This is a very controversial topic and perhaps one that will land me a lot of enemies. I nonetheless believe that it is imperative to understand the point: the existence of game remasters is a symptom of a larger problem, that will soon become all-encompassing. It is my firm belief that we as a species have blundered our way to this problem unknowingly, and that we can still fix it… albeit with a lot of effort, specifically pertaining to unlearning some of the bad habits that we have accumulated.

This rant will prominently feature Nightdive studios as a punching bag. The reason is twofold. One is that Nightdive clearly cares much about the games that it remasters, and obviously about game preservation in general. The second reason is, because they are contributing precisely nothing to the solution, and ensuring that the problems which cause many games to become unplayable, are made worse, with the possible exception of the games that they have personally worked on. It indirectly relieves the people who got us into this mess in the first place, from any form of responsibility, as well as de-values the work done by the fan-bases. In other words, Nightdive is used as pressure release valve, ensuring that the real problem is never solved, and only somewhat addressing some of the symptoms.

So without further ado.

Games are art Link to heading

And I stand by it. Games can tell stories and convey emotions that you wouldn’t be able to experience otherwise. I can demonstrate this on the example of the original Deus Ex which is art even by the standards of conventional art forms, due to its narrative structure, the fact that it makes one think and astute symbolism (the area which did not get transferred to the modern remakes very well). I would argue that Star Craft tells a very impersonal unattached story that allows one to retain perspective on the larger things that matter, divorcing oneself from concrete emotions, without simultaneously being completely emotion-free. It is a narrative experience that is only possibly closely mirrored by the Song of Ice and Fire series, in the sense that it forces the person experiencing the art to not get attached to concrete people, but to still have a sense of right and wrong. Finally, a common sentiment when Star Trek: Picard season 1 came out was that the Frankenstein’s monster narrative as regards synthetic life was handled far better in the Mass Effect series.

As such I stand by the statement that games are art. Even if your reference point is the mindless narrative-free first-person shooters of the 90’s, they have a close parallel in sculpture, as the three-dimensional environments have a distinct personality to them. In addition to the obvious appeal to the medium itself and the undeniable experience that is either familiar to the reader of this article, or a crucial missing piece, consider that the staunchest of critics of gaming have themselves contributed to them: George R. R. Martin, perhaps the most important example, given that he is the only one with viable criticism, the fact that he himself contributed to the Elden Ring, speaks volumes to the medium’s capabilities.

So why is this important to the question of remasters?

Well, for a few reasons. Most art is quite accessible from a sensory standpoint. Mainly, if it is an art gallery all one needs is some eyesight. It is not even good eyesight, because some forms of art, particularly late works impressionist painters are best appreciated only if the sight has deteriorated somewhat, in line with the deterioration of the eyesight of the artist himself. Statues are things that can be appreciated by the blind, as is music.

All one needs is the art in front of them. And to see the The Winter by Pieter Brueghel, one does not need to travel to Amsterdam: copies of it exist, and can even decorate one’s home. Art preservation is thus a simple act of keeping a painting in your house. Sure, there are museums, and different copies are different in different ways. Some copies retain a lot of material which was in contact with the author, but most paintings have to be restored. The thing that is being called the original and is posted in Musee D’Orsay is no more the original, than a very good copy made by a professional artist. The vast inflation in the price of art is the only reason why your Lautrec is different to the one in the museum, is that the snobs believe that their ship of theseus has more connection to the work of the author than yours does. For all intents and purposes the art of those art pieces is completely identical.

Film is a slightly more modern (on historical time-scale, very very recent), form of art that has slightly different requirements for preservation. Film deteriorates faster than paintings do. Not faster than many of the crafts that we haven’t preserved, there’s some survivor’s bias going on, but still faster than one can antiicipate. Furthermore, the form in which film is being preserved is in flux comparable to that of the early renaissance. So, preserving film at the end of the 20-th century meant preserving a bunch of reels in a mine. Preserving film today, most likely only involves preserving a few digital files, that take up a fraction of the space, and are completely lossless if preserved correctly. We now know, as well, that since we can make lossless copies of recordings of Richter playing in Vienna, or the masters of Star Trek: The Next Generation, that preserving is not a matter of guarding well a single copy, but rather ensuring that the copies are proliferated well, and that there are no self-imposed losses. Digital storage is abundant, cheap and universal.

