Massive Console History update
Scott H
Pilgrimage ( www.gamepilgrimage.com/consolehistory.htm ). All of the
console spec sheets have had significant amounts of data added to them
(for you nitpickers out there) and the redesigned formatting *should*
make the specs more easily read and understood than they were with less
than half the info listed before.
Comments and suggestions (better sources and more detailed information)
are welcome.
I've attempted to balance the facts with what is perceivable, and
compare that to popular/media opinion. If a statement seems abrupt,
it's because it is in answer to another equally abrupt popular
sentiment. In other words, each comment has a context. I have not
attempted to be exhaustive, but have attempted to summarize the facts
and present each console in light of all other consoles. Gamecube and
Xbox sections are in the works.
Significant additions to each section are as follows:
All sections now have sound information and all RAM (some specs missing
cache info for individual processors).
All sections now have storage capacity for software mediums used, and
the average size of carts.
The Saturn and PS1 sections have nearly double the information,
including the full span of resolutions supported (which apparently were
*all* used in actual software), processor marketing performance
information and other more system specific info.
The Dreamcast and PS2 section were updated in a similar fashion to the
Saturn and PS1 sections, and a quote was added which deals with
(hopefully conclusively) the graphical performance strengths and
weaknesses of each platform.
barrywa...@yahoo.com
First, you mention the max pps count of PS1 as being 180,00. While this
is true as far as official spec sheets are concerned, it was surpassed
in game according to the developers of "Iron and Blood", a fighting
game based on the dungeons and dragons universe. On the back of the
case, it mentions the pps rate as being 225,000. Furthermore, in an old
Gamefan magazine interview, I recall them stating that each character
was composed of around 3,700 gouraud shaded polygons, which at 30
frames comes to about 225,000 pps. On a side note, they also discussed
the fabled M2 version, which according to the developers, would feature
models built with over 10,000 polygons! So I think the fact that PS1
surpassed the original theoretical limit should be mentioned.
My second suggestion concerns your Snes history page. You fail to
mention one of the Snes's most significant hardware features.
Transparency. It was a distinct advantage that the Snes had over
Genesis (which had no hardware support for it). Even the much vaunted
Neo Geo lacked this feature in hardware. Like the Genesis, it had to be
faked with "meshes" on Neo Geo.
Tons of Snes games used it to create all kinds of neat special effects
(especially water effects). Games such as Chrono Trigger, Final
Fantasy, Secret of Mana, Killer Instinct, Doky Kong Country, and even
unpopular titles like Indiana Jones Trilogy, all used it to great
extent. One game in particular, EarthWorm Jim, had beautiful lens flare
on the first level that was produced using this effect. But it was
completely absent from the Genesis version. This feature was also used
to artificially increase color counts and simulate light sourcing
(example, Fulgore's stage in Killer Instinct. There is a swinging lamp
with a translucent sprite that "lights" up fighters and projectiles). I
think this should be mentioned.
Third, corcerns your claim that Sonic 2 has 7 or more background
planes. I don't think this is accurate. I remember reading somewhere
that most of the "planes" in Sonic were created using raster effects,
but I can't remember where I read it unfortunately. However, using a
feature in Snes9x that allows you to turn on or off the four snes
background planes, I think I found information that supports the idea
of Sonic 2 not being a background scrolling beast, at least
technically, though visually it is. In Mortal Kombat 1 for Snes, there
appears to be 5 background layers in the first stage, which exceeds the
official specs, BUT, after turning on and off the four background
layers with the 1,2,3,4 keys, I discovered that it is using only three,
with the third being the health bar and statistics. There also appears
to be, at least visually, 5 background layers in the Pit stage. Three
of these scrolling "backgrounds" are cloud layers, and by turning off
background 2, all three disappear. This points to some other 16-Bit
hardware effect that's capable of artificially increasing the visual
parallax scroll. Perhaps the raster hardware is responsible for this
neat effect.
Scott H
> Well, I have a few suggestions.
>
> First, you mention the max pps count of PS1 as being 180,00. While this
> is true as far as official spec sheets are concerned, it was surpassed
> in game according to the developers of "Iron and Blood", a fighting
> game based on the dungeons and dragons universe. On the back of the
> case, it mentions the pps rate as being 225,000. Furthermore, in an old
> Gamefan magazine interview, I recall them stating that each character
> was composed of around 3,700 gouraud shaded polygons, which at 30
> frames comes to about 225,000 pps. On a side note, they also discussed
> the fabled M2 version, which according to the developers, would feature
> models built with over 10,000 polygons! So I think the fact that PS1
> surpassed the original theoretical limit should be mentioned.
Okay, I'll add that this game surpassed the (revised) official specs.
> My second suggestion concerns your Snes history page. You fail to
> mention one of the Snes's most significant hardware features.
> Transparency. It was a distinct advantage that the Snes had over
> Genesis (which had no hardware support for it). Even the much vaunted
> Neo Geo lacked this feature in hardware. Like the Genesis, it had to be
> faked with "meshes" on Neo Geo.
Good point, I completely forgot about this. I think the *only*
Genesis game to do transparency in software was Sonic & Knuckles, though
even that might be some graphical illusion that just doesn't produce any
artifacts like dithering does.
This is the same as the TG16's capability to do more than 1 background.
I'm very certain that it wasn't "forcing" more actual backgrounds to
scroll, but the point was that both the TG16 and the Genesis were able
to do more parallax (atleast for all intents and purposes) than their
specs indicate, while the Snes, as far as I had seen had not. Whatever
software effects it might be using, abundance of parallax planes in
Lightening Force and certain levels of Sonic 2. According to this
though, you've seen the Snes scroll 5 planes in MKI, so that's at least
one game to go over the limit for the Snes, at least visually.
--
Scott
Scott H
>
> This is the same as the TG16's capability to do more than 1 background.
> I'm very certain that it wasn't "forcing" more actual backgrounds to
> scroll, but the point was that both the TG16 and the Genesis were able
> to do more parallax (atleast for all intents and purposes) than their
> specs indicate, while the Snes, as far as I had seen had not. Whatever
> software effects it might be using, abundance of parallax planes in
> Lightening Force and certain levels of Sonic 2. According to this
> though, you've seen the Snes scroll 5 planes in MKI, so that's at least
> one game to go over the limit for the Snes, at least visually.
>
>
There should have been a "there is an" in there before abundance I
think. ;)
--
Scott
barrywa...@yahoo.com
scroll"
OK. It just sounded like you were saying that the Genesis could
actually do more than 2 true background layers technically which I
don't think is the case. More likely it was faking it visually with
some hardware or software feature.
"According to this though, you've seen the Snes scroll 5 planes in MKI,
so that's at least
one game to go over the limit for the Snes,"
Yeah, visually, there are clearly 5 independent scrolling background
"planes" in the first MKI level (not counting the health bar which
would bring it to six), but in actuality it's only using TWO of the
background planes for the main stuff, and one for the health bar. The
fourth is not even used! Somehow the "planes" are being faked, but it
doesn't really matter since the end result is still five background
layers visually.
