Commercial Critique - Pokemon Red/Blue

[Corinthian Baby]

I think we should be looking at these sprites in the initial GB 4 color greys first, and then looking about how we can apply 4 color preset pallets to them efficiently. Just to kind of see how we would do it vs. how they did it.

I think looking from the green sprites to the red/blue we're more familiar with, we can see what kind of criticisms they received and how they tried to improve them. One interesting thing about the sprites is because of the limited room, they used perspective and foreshortening to fit some of their designs in place. While it adds depth, it can also cause readability issues, so it's a double edged sword. I really think it works for Sycther and Hitmonlee. Others like Machamp and Hitmonchan look more confusing at first glance.

They also look much more zany and cartoony in the green sprites, the poses are more dynamic, aggressive and tensed in the red/blue ones. Maybe they studied how composition works best either directly before or after the action. That's why those perspective ones look kind of out of place because it's somewhat during the action.

Look at green version Kabuto, from a design perspective, pretty readable. Compared to red/blue Kabuto, a bit more confusing, especially if you don't have the comparison, and are just going from the sprite. I remember playing the original and thinking how weird he was, I thought those 2 dots on his shell were his eyes. So changing the sprites to be more battle ready also created new issues I geuss. But look how well it works for green Porygon to r/b Porygon. One is like testing the waters to get comfortable with the design, the other pushes it into the 3rd dimension.

For the record I still don't know what's going on with Trainer Cuball's crotch protector.

[PypeBros]

May I ask for some hint (excel-like "C3" coordinates, for instance) for identifying which picture you're referencing to ? I'm afraid pikachu, salamčche and bulbizar are the only one I could name...

[Ryumaru]

For reference, here are the pokemon yellow sprites that were created once the look and convention of pokemon was in place from the anime and other media. I think these sprites are quite beautiful and really handle the restrictions well; however, I can't help but feel there is a certain charm from the red/blue sprites that is lost in the more commercial and consistent yellow counterpart.



edit: The activity thread is up.
I'd especially like clarification on the palette swapping that the gameboy color could do with the original graphics. Can anyone shed more light on this? Almost every pokemon has it's own separate, relatively saturated palette, but the overworld is almost all a washed out, pastel couple of palette entries. Was there any reason the hardware couldn't separate from different assets such as trees or buildings? Was this perhaps just something that wasn't fully utilized by Gamefreak due to time or resource restraints?

Also I'm very interested in how the fonts work as well. Arne, you said that most often they only used 1 bit channel to save a bit on memory. do you think that it would have been possible to create fonts with some single layer AA to help smoothen and increase readability? This game is actually pretty text heavy so this is an asset that could surely use some sprucing up.

It may be beyond the scope of the activity, but it would be interesting to find out how much space the pokemon game took up from it's available cartridge. Perhaps at the least we could theorize of space saving techniques that would allow for some of the more "extravagant" things such as AAed fonts.

[Cyangmou]

I really would love to have a new commercial critique but I personally don't think that Pokemon Green/Red/Blue is a good subject for a commercial critique.
I don't want to be the kill-joy here but the things which are concerning me:

-Pokemon Green was the first version of the commercial most successful videogame series nowaday (there are in total 66 games, 23 games are from the main series). All first versions have some weaknesses which the developers usually try to avoid in the next versions.

-the style and art direction of a modern "classical" pokemon game looks and feels really similar to the first versions. Every new version "perfectionized" another aspect of the game while maintaining the roots of the series (graphics and gameplay).

-the graphic was functional, but the designs were good. The game definitely had a well thought-out style - there will always be smaller inconsistencies, but the normal player don't cares.
There were a lot of gameboy games which you could play through in about 2 hours. Pokemon managed to provide a world which took you dozens of hours to discover. The developers definitely sacrificed aesthetical value due to gameplay (I don't say that details could have been solved better, but they fixed them with the new versions anyways).
I played Pokemon Red as I was 5, but the graphics never feeled wrong to me - The cool part for me was always that you can catch, train and evolve cool monsters which are all different. I recognized a city as a city, a path as a path and a cave as a cave.

-(nearly) everybody has played one part of the series. So everybody has his own nostalgical memories of the game. The nostalgy factor usually kills the unprejudiced argumentation.

All in all I think everybody has made up his mind about the game so far, of course we can collect the opinions and try to "fix" stuff which is already fixed. There are a lot of games which changed from 2D to 3D like Metroid (I know that there are a few GBA titles) and where there would be place for quite some improvement (the newer 2D Castlevanias prooved it).
And there are lots of games without sequels and bad graphics which are also worth a look from the "critique" point of view.