Art preservation, particularly digital art, whose abstract nature inherently kes preservation easy and controllable, is a solved problem. Sure, we can’t exactly convert the Starry Night into a JPEG and call it a day, because the brush strokes retain a lot of information, but with enough effort, it is possible to produce a digital clone of the painting that retains everything that an art snob would consider important, and ignore everything else. True that scan might take up as much space as the original painting, if stored on outdated hardware, if sufficiently detailed and sufficiently poorly compressed; but that scan can be sent very quickly via the internet, and reproduce the exact same painting halfway across the globe. If there is a fire, you have a backup from which it is possible to restore the “original”.

But despite the advances in technology, we still have a problem with game preservation. And while the reasons are quite similar to those of film, the imposition of additional restrictions by less than intelligent individuals, has resulted in games becoming a lossy medium that is for some reason harder to preserve than decaying dying organic material attached to a mutilated corpse of a plant with ostensibly imperfect checmicals. And believe it or not, egg-based paint has more longevity than Doom.

But why is that.

A brief history of Nightdive studios Link to heading

This, much like any good story began with a bit of frustration with the status quo. One regular person was on a trip, and lo and behold they couldn’t play their favourite good-old-game, System Shock.

There are a multitude of reasons for why you can’t play system shock these days. The game is old, and I’m not just referrring to the fact that it’s old fashioned or has accumulated a lot of technical differences: the way it was written has a few assumptions about the underlying hardware that no longer hold true. It is a game that didn’t do too well either. It didn’t give its authors the confidence to make their game engine open source, as was common with Id Tech games, and the game wasn’t super profitable to make the original publishers stay afloat, or this particular IP to be their prized possession.

As a consequence, if you want to play the original System Shock, you could still do it, but you’d have to jump through a lot of hoops. The systems for which it was built no longer exist, and it is not quite easy to reverse engineer what it was meant to do… Or at least not as easy as with other games, given how unconventional the game was, and how different its engine.

So naturally, there was some idea that it would be possible to bring those games into the 21st century. That it would be possible to update the assets and game engine, without affecting the core gameplay. And Nightdive stuios was largely inspired by that goal.

Their roster of games is quite interesting, and I should probably talk about the specifics of how they do what they do. Their first foray into remaster territory was of course System Shock 2 in 2013, which was released on Steam and was the best, at the time, way to play that game. They had also fixed up I have no Mouth and I Must Scream, and a few other smaller titles. This was done largely in an effort to bring those games back to life from the dead.

The reality of the way in which it had to be argued is quite revealing: one had to email the rights holders of the “intellectual” property in question. Naturally this includes almost nobody involved in the original game, and a bunch of suits and CEOs. The argument was not “we should preserve these games for the sake of preservation”, but “there are N people who have wishlisted this game on GOG or Steam, and a re-release would be profitable”. Naturally, this very much is the case with System Shock. In fact, up until very recently I was salivating over the System Shock remake done by Nightdive, and this particular remake re-contextualised why I think Nightdive isn’t helping.

A world without remasters Link to heading

System shock is far from the only game that came out during that time, that people would like to play and play religiously. I have played Blood, Quake 1, Quake 2, FreeSpace, Half Life and quite a few other games from that same era. But these were games I could play for the longest time, thanks to one particular peculiarity.

Half Life, is very much a brainchild of Valve. It is their roots. And unlike many other companies, (ahem, Epicfail, ahem), they respect where they came from, to the point of continuously updating the game. Sure, it is quite a fallacy to believe that these updates were universally an improvement, but one cannot deny that they could play a reasonable approximation of Half Life on modern hardware for the vast majority of its existence. This was not free of blunders, like making the “HD” models not just the default, but the only way to play the game, and the infamous Half Life Source, that wasn’t nearly as broken initially as it was by the end of its miserable existence.

But thankfully, Valve also respected others imitating and replicating their work. Black Mesa, was a free mod for the Source Engine, that allowed one to play through the vast majority of the good parts of Half Life. I distinctly remember liking that mod a lot more than I ended up liking the retail game, but that has laregly everything to do with the fact that Valve had a clearer understanding of games, than the Crowbar collective. Still, having the option to play a pretty-much-fan remake of Half Life, is a good one to have.