Scott H
> "I'm very certain that it wasn't "forcing" more actual backgrounds to
> scroll"
>
> OK. It just sounded like you were saying that the Genesis could
> actually do more than 2 true background layers technically which I
> don't think is the case. More likely it was faking it visually with
> some hardware or software feature.
That's the distinction I hoped to achieve by first listing the hardware
limit, and then the qualifier "in software" after the higher number
achieved in game.
>
> "According to this though, you've seen the Snes scroll 5 planes in MKI,
> so that's at least
> one game to go over the limit for the Snes,"
>
> Yeah, visually, there are clearly 5 independent scrolling background
> "planes" in the first MKI level (not counting the health bar which
> would bring it to six), but in actuality it's only using TWO of the
> background planes for the main stuff, and one for the health bar. The
> fourth is not even used! Somehow the "planes" are being faked, but it
> doesn't really matter since the end result is still five background
> layers visually.
>
I don't think it matters as long as the layers are clearly
separate from each other. For example, the first level of Sonic 2 has
those yellow dots with a green background that scroll in a smooth way,
like the floors of SFII levels do, as does the water in Sonic 1. I
don't think that this effect is the same as actual parallax. I've also
seen raster effects (animated palette scrolling) used on many occasions
(such as the pirate ship in TMNTIV) to make it hard to count how many
actual layers there are. I count this as an entirely different effect
as something like Shadow of the Beast on TG16 and Genesis, or the
aforementioned Lightening Force 1st level, the cloud's in a later Sonic
2 level, or the parallax in Shinobi III. In other words, trying to play
tricks on the eyes, and actually fooling hardware into displaying more
independently scrolling planes are two different things, as are "line
scrolling" planes. On a side-side note, it is also typically considered
more hardware intensive to achieve an effect in software (even with
"tricks") than the same effect rendered using a hardware feature. Using
this principle, the TG16 and Genesis displaying more parallax (faked or
no) on screen than the Snes means that they are actually demonstrating
far more technical muscle in doing so, because it is being done in software.
Also on that note, I've noticed that a great deal of PS1 games
with lighting (presumably gouraud shading) also use a noticeable amount
of non-textured polygons. Dead or Alive and Tobal no. 1 are the most
obvious examples. I would suspect that using a hardware supported
effect like "lighting" on a number of non-textured polygons on screen
would be somewhat of a middle ground between flat shaded polygons and
texture mapped and gouraud shaded polygons. Meaning that if Iron and
Blood doesn't have texture mapping, or in some other way lessens the
memory required for texture mapping (lower resolution or bit depth) then
it is not achieving a higher polygon count, but is simply tapping into
the "flat shaded" polygon figure while still texture mapping some things.
Similarly, the Tekken games use flat floors (not rings) and flat
backgrounds. Texture mapped polygons are used for the floors and the
backgrounds, as well as the characters. Fighter's Megamix and Fighting
Vipers use a similar graphical style while using 2D floors and
backgrounds. This means that the Tekken game's polygon counts would
include more than just the characters while the Saturn fighter's polygon
counts are limited to just the fighters (and the 2D effects naturally
take up VRAM along with the 3D). Since fighters in FMM and FV look much
more rounded than even characters in Bloody Roar 2 and Tekken 3, I would
say that the use of 2D grounds and backgrounds was definitely preferable
for the Saturn, if it was going to also do lighting effects. In other
words, both systems were similarly powerful in two very different
approaches to 3D graphics. A flat polygonal plane is just as "2D" as a
2D plane being scaled and rotated in sync with polygonal characters.
The reason I bring this up is to draw out the actual point, slant and
bias of my website. That is, exactly what I state on the front page:
"Knowing the strengths and weaknesses in hardware and software of each
console can help you to know which console is likely to have more games
that suit your tastes. Knowing that your tastes are not the end all, be
all to what is and isn't a good game is a virtue."
--
Scott
Scott H
> I've completed a virtual redesign of the console history page on Game
> Pilgrimage ( www.gamepilgrimage.com/consolehistory.htm ). All of the
> console spec sheets have had significant amounts of data added to them
> (for you nitpickers out there) and the redesigned formatting *should*
> make the specs more easily read and understood than they were with less
> than half the info listed before.
>
> Comments and suggestions (better sources and more detailed information)
> are welcome.
> I've attempted to balance the facts with what is perceivable, and
> compare that to popular/media opinion. If a statement seems abrupt,
> it's because it is in answer to another equally abrupt popular
> sentiment. In other words, each comment has a context. I have not
> attempted to be exhaustive, but have attempted to summarize the facts
> and present each console in light of all other consoles. Gamecube and
> Xbox sections are in the works.
Correction, Gamecube and Xbox sections are up.
--
Scott
barrywa...@yahoo.com
with lighting (presumably gouraud shading) also use a noticeable amount
of non-textured polygons."
True for some games that use both light sourcing, transparency and
gouraud shading in 512x480. Below that, almost all PS1 light sourced
games are fully texture mapped (well the ones I've played, and that's a
lot). When you have those three VRAM eating features in one scene at a
high resolution like 512x480, there's not going to be a great deal of
space left for textures. So the Tobal's had to make due with a low
amount of texture detail, but I think they made up for it with those
effects and the addition of complex 3D backgrounds. Having said that,
remix DOA is for the most part texture mapped. Just look at that pick I
sent you of that black guy Zack. He's texture mapped everywhere, even
the SHOES!. True, the woman (Kasumi?) he's fighting has less, but
that's because she's supposed to. There's no need to put harsh muscles
textures on female bodies because their bodies are typcially curvy and
muscleless. Also, light sourcing is not the same as gouraud shading,
but it seems like you're suggesting that it is. These are separate
effects.
Light sourcing changes the color tone of the polygons depending on its
distance and angle to the preset light source, which is can be seen in
games like Tekken 2 where the color tone of the polygons are constantly
changing as the fighters move around the ring. Gouruad shading is the
process of blending the colors at the intersections of polygons. Tobal
2 and Soul Blade use this feature extensively, blurring the
intersection lines and giving the characters a smoother appearance.
Tekken 2, however, does not employ gouraud shading, and this is why the
characters have a more "polygon" look than the models in say, Soul
Blade. But I would say that the polygon count (models) is roughly eual
between those two.
"I would suspect that using a hardware supported
effect like "lighting" on a number of non-textured polygons on screen
would be somewhat of a middle ground between flat shaded polygons and
texture mapped and gouraud shaded polygons. Meaning that if Iron and
Blood doesn't have texture mapping, or in some other way lessens the
memory required for texture mapping (lower resolution or bit depth)
then
it is not achieving a higher polygon count, but is simply tapping into
the "flat shaded" polygon figure while still texture mapping some
things."
