Musings on the ethics of shooting screens

Ever since I first hold my print copy of HG101 Presents: The Guide to Classic Graphic Adventures in my hands, there's one anachronism that personally keeps bugging me about it. It uses an old-ass CRT monitor alright, but the screenshot of King's Quest that's 'shopped onto it looks like it would be displayed in DOSBox or ScummVM (with default settings):

Instead of filling out the entire screen, it leaves huge black borders at the top and bottom. That is not the way the game was meant to be viewed, but it's the way we nowadays play those old games with our modern monitors, where every pixel is always as high as it is wide. At least it is this way whenever we play old games the "pure" way, without additional scaling and filters applied, when pixels are as blocky as they should be. Or shouldn't they?

It hasn't always been like this. In fact, the age of pixel-perfect screenshots only really came upon us with the widespread of emulators around the 2000s and cleaner printing in the magazines. In the very beginnings, long before even AV-capturing was available, creating screenshots used to be an art; not for skillfull photographing techniques or the timing to get the right screenshots, no, "screenshots" were literally drawn by artists, often with greater-than-life proportions and situations that could have never appeared like that in the game. In times of limited magazine space and the laborous process of having to hand-draw each screens, as much action as possible was often squeezed into these images. The philosophy behind this is best exemplified in this reader letter to the late Bill "The Game Doctor" Kunkel in Electronic Games March 1983:

Q: Back in your second issue, on the last page, there was an advertisement for Midway games. Notice on the Pac-Man screen that the gobbler's on the apple but he's only got'460 points—not nearly enough to reach the apple rack. How did they do it? (Grey Oliver, Austin, TX)

A: Most of the playfields seen in magazines and advertisements, Grey, are not photographs but artist's reproductions of what you actually see on the monitor. There are several reasons for this, most prominent being that videogame screens are quite difficult to photograph clearly (as enyone who has ever tried to record a high score can testify). Even the best pictures tend to feature considerable distortion as the glass tube is slightly domed, producing a "bubble" effect.

Also, artistic renderings of playfields allow a single picture to tell a more complete story than any single frame of actual game play. Seldom do all of a game's most picturesque objects appear on screen simultaneously, but in a drawing, the potential player gets a better look at all the contest's elements.

Especially (but certainly not exclusively) in advertisements, this often led to some creative liberties taken with the game, showing off bigger sprites and more colorful effects than the machine could ever display compare with the above "screenshot" of Donkey Kong from Video Games these other two out of Video Invaders (left) and Invasion of the Space Invaders.

Especially the latter's open discontent for the game seems to be represented in the way Mario is shown:

Obviously, this method had to go away with the rising complexity of games' visuals, although a magazine full of skilled artists' representation instead of actual screenshots would be definitely interesting.

But let's get back to the initial issue: Nowadays the blocky pixel-perfect image has become a weird anachronist ideal in retro gaming, despite the fact that the actual designs were neither displayed, nor conceived in this way. Taking the risk of overusing this comparison, have a look at the abstract concept for the way Pac-Man was supposed to look in the minds of the player, how he actually would have looked on an arcade TV screen in the 1980s, and how most players experience him now via emulation in MAME (left to right).

Now take a look at this villain from Might and Magic 5:

He used to have a much larger and more impressive build when meeting him in 1994. Back in the era of SD television, most games in Europe ended up compressed in a similar way, due to the PAL video standard havig a higher vertical resolution than NTSC. While simply unavoidable with 2D-games, not properly converting the resolution turned to an important point of critique in the 32-to-64-bit era. But nowadays, everyone seems to take all these characters suffering from dwarfism at face value. It's the other way around with NES and SNES games, where the pixels are actually more wide than high. The pixels, it seems, have become more important than the overall impression. But what do we make of extreme cases like Ganbare Goemon: Kuru nara Koi! Ayashige Ikka no Kuroi Kage on the PSX, whose original resolution looks like this?

Another infamous issue is dithering, which many old games abused to make a game look like it had a higher color palette or even transparency effects. Even back in the day, this could easily be spotted on a big enough screen, but from a big enough distance a game like Ristar may have looked closer to this:

As opposed how we show it at Hardcore Gaming 101:

Let's take a look at Ninja Gaiden II on the NES for an example of what can go wrong when documenting it with the modern methods. A random pixel-perfect screenshot taken from the game in an emulator may look like this:

What's immediately off-putting is the fact that only one of the two ninja shadows is visible, while only the shuriken thrown by the invisible one is on display. The next frame explains the mess:

In the good old days, sprite flickering wasn't simply a glitch, often programmers used it willingly to simulate transparency. On a really bad old TV, the shadows (and shurikens) wouldn't really be viewed as constantly phasing in and out of the picture, but as one always visible, if unsteady, shape. A perfect emulator screenshot in this case produces a very wrong image of the scene. Actually the impression would be more like this:

Of course there are still problems left with that, as the elements from the second screen are always one frame of animation ahead of the first, which becomes a real problem when one of the permanently visible elements moves in between, but there's really no way around this.

Next, NES games, while displayed in 256x240 resolution, were always made under the assumption that the upper and lower 8 lines would be cut-off on a standard NTSC screen, which brings us to this:

Then we would have a 4:3 screen ratio:

But we're still not done: Many emulators actually allow options for scanlines, resulting from the digital-to-analog conversion for contemporary TV sets. Of course none of these options really look like the real deal, and usually turn out completely wrong when resizing to the thumbnail size of 320x. If you've got a really old TV, its frame may also cut off the corners:

There's yet one more problem when it comes to preview material and beta version. Obviously, we can't take our own screenshots (unless the versions got leaked) so we're often stuck with either:

-a low-quality old magazine shot:

-a watermark-spoiled shot from a shitty irrelevant website:

Well, since in this case I'm lucky enough to have both, I won't put up with any of that bullshit, and instead go on restoring the best possible variant (that I can achieve with my limited skills).

It may not look perfect (but it can get close when done by someone who actually knows her stuff with image editing), but turns out far better than any of the other available options.

N64 screenshots in general are an issue, as the emulators still haven't gotten (and maybe never will get) the effects quite right. Besides the typical hardware blur, the fog looks also much different on a real system, and on some bigger textures one can see the seams in between tiles. With a capture card, on the other hand, one has to live with additional AV blur and off colors, which hare always dependent on individual TV settings.

The seemingly simple question of selecting "good" screenshots also often bothers me—I usually just take hundreds of screenshots randomly while playing (unless there's something specific I want to show), and then select those I deem the most interesting. But do those really represent the game properly when I have 20 interesting screenshots and 300 boring ones? Let's assume you write a regular column of game introductions, and you get 3 screenshots to chose for each. My impulse for a game with three different levels would be to get a screenshot from each to show the most variety. But that approach would also make it impossible to do the game with 30 different stages justice. When 90% of the time in an RPG is spent wandering around bleak environments and dull dungeons, can I "rightfully" show 50% battle scenes?

There are so many possible ways to represent (old) games in screenshots. Every site or magazine has its own policies to deal with these issues, Hardcore Gaming 101 sure has its own. But is there a "right" way to do screenshots? Which do you consider the "best" one? Are there some edits you consider immoral/wrong? Let's see some opinions.