monochrome mode (used in Shapeshifter
   when pausing) which won't work on its RGB output.
   The composite artifact filter effects (that could potentially be used to
   make an interlaced picture) also doesn't work on RGB. The PC-Engine S-video
   mod might not have these problems though, I don't know.
User avatar
dougeff
Posts: 3079
Joined: Fri May 08, 2015 7:17 pm

Re: Does Jurassic Park use pseudo high-res mode?

Post by dougeff »

I'm basically asking if you can use colour math effects on more than one BG layer at the same time and without them affecting each other.
With regards to color math, consider there to be 2 layers.... the main screen and the sub screen. You can decide which BG layers (and sprites) go on each of these, but there is only one blend going on (usually addition with half for transparency).

You can also use windows to prevent color math from working on part of, or ALL of, a particular BG layer.

Optionally, if nothing is on the sub screen, or if you flip a specific control bit, only the fixed color will be used for the sub screen part of color math.
nesdoug.com -- blog/tutorial on programming for the NES
User avatar
rainwarrior
Posts: 8732
Joined: Sun Jan 22, 2012 12:03 pm
Location: Canada
Contact:

Re: Does Jurassic Park use pseudo high-res mode?

Post by rainwarrior »

Pokun wrote: Mon Jun 27, 2022 5:49 pmI'm basically asking if you can use colour math effects on more than one BG layer at the same time and without them affecting each other. Kirby 3 may have transparent water on the midground (which is BG 1 or 2 I guess) and also a transparent foreground with falling petals, bushes, rocks and other stuff (which I guess is BG 3) at the same time. I don't know if this is the case, nor have I checked if it actually uses two transparent BG layers at the same time.
If "same time" is on different scanlines, yes you can do it. You can basically rearrange the whole setup on every scanline. However, on the same scanline, you can't. There is only one blend step, so there isn't a way to get a second blend on top of the first one.

(Though, you can use a stack of layers as part of the blend, i.e. you could put a leaves layer on top of a water layer, and then blend them both with a background layer... just the leaves will be "part" of the water instead of blended separately with it, if that makes sense? The leaves would be transparently blended with the other background, but they wouldn't be blended with the blue of the water.)

There also isn't really an option to do a "second" blend with pseudo-hires, because hires mode actually takes over the blending hardware and replaces its function with the hires thing.

COLDATA gives another option. It replaces one of the two "screens" for blending with a solid colour. That's not super useful by itself, but if you change COLDATA per-scanline, it can make a nice gradient for things like a fading sky horizon. However, it uses hardware blending to do it, so it's not something you can do on top of another blend.

...but pseudo-hires makes an exception for it. Even though it's taking over the two screens normally used for blending, the COLDATA solid colour is still available and you can use it. So, at least under a composite output you can do the pseudo-hires "blend" at the same time as applying a COLDATA gradient.

It's a pretty minor ability, as there's many other ways to accomplish similar things, and we don't have any examples of any game actually doing this, as far as I know, but it's a marginal exception to the rule of only being able to do one blend step, as long as you can accept the pseudo-hires compromise.
iNCEPTIONAL

Re: Does Jurassic Park use pseudo high-res mode?

Post by iNCEPTIONAL »

dougeff wrote: Mon Jun 27, 2022 6:48 pm
I'm basically asking if you can use colour math effects on more than one BG layer at the same time and without them affecting each other.
With regards to color math, consider there to be 2 layers.... the main screen and the sub screen. You can decide which BG layers (and sprites) go on each of these, but there is only one blend going on (usually addition with half for transparency).

You can also use windows to prevent color math from working on part of, or ALL of, a particular BG layer.

Optionally, if nothing is on the sub screen, or if you flip a specific control bit, only the fixed color will be used for the sub screen part of color math.
It's strange because even though technically the SNES isn't supposed to be able to do semi-transparency on top of semi-transparency (so I'm constantly told), the effect in the following example achieves exactly that at a casual glance and actually looks really good too imo (which is all that really matters to most people just playing the game):

https://youtu.be/5WxHLB0Kc10?t=8399

There's moments there where it clearly looks like there's semi-transparent ghosts passing over semi-transparent ghosts on top of a semi-transparent background all at once, even is that is not what is actually happening on a purely technical level.

If that's the SNES not being able to have multiple overlapping semi-transparent effects then it's more than good enough for the kind of stuff I would like to do visually in my game with semi-transparency.
Zonomi
Posts: 69
Joined: Wed May 09, 2007 12:45 pm

Re: Does Jurassic Park use pseudo high-res mode?

Post by Zonomi »

rainwarrior wrote: Mon Jun 27, 2022 10:04 am I very explicitly included emulators in the point I was making, but yes, outside of them RGB is also something I'd call common now. It's not the most common hardware setup, but lots of people have RGB output with their SNES now, as opposed to 1992 where basically nobody did.
Well, everyone in France who owned a SNES could benefit from RGB output. :D
Post Reply