Yeah, but Iron & Blood is heavily texture mapped and gouraud shaded,
and I would say it's running a Tekken or Toshinden type resolution. It
looks very clear on my real PS1. But I can't get the ripped contents to
run in ePSXe, so I can't be 100% certain of the res, but I'm 99%.
There's also LOTS of color in the game. A very detailed look. Way
beyond DOA's color count.
"Similarly, the Tekken games use flat floors (not rings) and flat
backgrounds. Texture mapped polygons are used for the floors and the
backgrounds, as well as the characters. Fighter's Megamix and Fighting
Vipers use a similar graphical style while using 2D floors and
backgrounds. This means that the Tekken game's polygon counts would
include more than just the characters while the Saturn fighter's
polygon
counts are limited to just the fighters (and the 2D effects naturally
take up VRAM along with the 3D). Since fighters in FMM and FV look
much
more rounded than even characters in Bloody Roar 2 and Tekken 3,"
The models in FV don't appear more rounded because of more polygons,
but because that game uses gouraud shading which helps to conceal the
polygon intersections. Hell, even the models of the original Toshinden
look more rounded thanks to the addition of gouraud shading, but
officially, they're only using 800 per model whereas BR2 uses *at
least* a thousand average, most likely more. BR2 just uses little
gouraud shading in comparison to these games, resulting in more "lego"
looking models. If you recall, VF2 in the arcade lacked gouraud shading
too, and as a result, had the same "lego" look that even the Saturn
version didn't have. This was because Saturn VF2 had no lighting unlike
the arcade version, but if you turn it on in the PC version, the Saturn
models will begin show their polygon intersections. But the amount of
intersections are noticeably less (look at the shoulders) than either
the Tekken 2, VF2 or the BR2 models, demonstrating a lower polygon
count for Saturn VF2.
But the addition of heavy gouraud shading is an advantage that FV has
over BR2. Even the arcade version didn't have it. Model 2 didn't
support it in hardware.
"I would say that the use of 2D grounds and backgrounds was definitely
preferable
for the Saturn."
It was definitely an advantage. Constructing floors with polygons
certainly ate up more clock cycles on the PS1 than if it had used a
warped bitmap.
barrywa...@yahoo.com
game is 98% texture mapped, with the exception of female characters
whose legs and arms lack the texture quality of the male characters,
but as I said before that's because they're supposed to. Women have
curves, not edgs and muscles. Male characters are textured almost
everywhere, even the bottom of their shoes have textures. But the shots
on your comparison page don't do them justice, especially the Jet-Lee
comparison in which the PS1 pic has intense light sourcing that darkens
and hides the textures. Plus, he's not even wearing the same costume as
the Saturn Jet-Lee. In the blue pants costume, the PS1 Jet-Lee has
nearly as much texture maps as the Saturn version. Biceps, abdomens,
pecks, etc are all represented with textures. Overall, it's comparable
to Saturn's texture quality IMO.
bwahhhh
@f14g2000cwb.googlegroups.com:
The lack of texture mapping on character skin (at least the girls) in
DOA1 on the PS1 was a design choice, I think, I agree with you. I don't
think it was because the PS1 couldn't do or anything like that. It
wasn't done on the Saturn (or Arcade) because the gouraud shading
couldn't be done on their hardware, (or in the Saturn's case, at that
resolution) so both the Saturn and Arcade versions have the skin heavily
textured because flat shading just wouldn't look right.
The reason I think this is the case is because DOA2 on the Dreamcast did
not have that heavily textured look like DOA1 Saturn/Arcade did, but
instead followed the path of DOA1 on the PS1, gouraud shading the skin
and giving the girls that 3D anime Barbie doll look that Itagaki seems to
like. (even DOA4 on the Xbox 360 does the same) When I look at DOA2 and
3 they look more like DOA1 on the PS1 aesthetically than DOA1 on the
Saturn/Arcade, at least to me, leading me to believe that they
specifically wanted it to look that way on the PS1.
Scott H
> Also on that note, I've noticed that a great deal of PS1 games
> with lighting (presumably gouraud shading) also use a noticeable amount
>
> of non-textured polygons."
>
> True for some games that use both light sourcing, transparency and
> gouraud shading in 512x480. Below that, almost all PS1 light sourced
> games are fully texture mapped (well the ones I've played, and that's a
> lot). When you have those three VRAM eating features in one scene at a
> high resolution like 512x480, there's not going to be a great deal of
> space left for textures. So the Tobal's had to make due with a low
> amount of texture detail, but I think they made up for it with those
> effects and the addition of complex 3D backgrounds.
Well if we're talking about aesthetics then the only answer is
that all of these approaches were stop gaps to the current generation of
rendering, as all of them have been replaced with more complex (and
realistic looking) rendering techniques. What I was pointing out was
only that if the PS1 fighter wasn't texture mapping any polygons on the
screen, then the polygon count would be higher than if everything was
equally textured.
> Having said that,
> remix DOA is for the most part texture mapped. Just look at that pick I
> sent you of that black guy Zack. He's texture mapped everywhere, even
> the SHOES!. True, the woman (Kasumi?) he's fighting has less, but
> that's because she's supposed to. There's no need to put harsh muscles
> textures on female bodies because their bodies are typcially curvy and
> muscleless.
The effect works, but the textured Saturn version has far more fine
detail from the higher resolution texture mapping (uniform across all
characters), and allows for more close-ups as a result. In addition,
the Saturn version is higher resolution, and *if* it has higher
animation is doing more still that indicates to me that both systems
were being pushed in their own unique ways.
http://www.gamepilgrimage.com/DOAc8.htm
> Also, light sourcing is not the same as gouraud shading,
> but it seems like you're suggesting that it is. These are separate
> effects.
> Light sourcing changes the color tone of the polygons depending on its
> distance and angle to the preset light source, which is can be seen in
> games like Tekken 2 where the color tone of the polygons are constantly
> changing as the fighters move around the ring. Gouruad shading is the
> process of blending the colors at the intersections of polygons. Tobal
> 2 and Soul Blade use this feature extensively, blurring the
> intersection lines and giving the characters a smoother appearance.
> Tekken 2, however, does not employ gouraud shading, and this is why the
> characters have a more "polygon" look than the models in say, Soul
> Blade. But I would say that the polygon count (models) is roughly eual
> between those two.
Gouraud shading has the effect of blending edges, however in this
generation it was used to simulate lighting, because the systems could
not do "lights" like 3D accelerated hardware started doing later. If
there was some other technique in addition to Gouraud shading and flat
shading used on these systems I've found no description of how it works.