With Half life, one had the GoldSource version that was faithful-enough and came from Valve directly, in fact more faithful with the latest patch. The Source version which was hopelessly broken, but was nonetheless a somewhat of an upgrade at the time it came out. The Source fan-made mod, which was a short, but faithful adaptation. And a Source-based remake; one which is less broken than the official Valve product, has some direct upgrades, and some blunders for good measure. Still, one cannot say in 2024, that it is impossible to play Half Life, in fact one is spoiled for choice.

Still another game from a similarly long bygone era, is freespace (1&2) from 1998 and 1999 respectively. It is a game with quite a lot of depth and a surprising amount of emotional investment for what is ostensibly a slightly sci-fi military piloting simulator. It’s a fun game, that I quite liked when it was modern. Thankfully, I can still play it. The reason is the hard light project. The basics of it come down to: the source code for the program is open. The so-called MediaVPs are not, but you can, if you own the game, refurbish the assets and play them with the new engine. And in some cases there are improved and upscaled (manually) assets. The game has a plethora of gameplay improvements. The number-based command dispatch was replaced with Windows speech recognition, a feature that I used quite extensively. There are multiple campaigns, and multiple missions packs that could be run on the same engine. The game is just fun in ways in which the original could not be. With freespace the game is preserved by virtue of being open.

With Quake, the situation is even more interesting. When the game was completed, the Source code for it was released under the GPL; the best license for free software. There are a few contributing factors to why Quake is such a wonderful game to keep updated to run on modern hardware. Firstly the game code is relatively small. It was one of the earliest game engines, so the amount of code that needs to be ported is itself small. Secondly, the code itself is of good quality. Sure it’s C code, which imposes some limitations on what can be done effectively with it, but it is very good C code. The third aspect is that Quake’s source code had been available for a very long time. The latest GPL release had been out for 13 years. There were previous releases long enough so that code modifications could be lost. So where we are now, very much falls into the previous two. However, it should be noted that Quake to a far greaterextent was designed to foster the creativity of its userbase. Even before the source code had been released, great effort was taken to ensure that the game was easy to modify, easy to create with and within the confines of its genre, created the necessary preconditions for the creation of the aforementioned Half Life. FreeSpace was not as forthcoming, despite the premise being more suitable, the malleability of it was less, by virtue of its fanbase being much smaller.

These games were in no need of a remaster, and they lived on because they were good, but also because their creators had made the decisions that made these games a permanent fixture of modern society. There are calculators running Doom, which is a far cry from the IBM PCs that they were intended to run on. Meanwhile getting something like Assassin’s Creed: Brotherhood to run on the intended platform is a challenge; so games preservation is also very much a function of the game that is being preserved.

So let us now turn to yet another example of a game that has been preserved well. Quake 2, which for some unfathomable reason is not called Wor, is a game that had been used to demonstrate the effects of Ray tracing and path tracing.

The reason why it was chosen was threefold: it was a game with an aesthetic that could support the kind of lighting that nVidia wanted to show off. Secondly its models were sufficiently simple, so that path tracing was not prohibitively expensive. Thirdly, and perhaps most importantly, the code was GPL, meaning that nVidia would have been capable of adding on their RTX renderer, wityhout fear of some of their competitors taking advantage of the code, although RTX itself being proprietary opened a whole other can of worms that GPL could have exacerbated. Nonetheless, the kind of game that Quake 2 RTX ended up being is quite an interesting conundrum.

On the one hand, there is a popular opinion that Quake 2 RTX veers too far away from the art style of Quake 2, and if RTX were considered an official remaster, or e.g. a replacement for the Quake 2 that one can play via Steam, or by using the retail CD, I’d argue that mayhaps this would be a valid critique. But it is also an experimental title. It is a test of how well, can the new features integrate into an older engine, and how much work would need to be done to do so. The reason why that in and of itself is a good thing should be obvious, but in case it is not, consider that in some games, screen space reflections are so well-tuned that it is difficult to tell if path tracing is applied at all. Yet, we can experience Quake 2 in that way too! How fun!

So do these games need officially licenced remasters? I believe not, and I would argue that this imposes an evolutionary selection criterion: if you want your game to last longer, make it open source, make it replay-able and give players choices as to how to experience it. If you don’t, your game will disappear. Simple as that.

Rebastardisation Link to heading

We should talk a little about the process of remastering and how good intentions lead to inferior results.

The first remaster that I thought about getting was that of Homeworld 1&2. It is a game that I enojyed very much when I was but a boy, and for quite good reasons. The game was ingenious, and despite the derision for the sequel, a very interesting experience.