A much more interesting attempt would be to look what the developers changed with the newer versions, especially graphics-wise. There were a lot of changes and a lot of changes to the good side. The last version I played was Ruby and  it was much better, although it looked and feeled still the same.
just some thoughts from my side

[Ryumaru]

I really would love to have a new commercial critique but I personally don't think that Pokemon Green/Red/Blue is a good subject for a commercial critique.
I don't want to be the kill-joy here but the things which are concerning me:

-Pokemon Green was the first version of the commercial most successful videogame series nowaday (there are in total 66 games, 23 games are from the main series). All first versions have some weaknesses which the developers usually try to avoid in the next versions.

-the style and art direction of a modern "classical" pokemon game looks and feels really similar to the first versions. Every new version "perfectionized" another aspect of the game while maintaining the roots of the series (graphics and gameplay).

-the graphic was functional, but the designs were good. The game definitely had a well thought-out style - there will always be smaller inconsistencies, but the normal player don't cares.
There were a lot of gameboy games which you could play through in about 2 hours. Pokemon managed to provide a world which took you dozens of hours to discover. The developers definitely sacrificed aesthetical value due to gameplay (I don't say that details could have been solved better, but they fixed them with the new versions anyways).
I played Pokemon Red as I was 5, but the graphics never feeled wrong to me - The cool part for me was always that you can catch, train and evolve cool monsters which are all different. I recognized a city as a city, a path as a path and a cave as a cave.

-(nearly) everybody has played one part of the series. So everybody has his own nostalgical memories of the game. The nostalgy factor usually kills the unprejudiced argumentation.

All in all I think everybody has made up his mind about the game so far, of course we can collect the opinions and try to "fix" stuff which is already fixed. There are a lot of games which changed from 2D to 3D like Metroid (I know that there are a few GBA titles) and where there would be place for quite some improvement (the newer 2D Castlevanias prooved it).
And there are lots of games without sequels and bad graphics which are also worth a look from the "critique" point of view.

A much more interesting attempt would be to look what the developers changed with the newer versions, especially graphics-wise. There were a lot of changes and a lot of changes to the good side. The last version I played was Ruby and  it was much better, although it looked and feeled still the same.
just some thoughts from my side

What you find as reasons that a critique is unnecessary or unfruitful, I find to be the best reasons to have a critique. the success of this game is huge, and it may very well have been to the expansive world that the red and blue versions gave ( comparative to other GB games as you say) but at the end of the day, that world is made up of 8x8 tiles, all of which can include 4 shades of grey; the beautiful thing about this is that the gameboy doesn't care how those shades are placed inside it which means that if we are using the same amount of tiles for an asset and can make it better in terms of pixel art technique as well as traditional art values and game art specific concerns, why not do it?

The charm and nostalgia that surrounds this game actually gives depth to the critique as there is a necessity to balance technical qualities with the more abstract qualities that are inherent in the less than " perfect" graphics.

It is not of our concern whether the graphics were "fixed"in next generations on the gameboy advance and the like ( ruby and sapphire actually have some pretty bad graphics too, employing many of the poor conventions this game established) . I think many of us agree that there are some things that could be fixed in this original generation with the hardware it had available that could lead to a more aesthetically pleasing game.

Critique is not necessarily about a a solution or an answer either. if, after the activity, we find that we've only moved laterally- we may realize that this game full of mismatched GUI, mixed resolution battle scenes, and sometimes deformed looking creatures, hit on some special nerve of perfection within imperfection. Finding that out through exploration of the game's graphics and attempting actual revamps is much more fulfilling I think, than just saying " nah, it was good enough for the time".

[Arne]

Pokémon Yellow is 1 megabyte and has tons of unused space between data segments from what I can tell. R/B are probably quite a bit smaller but I don't have them. Uncompressed 8*8px 2BP tiles take 16 bytes on both the NES and Gameboy (color data is stored elsewhere). A full 16*16 tile NES tile table thus takes 256 tiles * 16 bytes. Sometimes NES games have loose tile segments scattered about, and the GB games I've looked at don't seem to have full tile tables at all (in ROM). In PKMNY there are little segments with SS.ANNE and furniture and places. These go into the video memory when needed, alongside common/shared tiles no doubt.

As for the 1 bitplane (b/w) stuff. Here's the font right under the Japanese logo. You'd think it has anti-aliasing judging by the 2BP mode view, left, but it looked suspiciously wonky so I wrote a 1BP view mode which revealed the truth. There can be no anti-aliasing in 1bp mode. (The gfx underneath appears garbled because it's on a different offset multiple.)



It's a fixed width font of course. You can't easily have kerning on these old tile based systems (and you need annoying extra code for stuff like end of line detection).

On the NES, 1bp stuff tend to look overlayed and not squished, like this:



If I recall correctly, the monster designs were established fairly early on, it's just that the R/B pixel art was a little off. Meaning, they didn't "find" the designs with the later games. They just did better pixel representations (probably pretty taxing to do graphics for 150 monsters the first time around with a smaller team).