Thus have assumed that Gouraud or flat shading are being used in all
cases to simulate lighting's effects. So basically, if the game is
using what appears to be a "light" to shade characters or scenery, it is
using either Gouraud or flat shading to achieve the effect, if you can
see polygonal facets caused by the effect (
http://www.gamepilgrimage.com/VF2PClight.htm ) then it's flat shading,
if the polygons look more rounded
(http://www.gamepilgrimage.com/FMM004.htm ) then it's Gouraud. Again,
if there is an additional effect being used, I'd like to know how it
works and where it's listed in regards to these systems. I think the
assumption has been that because later 3D hardware uses "lights" as a
source of the effect, that these were as well.
http://en.wikipedia.org/wiki/Gouraud_shading
http://en.wikipedia.org/wiki/Flat_shading
In addition to this, I found the following comment on the differences
between how the Saturn and PS1 handle Gouraud shading. Apparently the
Saturn uses additive and the PS1 multiplicative. The former is less
processor intensive and not preferable as it causes a higher amount of
saturation in the effect, almost completely wiping out the underlying
colors (seen in Powerslave, and Quake, though Duke Nukem seems to handle
it better).
Saturn uses additive lighting, PS1 multiplicative.
http://www.shinforce.com/saturn/information/3D-Capabilities.htm
Additive and Multiplicative
http://groups.google.com/group/3dfx.glide/msg/991092ecc25236c0?hl=en&
>
> "I would suspect that using a hardware supported
> effect like "lighting" on a number of non-textured polygons on screen
> would be somewhat of a middle ground between flat shaded polygons and
> texture mapped and gouraud shaded polygons. Meaning that if Iron and
> Blood doesn't have texture mapping, or in some other way lessens the
> memory required for texture mapping (lower resolution or bit depth)
> then
> it is not achieving a higher polygon count, but is simply tapping into
> the "flat shaded" polygon figure while still texture mapping some
> things."
>
> Yeah, but Iron & Blood is heavily texture mapped and gouraud shaded,
> and I would say it's running a Tekken or Toshinden type resolution. It
> looks very clear on my real PS1. But I can't get the ripped contents to
> run in ePSXe, so I can't be 100% certain of the res, but I'm 99%.
> There's also LOTS of color in the game. A very detailed look. Way
> beyond DOA's color count.
I wish that I had ever seen the game in action. The tiny
screenshots on Gamespot do seem to reveal that the characters are
textured. I wonder if there's a connection to games like Iron and Blood
not being good games and their technical prowess. I also wonder how
much memory VF2's AI routine takes up.
I think that the Tekken games and BR games are pushing more polygons in
the characters than any Saturn fighter. I also think that FV and FMM
are just plain ugly, and I can't decide if it's just ultra low
resolution, or if the panning camera combined with Gouraud shading is
causing the characters to look so cruddy. With that said, take a look
at FMM's Dural, check out the facets in the leg, arms, torso and face,
that model is not pushing a low polygon count for this generation.
>
> But the addition of heavy gouraud shading is an advantage that FV has
> over BR2. Even the arcade version didn't have it. Model 2 didn't
> support it in hardware.
I'd personally take the higher resolution of the arcade version
over the Gouraud shading. I much prefer the fact that you can see fine
detail in VF2 and DOA for Saturn than that characters appear rounded and
can a black back and white front in FV and FMM.
> "I would say that the use of 2D grounds and backgrounds was definitely
> preferable
> for the Saturn."
>
> It was definitely an advantage. Constructing floors with polygons
> certainly ate up more clock cycles on the PS1 than if it had used a
> warped bitmap.
>
I wouldn't go so far as to say the approach was superior to the
PS1's approach, just that it was preferable for the Saturn hardware to
do so, and produced unique effects and results.
--
Scott
barrywa...@yahoo.com
between how the Saturn and PS1 handle Gouraud shading. Apparently the
Saturn uses additive and the PS1 multiplicative. The former is less
processor intensive and not preferable as it causes a higher amount of
saturation in the effect, almost completely wiping out the underlying
colors (seen in Powerslave, and Quake, though Duke Nukem seems to
handle
it better). "
That's all very interesting. I've never heard of gouraud shading being
to simulate light sourcing before. I had always assumed that PS1 style
light sourcing was completely separate form gouraud shading or flat
shading. I stand corrected.
Anyways, I went to that Shinforce link, and found something that I take
issue with. It's his comment about polygon seaming being a flaw unique
to the PS1 hardware. This is very false. I don't know about the Saturn,
but the N64 had it in many games. The Turoks for example had seaming in
many places. Some levels in Mario 64 were plagued with it. And it
wasn't limited to that generation either. Seaming appears often in the
current generation consoles, even in top of the line games like Doom 3
for the XBOX. In fact, the seaming in Doom 3 is terrible, almost as bad
as Tomb Raider on the PS1! Here is a game that employs so many advanced
features to achieve an ultra-realistic look, but has a horrible amount
of seaming that nearly destroys the illusion. I really hope the next
generation of consoles are able to eradicate once and for all.
"I wish that I had ever seen the game in action. The tiny
screenshots on Gamespot do seem to reveal that the characters are
textured. I wonder if there's a connection to games like Iron and Blood
not being good games and their technical prowess."
In this case, yeah. They focused so much on the graphics that they
forgot about the gameplay. :) Seriously though, it's not THAT bad. It's
just a tad below Toshinden, on par with Criticom. Ok, it is that bad.
But you won't be playing it much anyways. You'll be doing more staring
at the loading screen than actually playing. The load times are
painfully long.
"The effect works, but the textured Saturn version has far more fine
detail from the higher resolution texture mapping (uniform across all
characters), and allows for more close-ups as a result."
It's more detailed on the Saturn, but I don't think the difference is
really that big. There is plenty of fine detail on Zack for instance.
Look at his bicep and abdomen textures. That looks as good as Tekken 2.
"In addition,
the Saturn version is higher resolution,"
That's debatable. Has any authorative source stated its exact
resolution? Such as one of the game's programmers? I'm also skeptical
of VF2's alleged 704x480 resolution. I do recall a Sega rep saying it
ran in that res, but that was just a rep. I don't think he worked on
the game. He may have just assumed it. And in none of the interviews
with AM2 (the ones I've read) did I ever see them state the exact
resolution. Just that it was hi-res. I would think that such a
resolution would be counter-productive since as far as I know both the
Saturn and Playstation were limited to 8-Bit color in 640x480 and up.
In old PS1 spec sheets, they gave the maximum color depth of each
resolution, and 640x480 was limited to 256 colors. Regarding the
Saturn, there was an interview (which is still archived on the net I
think) with the designers of the Saturn net-link browser, and according
to them, 640x480, which was the res that the browser ran in, was
limited to 8-bit color depth. Now VF2 looks like it's showing a heck of
a lot more colors than 256. I can't see any obvious color gradients in
the distant bitmap for example. VF2 gives the impression of lots of
color. If it really is using only 256 colors, then it's using them very
well.
"and *if* it has higher
animation is doing more still that indicates to me that both systems
were being pushed in their own unique ways."
Saturn is doing more in the backgrounds (if the XBOX version is an
accurate port). They look much, much better. The PS1 has probably the
worst backgrounds of any PS1 fighter I have ever seen. They're so dull.
But regarding the models, I have to give the edge to the PS1, at least
as far as "doing more". You may prefer the higher texture detail
though. The PS1 has full gouraud shading, simulated light sourcing, and
polygon skin meshing (it does, I recently checked, it's just not as
good as Tekken 3). The Saturn just has higher texture detail, but I
don't see a very huge difference on the male characters. It's not like
it's night and day. As far as polygon count is concerned, I would say
that they're roughly equal. Neither appear to be pushing as many
polygons per model as Tekken 2. Despite the heavy gouraud shading, I
can still occasionally see the polygons facets (yeah, that's the
correct word) on the shoulders of certain characters, and they clearly
lack the complexity of the models in Tekken 2. Both games are probably
using a Saturn VF2 style polygon count. Maybe a little better.
"I'd personally take the higher resolution of the arcade version
over the Gouraud shading. I much prefer the fact that you can see fine
detail in VF2 and DOA for Saturn than that characters appear rounded
and
can a black back and white front in FV and FMM."
Yeah, most people were of that opinion when it was first released. If
it's running in the same res as Virtual On, then it's pretty low. VO
has that same muddy look that PS1 Ridge Racer has, even RR4. The arcade
version wasn't that much of a looker anyways. One of the least
impressive Model 2 games IMO. Back in the day, my mall arcade had a VF2
machine right next to a FV machine, and I remember the difference in
detail being big.
Scott H
> "In addition to this, I found the following comment on the differences
> between how the Saturn and PS1 handle Gouraud shading. Apparently the
> Saturn uses additive and the PS1 multiplicative. The former is less
> processor intensive and not preferable as it causes a higher amount of
> saturation in the effect, almost completely wiping out the underlying
> colors (seen in Powerslave, and Quake, though Duke Nukem seems to
> handle
> it better). "
>
> That's all very interesting. I've never heard of gouraud shading being
> to simulate light sourcing before. I had always assumed that PS1 style
> light sourcing was completely separate form gouraud shading or flat
> shading. I stand corrected.
I think it's more accurate to say that there just isn't
information to the contrary, since popular thought was that lighting of
some sort was used with or without gouraud shading.
> Anyways, I went to that Shinforce link, and found something that I take
> issue with. It's his comment about polygon seaming being a flaw unique
> to the PS1 hardware. This is very false. I don't know about the Saturn,
> but the N64 had it in many games. The Turoks for example had seaming in
> many places. Some levels in Mario 64 were plagued with it. And it
> wasn't limited to that generation either. Seaming appears often in the
> current generation consoles, even in top of the line games like Doom 3
> for the XBOX. In fact, the seaming in Doom 3 is terrible, almost as bad
> as Tomb Raider on the PS1! Here is a game that employs so many advanced
> features to achieve an ultra-realistic look, but has a horrible amount
> of seaming that nearly destroys the illusion. I really hope the next
> generation of consoles are able to eradicate once and for all.
The guys at Bleem! used to have a page up explaining to customers that
the splitting of polygons in PS1 games was a flaw of the hardware itself
which showed up in virtually every game with a moving camera. In
absence of that, I've had shots of this effect up for some time. All of
the shots on the left show what I'm talking about. On the Saturn or N64
if there were visible seams between polygons, it was a software issue
and the fault of the developer, on the PS1 it was virtually unavoidable.
http://www.gamepilgrimage.com/PS1flaws.htm
> "I wish that I had ever seen the game in action. The tiny
> screenshots on Gamespot do seem to reveal that the characters are
> textured. I wonder if there's a connection to games like Iron and Blood
>
> not being good games and their technical prowess."
>
> In this case, yeah. They focused so much on the graphics that they
> forgot about the gameplay. :) Seriously though, it's not THAT bad. It's
> just a tad below Toshinden, on par with Criticom. Ok, it is that bad.
> But you won't be playing it much anyways. You'll be doing more staring
> at the loading screen than actually playing. The load times are
> painfully long.
There's a point on both of these systems where graphics and things like
gameplay and AI have to balance, and going overboard on one will cause
the others to suffer due to limited resources. Now, even today, limited
development time can do the same thing, so the point might be moot.
> "The effect works, but the textured Saturn version has far more fine
> detail from the higher resolution texture mapping (uniform across all
> characters), and allows for more close-ups as a result."
>
> It's more detailed on the Saturn, but I don't think the difference is
> really that big. There is plenty of fine detail on Zack for instance.
> Look at his bicep and abdomen textures. That looks as good as Tekken 2.
It might look as good as Tekken 2, but it's still less than DOA Saturn.
Just look at the close up in the last picture here (
http://www.gamepilgrimage.com/DOA7.htm ). Then, not to be a perv, but
check out the many clevage shots of Kasumi on the various pages and try
to find one that shows that kind of detail in the PS1 game. Then
there's also more texture detail in the facial detail. When I was
taking the shots I was frustrated by the fact that the PS1 game never
evenzooms the camera in as closely as the Saturn game.
> "In addition,
> the Saturn version is higher resolution,"
>
> That's debatable. Has any authorative source stated its exact
> resolution? Such as one of the game's programmers? I'm also skeptical
> of VF2's alleged 704x480 resolution. I do recall a Sega rep saying it
> ran in that res, but that was just a rep. I don't think he worked on
> the game. He may have just assumed it. And in none of the interviews
> with AM2 (the ones I've read) did I ever see them state the exact
> resolution. Just that it was hi-res. I would think that such a
> resolution would be counter-productive since as far as I know both the
> Saturn and Playstation were limited to 8-Bit color in 640x480 and up.
> In old PS1 spec sheets, they gave the maximum color depth of each
> resolution, and 640x480 was limited to 256 colors. Regarding the
> Saturn, there was an interview (which is still archived on the net I
> think) with the designers of the Saturn net-link browser, and according
> to them, 640x480, which was the res that the browser ran in, was
> limited to 8-bit color depth. Now VF2 looks like it's showing a heck of
> a lot more colors than 256. I can't see any obvious color gradients in
> the distant bitmap for example. VF2 gives the impression of lots of
> color. If it really is using only 256 colors, then it's using them very
> well.
>
There is this link that seems to imply that the VDP1 was maxed out at
512x512x8 or 512x256x16.
http://en.wikipedia.org/wiki/VDP1_32-bit_video_display_processor
Then this discussion of the way TV screens work kind of flushes out the
idea that the rendering internally might be different than what the
system actually outputs. The same discussion also fails to question the
Saturn running multiple games at 704x480 at 60fps though.
http://groups.google.com/group/rec.games.video.nintendo/msg/a357b8b8ec953779?hl=en&
> "and *if* it has higher
> animation is doing more still that indicates to me that both systems
> were being pushed in their own unique ways."