Conceit> I like the designs to various degrees, and sometimes I just like the one aspect of a design but not another, or I like how a design looks but not how it relates to other evolutions. I've never liked Mankey and Primeape much. Primeape feels like an unnecessary evolution. But in light of the newer monkey Pokémon I think Mankey is a fairly memorable design somewhere at heart since it has the quirky ball-body going on and is not just a cartoon ape with stuff on it. Anyways, it's really hard for me to pick a specific set of 30. It depends on what the filter is... Cute, Cool, interesting double function of details, memorable, nice evolution set, fun-ugly, great familiar made unfamiliar combo.

[Ryumaru]

If pokemon yellow had extra space then there must have been the ability to make room for the full res back sprites the game so desperately could have used.

Do you know how text and the GUI elements are treated in game? I think I remember reading they were on a special priority level or something of the sort. Was there a special mode or function these assets used that made multiple colors for anti aliasing impossible?

If so, could it perhaps have been worked around with tiles ( seems especially doable with portions of the battle screen that stay static)?

[Corinthian Baby]

Not quite as related to memory space, but I was looking over the tiles and I think they are fairly succinct and readable for a GB game. I know that when I played as a kid, I didn't really have to think, "what is that?" Of course there are the infamous barrels everywhere that literally could have been anything else, but for the most part the buildings are recognizable, the grasss, the trees look like trees. It's just the crazy pallets being used to render if you played on GBC (Which I did). Everything was either all red or all green, I remember that, and I think the GB chose more red pallet stuff because I was playing "red version", and same for blue and blue.

Although as time went on and I got to the GBA games, I thought to myself "I could make this better, I would change this map, etc." By the time it got to the DS games, I was like "good god get me some AA". And there are a lot of graphical changes I would make to those, but it also is psedo 3D in that it changes/shifts the view to stay within perspective instead of being fixed. But those are critiques for another time, perhaps.

[Arne]

You can see the dialog box panel border with the pokéballs in my dump above. It's in 2bp mode, but was kept B/W because it was more graphical. I think the 1 bitplane stuff is done to save game data space, not video memory space. I'm guessing anything which goes into video memory is translated to 2bp but I could be wrong. SML had 4 color/shade fonts (just regular tiles). Personally I often prefer crisp text without AA on these old systems.

I don't mind the 2x backsprites much. The extra line thickness sort of works, puts the figure on a different plane than the opponent.

I think the backsprites were compressed along with the front sprites and can't easily be expanded to normal rez. And if they could, there would have to be room in the video memory tile table as well.. a lot more room since the images would be 4 times larger to cover the same area. 4*4 to 8*8? This would take... 48 tiles away from the other stuff which needs to be there, such as font, gui, opponent, battle fx... The GB had room for 256 tiles at once but could apparently switch banks mid screen and get another 128 out of it (move half a tile table down). Perhaps this is why the Pokémon are placed at the upper and lower half? Or maybe bank switch is made under both monsters to get lower case font etc. Looking at the tile tables in an emulator might answer this. I'm thinking since there's text and stuff at either halves it would have to exist in both banks, so you won't get that much extra space after all (I think you use the 128 tile overlap for shared tiles). Would be extra difficult to also squeeze in your grass tiles, but I might be wrong, it's not entirely uncommon to see some wasted space.

You can do a test, make a 16*16 tile canvas, plop in all the elements which might show at once, see if it fits. I'm thinking it could be done with some space over (Rough calculation in head gives me as much as 50-100 tiles to spare but maybe I'm missing something). Some tiles could be streamed in when needed, such as new battle effects (bubble beam and whatever). Trainer and monster never appears at the same time iirc. The HP bar and arrows can be seen in my dump above. They used tiles for the bar since drawing rects is a no-no.

Hmm, I just realized I developed a tool which automates tile counting a while back (2007). It takes an image, compares 8px tiles to find unique ones. Calculates visual similarities and number of copies (useful when optimizing tile count), and even makes a game map (since it now recognizes the tiles.). If anyone's interested I could upload it. It can count the number of tiles used in an image easily.
Balding's Quest tilemap tool

[Ryumaru]

If anyone here has access to an emulator that can look at the tile tables and would like to share, that would be great! Unfortunately KiGB on a mac does not have that function. It would be interesting to get a closer look on how things are implemented and it might better inform our critique and the activity.

Arne: I think you're already talking a bit over my head as I didn't know there would be a difference between game data memory and video memory. Is there a place online you know of that has more information about what you're speaking of? Preferably gameboy specific.
Would all 151 pokemon backsprites have to fit on these tile tables?

I'm sure many would find use in your tile counter tool. Feel free to upload it :]