>
> Saturn is doing more in the backgrounds (if the XBOX version is an
> accurate port). They look much, much better. The PS1 has probably the
> worst backgrounds of any PS1 fighter I have ever seen. They're so dull.
> But regarding the models, I have to give the edge to the PS1, at least
> as far as "doing more". You may prefer the higher texture detail
> though. The PS1 has full gouraud shading, simulated light sourcing, and
> polygon skin meshing (it does, I recently checked, it's just not as
> good as Tekken 3). The Saturn just has higher texture detail, but I
> don't see a very huge difference on the male characters. It's not like
> it's night and day. As far as polygon count is concerned, I would say
> that they're roughly equal. Neither appear to be pushing as many
> polygons per model as Tekken 2. Despite the heavy gouraud shading, I
> can still occasionally see the polygons facets (yeah, that's the
> correct word) on the shoulders of certain characters, and they clearly
> lack the complexity of the models in Tekken 2. Both games are probably
> using a Saturn VF2 style polygon count. Maybe a little better.
The Xbox game is more blurry than the original, thanks to the texture
filtering and Xbox display in general. Fine details that were sharp in
the Saturn original are much more subdued in the Xbox version. The
screen ratio is also slightly different, and it's obviously running in
emulation.
I agree that the polygon counts of the PS1 and Saturn games are roughly
equal, but I thought we had established that the Saturn version is
640x480, while the PS1 version is 512x480. I've seen other reports
stating that DOA is actually running on VF2's engine and is also
704x480, I've never seen any lower resolution quoted for either DOA or
VF2, although certain texturing issues and aliasing I've noticed leads
me to believe that some kind of "trickery" is being done to achieve the
resolution.
> "I'd personally take the higher resolution of the arcade version
> over the Gouraud shading. I much prefer the fact that you can see fine
>
> detail in VF2 and DOA for Saturn than that characters appear rounded
> and
> can a black back and white front in FV and FMM."
>
> Yeah, most people were of that opinion when it was first released. If
> it's running in the same res as Virtual On, then it's pretty low. VO
> has that same muddy look that PS1 Ridge Racer has, even RR4. The arcade
> version wasn't that much of a looker anyways. One of the least
> impressive Model 2 games IMO. Back in the day, my mall arcade had a VF2
> machine right next to a FV machine, and I remember the difference in
> detail being big.
>
On the upside, both FV and VO are doing more than most games of
that era, in the body armor being destructible and the full 3D nature of
the gameplay respectively. It kind of shows the limitations in hardware
and development time in the day that VF2 and DOA in the arcades look so
much better, and games which are more complex in gameplay lack in
noticable ways. FMM on the Saturn also has full 3D dodging
capabilities, something which seemed to be very difficult in that era of
3D fighters.
--
Scott
LocalH
layer, yes. However, within the background layers, each tile can be set
to low or high priority, and sprites can be set the same way, so you
have any of these five combinations of priority (S=sprite, A=scroll A,
B=scroll B, G=backdrop): SABG, ASBG, BSAG, SBAG, ABSG. Within each layer
of priority the order is SAB (if all three priorities are set to high or
low, then sprites are always on top of scroll A which is always on top
of scroll B). Proper use of the layer priority can create the illusion
of increased parallax.
Also, Sonic 2 used both scroll layers - the foreground level art (what
Sonic runs along) is on scroll A, while the background is on scroll B.
Scroll B is generally low-priority. In levels like EHZ, the background
is given a sort of parallax-looking effect by using line-by-line
horizontal scrolling (although this is supported by hardware, so the CPU
doesn't have to stuff scroll values into RAM on every line of the
effect). On parts of the level where Sonic goes behind the scenery,
those tiles are high priority. Tiles that show up behind Sonic are
low-priority. True parallax can also be done using low-priority sprites,
given that the Genesis can display 80 64x64 sprites by default (20 per
line), but you can also reset the sprite link table mid-frame and create
what is known as a "multiplexer", which reuses sprites multiple times
per frame to increase the total onscreen sprites.
Also, the Genesis has 6 FM channels and 3 PSG channels, or you can swap
out one FM channel for the DAC, so you have 9 total. The Z80 (which
commonly runs the sound driver in games) has 2KB of RAM and can access
the cart ROM in 32KB banks. The are also actually 1536 total colors,
when you take into account the hardware shadow/highlight feature
(although there may be some duplicates), and software can display them
all simultaneously (although such code is not useful for a game but
rather only a demo).
The Genesis' true max resolution is 320x448 on NTSC and 320x480 on PAL -
there is an interlace mode which doubles the normal vertical
resolutions. There are also 256 pixel wide modes which were used for
ports from SNES as well as a general method for saving a bit of VDP RAM,
of which you only have 64KB. NTSC systems can't run at 240/480 lines
(I've tried on my Genesis, the screen loses vertical sync until you
switch back to 224/448 lines)
Scott H
> The Sega Genesis hardware has two background layers and one sprite
> layer, yes. However, within the background layers, each tile can be set
> to low or high priority, and sprites can be set the same way, so you
> have any of these five combinations of priority (S=sprite, A=scroll A,
> B=scroll B, G=backdrop): SABG, ASBG, BSAG, SBAG, ABSG. Within each layer
> of priority the order is SAB (if all three priorities are set to high or
> low, then sprites are always on top of scroll A which is always on top
> of scroll B). Proper use of the layer priority can create the illusion
> of increased parallax.
>
> Also, Sonic 2 used both scroll layers - the foreground level art (what
> Sonic runs along) is on scroll A, while the background is on scroll B.
> Scroll B is generally low-priority. In levels like EHZ, the background
> is given a sort of parallax-looking effect by using line-by-line
> horizontal scrolling (although this is supported by hardware, so the CPU
> doesn't have to stuff scroll values into RAM on every line of the
> effect). On parts of the level where Sonic goes behind the scenery,
> those tiles are high priority. Tiles that show up behind Sonic are
> low-priority. True parallax can also be done using low-priority sprites,
> given that the Genesis can display 80 64x64 sprites by default (20 per
> line), but you can also reset the sprite link table mid-frame and create
> what is known as a "multiplexer", which reuses sprites multiple times
> per frame to increase the total onscreen sprites.
Hmm, so that's how they worked around the parallax limit. I'm curious
how they scroll independantly if they aren't actually independant in any
way.
> Also, the Genesis has 6 FM channels and 3 PSG channels, or you can swap
> out one FM channel for the DAC, so you have 9 total. The Z80 (which
> commonly runs the sound driver in games) has 2KB of RAM and can access
> the cart ROM in 32KB banks. The are also actually 1536 total colors,
> when you take into account the hardware shadow/highlight feature
> (although there may be some duplicates), and software can display them
> all simultaneously (although such code is not useful for a game but
> rather only a demo).
What is the shadow/highlight feature in game. I've heard this refered
to in Sonic 1, but I'm not sure what it's actually affecting.
Screenshots from emulators reveal the actual in game limit of most
games, including Ranger X, as 50-55 colors at once.
> The Genesis' true max resolution is 320x448 on NTSC and 320x480 on PAL -
> there is an interlace mode which doubles the normal vertical
> resolutions. There are also 256 pixel wide modes which were used for
> ports from SNES as well as a general method for saving a bit of VDP RAM,
> of which you only have 64KB. NTSC systems can't run at 240/480 lines
> (I've tried on my Genesis, the screen loses vertical sync until you
> switch back to 224/448 lines)
How does that compare to the Snes's 256x224 mode, can it also do 448?
--
Scott
LocalH
parallax - all they're doing is scrolling each line separately such that
it looks like lower lines are scrolling faster. This is easy as pie -
all you have to do is, somewhere in your main game loop, calculate the
proper scrolling values for each line and stuff them into RAM, then
during vblank you can DMA this table of values to VRAM, and the VDP will
scroll each line as you told it to. And since you can scroll each layer
independently, this doesn't affect the foreground at all.
For true parallax, since the hardware can scroll each line independently
on each layer, you can get some pretty sophisticated effects if you do
it just right - creative use of tile placement and priority bits, along
with sprites, could create the illusion of several layers of parallax
when you only have three layers (S, A, and B).
A similar effect could be used vertically, but the hardware can only
scroll each 2-tile column separately, as opposed to every scanline. So
you didn't see this as much as the Sonic-style "parallax".
>> Also, the Genesis has 6 FM channels and 3 PSG channels, or you can
>> swap out one FM channel for the DAC, so you have 9 total. The Z80
>> (which commonly runs the sound driver in games) has 2KB of RAM and can
>> access the cart ROM in 32KB banks. The are also actually 1536 total
>> colors, when you take into account the hardware shadow/highlight
>> feature (although there may be some duplicates), and software can
>> display them all simultaneously (although such code is not useful for
>> a game but rather only a demo).
>
> What is the shadow/highlight feature in game. I've heard this refered
> to in Sonic 1, but I'm not sure what it's actually affecting.
> Screenshots from emulators reveal the actual in game limit of most
> games, including Ranger X, as 50-55 colors at once.
>
S/H is used in several games. Sonic 2 uses it as a debugging aid (you
may have heard of a so-called "night mode" cheat - this essentially
turns on S/H which can be used to check and make sure that the
priorities are right on the tiles). Castlevania: Bloodlines uses it
in-game for a quite nice effect where a light is shining on an object,
and everything behind the light "beam" is normal, while everything
outside the "beam" is dark (shadow). I think Ecco 2 also uses it in some
places. I'm sure there are many other games that use S/H that I don't
know off the top of my head. If you want to see the difference between
shadow, normal, and highlight, check out this page:
http://cgfm2.emuviews.com/old2001.php
Scroll down to the 5/27 update and you'll see a screenshot of this demo,
along with a ROM that runs on emulators or real hardware.
>> The Genesis' true max resolution is 320x448 on NTSC and 320x480 on PAL
>> - there is an interlace mode which doubles the normal vertical
>> resolutions. There are also 256 pixel wide modes which were used for
>> ports from SNES as well as a general method for saving a bit of VDP
>> RAM, of which you only have 64KB. NTSC systems can't run at 240/480
>> lines (I've tried on my Genesis, the screen loses vertical sync until
>> you switch back to 224/448 lines)
>
> How does that compare to the Snes's 256x224 mode, can it also do 448?
>
As far as I can tell, the Genesis 32 cell (256 pixels) mode is extremely
close to the SNES output in terms of the actual display (although I can
about guarantee that the timings are a bit different). Also, in 32 cell
mode, you're limited to 64 sprites instead of 80, and 16 per line
instead of 20 per line.
I'm pretty sure the SNES can output a laced 448 lines as well, and it
also has a 512 pixel wide mode that some games use.
Scott H
> Scott H wrote:
>> Hmm, so that's how they worked around the parallax limit. I'm curious
>> how they scroll independantly if they aren't actually independant in
>> any way.
>
> Well, the background in Sonic 1's GHZ and Sonic 2's EHZ is not true
> parallax - all they're doing is scrolling each line separately such that
> it looks like lower lines are scrolling faster. This is easy as pie -
> all you have to do is, somewhere in your main game loop, calculate the
> proper scrolling values for each line and stuff them into RAM, then
> during vblank you can DMA this table of values to VRAM, and the VDP will
> scroll each line as you told it to. And since you can scroll each layer
> independently, this doesn't affect the foreground at all.
>
> For true parallax, since the hardware can scroll each line independently
> on each layer, you can get some pretty sophisticated effects if you do
> it just right - creative use of tile placement and priority bits, along
> with sprites, could create the illusion of several layers of parallax
> when you only have three layers (S, A, and B).
>
> A similar effect could be used vertically, but the hardware can only
> scroll each 2-tile column separately, as opposed to every scanline. So
> you didn't see this as much as the Sonic-style "parallax".
Yeah, I knew that wasn't parallax, the effect I was refering to is in
the Hill Top Zone (I think that's it) in Sonic 2, where there are at
least six layers of clouds in the background scrolling independantly all
the way to the far background. Also, Sonic 1's Japanese version has two
cloud layers scrolling independantly of everything else on the screen in
the first level, Thunder Force five has something around 8 layers in the
first level alone, and Shinobi III does at least 6 in the first level.
What's always thrown me off about these levels is that they are solid
horizontal planes scrolling independant of each other, rather than a
sprite or a raster effect stuck on a flat color background to cause an
optical illusion. In other games I can also see one plane that's moving
slower than the one next to it jump pixels, and thus animate poorly and
possibly be "faking" the independance of the scrolling. But the games
listed either don't do that, or they covered it up so well I haven't
noticed it (I haven't recently glared at each layer to be sure).
Huh, that's really interesting. According to Gif Movie Gear, it's
displaying 253 colors. I need to get a level code to the last level of
Castlevania and take a screenshot....
>>> The Genesis' true max resolution is 320x448 on NTSC and 320x480 on
>>> PAL - there is an interlace mode which doubles the normal vertical
>>> resolutions. There are also 256 pixel wide modes which were used for
>>> ports from SNES as well as a general method for saving a bit of VDP
>>> RAM, of which you only have 64KB. NTSC systems can't run at 240/480
>>> lines (I've tried on my Genesis, the screen loses vertical sync until
>>> you switch back to 224/448 lines)
>>
>>
>> How does that compare to the Snes's 256x224 mode, can it also do 448?
>>
>
> As far as I can tell, the Genesis 32 cell (256 pixels) mode is extremely
> close to the SNES output in terms of the actual display (although I can
> about guarantee that the timings are a bit different). Also, in 32 cell
> mode, you're limited to 64 sprites instead of 80, and 16 per line
> instead of 20 per line.
Why would dropping the resolution cause the sprite limit to go down?
Were some other effects available that made this mode usefull, aside
from merely being "like" the Snes's output?
> I'm pretty sure the SNES can output a laced 448 lines as well, and it
> also has a 512 pixel wide mode that some games use.
Yeah, but as I understand it the Snes could only use that 512 mode with
16 colors and a single plane (no sprites).
--
Scott
LocalH
> Yeah, I knew that wasn't parallax, the effect I was refering to is in
> the Hill Top Zone (I think that's it) in Sonic 2, where there are at
> least six layers of clouds in the background scrolling independantly all
> the way to the far background. Also, Sonic 1's Japanese version has two
> cloud layers scrolling independantly of everything else on the screen in
> the first level, Thunder Force five has something around 8 layers in the
> first level alone, and Shinobi III does at least 6 in the first level.
> What's always thrown me off about these levels is that they are solid
> horizontal planes scrolling independant of each other, rather than a
> sprite or a raster effect stuck on a flat color background to cause an
> optical illusion. In other games I can also see one plane that's moving
> slower than the one next to it jump pixels, and thus animate poorly and
> possibly be "faking" the independance of the scrolling. But the games
> listed either don't do that, or they covered it up so well I haven't
> noticed it (I haven't recently glared at each layer to be sure).
>
Genesis can show 80 sprites in 40 cell mode, there can usually be some
set aside for effects like those. Sprites can also be visually and
programmatically linked, where you have two or more sprites that mesh
together for a large object, so for example you can use two sprites
side-by-side to get a 128x64 object at 16 colors.
> Why would dropping the resolution cause the sprite limit to go down?
> Were some other effects available that made this mode usefull, aside
> from merely being "like" the Snes's output?
>
Because when you change the resolution, I believe you're actually
changing the clock rate that the VDP runs at, and since a lower
resolution would mean a lower clock, you would naturally have less
sprites and less pixels-per-line. Sort of cyclical logic, really, but it
should do.
There weren't really any new effects available, but it did mean that a
full screen would take up less tiles (320x224 mode used 40x28=1120
tiles, while 256x224 mode used 32x28=896 tiles), and thus also less VRAM
for the patterns (assuming each tile used a distinct pattern, 320x224
would take up 1120*32 bytes per tile=35840 bytes, or over half the VRAM,
while 256x224 only takes up 896*32=28762 bytes, about 4K less than
half). 256x224 was used for ports from SNES so that art would not have
to be redrawn for the proper aspect ratio of the 320x224 mode (where
pixels are taller than in 256x224).
>> I'm pretty sure the SNES can output a laced 448 lines as well, and it
>> also has a 512 pixel wide mode that some games use.
>
> Yeah, but as I understand it the Snes could only use that 512 mode with
> 16 colors and a single plane (no sprites).
>
Now there, I don't know - my area of expertise is the Genesis, and I
don't know too much about the inner workings of the SNES PPU.
Scott H
> Scott H wrote:
>
>> Yeah, I knew that wasn't parallax, the effect I was refering to is in
>> the Hill Top Zone (I think that's it) in Sonic 2, where there are at
>> least six layers of clouds in the background scrolling independantly
>> all the way to the far background. Also, Sonic 1's Japanese version
>> has two cloud layers scrolling independantly of everything else on the
>> screen in the first level, Thunder Force five has something around 8
>> layers in the first level alone, and Shinobi III does at least 6 in
>> the first level. What's always thrown me off about these levels is
>> that they are solid horizontal planes scrolling independant of each
>> other, rather than a sprite or a raster effect stuck on a flat color
>> background to cause an optical illusion. In other games I can also
>> see one plane that's moving slower than the one next to it jump
>> pixels, and thus animate poorly and possibly be "faking" the
>> independance of the scrolling. But the games listed either don't do
>> that, or they covered it up so well I haven't noticed it (I haven't
>> recently glared at each layer to be sure).
>>
> Well, some of those effects can be done using sprites - since the
> Genesis can show 80 sprites in 40 cell mode, there can usually be some
> set aside for effects like those. Sprites can also be visually and
> programmatically linked, where you have two or more sprites that mesh
> together for a large object, so for example you can use two sprites
> side-by-side to get a 128x64 object at 16 colors.
Hmm, that'd only take four sprites to make something reach almost all
the way across the screen. I wonder how that'd work out on a scrolling
screen though where the layer just keeps cycling.
>> Why would dropping the resolution cause the sprite limit to go down?
>> Were some other effects available that made this mode usefull, aside
>> from merely being "like" the Snes's output?
>>
> Because when you change the resolution, I believe you're actually
> changing the clock rate that the VDP runs at, and since a lower
> resolution would mean a lower clock, you would naturally have less
> sprites and less pixels-per-line. Sort of cyclical logic, really, but it
> should do.
Well, that makes sense, not, I mean in regard to why they did it that
way, but that explains the lower performance. ;)
> There weren't really any new effects available, but it did mean that a
> full screen would take up less tiles (320x224 mode used 40x28=1120
> tiles, while 256x224 mode used 32x28=896 tiles), and thus also less VRAM
> for the patterns (assuming each tile used a distinct pattern, 320x224
> would take up 1120*32 bytes per tile=35840 bytes, or over half the VRAM,
> while 256x224 only takes up 896*32=28762 bytes, about 4K less than
> half). 256x224 was used for ports from SNES so that art would not have
> to be redrawn for the proper aspect ratio of the 320x224 mode (where
> pixels are taller than in 256x224).
I've noticed that in some multi-platform titles (mainly Weaponlord) have
a large black bar on the top of the screen in the Genesis version, and
they sometimes use this bar for energy bars and such which were
overlayed on the graphics in the Snes version. I had assumed that in
these cases the Genesis version was wasting the extra resolution height
by simply making it black (rather than stretching the image). Since the
resolution increase is horizontal, not vertical, I'm pretty sure my
assumption was wrong and I have no idea why the Genesis version has the
black bar at the top with no visible cropping of the image.
>>> I'm pretty sure the SNES can output a laced 448 lines as well, and it
>>> also has a 512 pixel wide mode that some games use.
>>
>>
>> Yeah, but as I understand it the Snes could only use that 512 mode
>> with 16 colors and a single plane (no sprites).
>>
> Now there, I don't know - my area of expertise is the Genesis, and I
> don't know too much about the inner workings of the SNES PPU.
Well that expertise is appreciated. It's good to know more about
what was going on in the games from a technical standpoint. Most of the
public perception of the SNES being vastly superior just doesn't seem to
be based on a technical viewpoint, but on more of a wishy washy
"perception" idea, propagated by media and crazed fans. It's just as
wrong as saying the Genesis had the absolute hold on "speed", it
depended on the software, and by and large software was great on both
platforms.
--
Scott