And the remaster looked amazing when it first came out. I spent a lot of days playing the campaign, still more playing instant action and figuring out that despite my best efforts, the AI was plain unfair.

So was it a good remaster. For the longest time I thought so. It did a lot to upscale the textures, it added some features that were missing from the original. I did like a few things, but there were decidedly problems with the way the games worked in the new engines.

The first game was affected the worst, the original Homeworld 1 had a very peculiar balancing, wherein the hit-check was based on a physical calculation. It was replaced with a random number generator. So as you can imagine ships that used to do very well in groups, like the guns on some large craft were no longer capable of shooting down fighters… which makes the peculiar balancing even more favourable towards smaller craft.

Another problem I found was that the lines were re-recorded. Homeworld 1 feels largely the same, as if alternative takes were used, but with no major differences. Homeworld 2, however, feels like a different game. Fleet intelligence in particular was a different character altogether whose lines no longer agreed with the delivery. In fact, it is rather challenging to tell whether the line delivery was off, or just the perception of the radically different voice… Does it make it a bad remaster? I suppose not. But it does make it harder to recommend.

Blood Link to heading

This is a game that is largely in the same camp as Quake and Freespace, despite the source code not being available, the widely revered build engine was reverse engineered in more than one way. There was Blood GDX, made with Java, and nBlood which was written in C++. Both were fan-made projects with quite a bit of good faith and a decent amount of performance too.

And then the nightdive remake came out. And it was … well, it was certainly a remake, the game didn’t look particularly better than either nblood or Blood GDX. It had a Vulkan renderer, and a few other goodies, but to be quite frank, it wasn’t particularly good. It was certainly a way to experience blood, because it didn’t require as much work to be done on Windows.

But I should also mention this. I’m a GNU Linux user. I do this because the kind of work that it requires one to know how that operating system works and being submerged into that ecosystem, makes economical use of my time. I had always adored the fact that games that had their source code reverse engineered or available, or in fact, open, worked on Linux quite well. Despite the many problems Blood GDX had that. nblood, despite being written in a lower level language, was also capable of running on my machine without any translation. The nightdive port could not.

Tell, me is that entitlement to expect people who have a full-time job, to produce something of at least quality parity as someone who had been doing the port in their spare time? One may well argue that asking for Linux support is asking for too much… And I may well argue that if someone did it for free, with better performance, and with some additional advantages, like the fact that the game wasn’t broken in many crucial ways, and on top of all of that, did so without access to any of the original assets, I suppose that it may also well be argued that Nightdive is not particularly competent.

Now, this all assumes that the game is otherwise identical, and performs similarly to nblood. Nosedive’s in-house rebastardisation engine managed to take a game that ran smoothly on turn-of-the-millenium machines, and make it run worse than the poorly optimised Black Mesa, or the extremely well-optimised, but demanding Doom (2016). It takes a particular kind of talent to take a game, add a Vulkan renderer, a bunch of completely unnecessary graphical “features” and make it run poorly. Am I being too harsh?

The Quake 1 remaster Link to heading

This is probably the most interesting remaster to cover. From an outsider’s perspective, it is a faithful remaster, which added co-opeartive multiplayer, remastered the game with good performance, while adding features such as dynamic lighting, among other graphical features, added in many of the official expansions, if not all of them, and to add more to the package, they made sure that the game retained many of the features that made other modifications work. And added a horde mode.

Well, almost. The story modifications that didn’t made any changes to the source code do in fact work. But most other things, like e.g. Quake 1.5, naturally don’t work anymore.

While the game is decidedly better than the original retail release, I’d argue that a lot had been lost. Due to the source code for the Kex engine not being freely available, the mods are limited to those that utilise the base quake code. So no fancy guns for you.

Secondly, I may be the only one irritated by the fact that a game that should load instantaneously, even if in a heavily modified engine with extreme upscaled assets, loads for a few minutes, and tries to connect to Bugthesda.net.

Another problem that stems from the fact that Kex is not publicly available, is that it doesn’t work as well on GNU Linux. For example the fancy co-op doesn’t, despite many fan-made mods for the original quake doing that just fine. The game is a windows-only executable.

The horde mode, being official is quite a peculiar thing… I almost remember there being some mod that did almost exactly that.

None of what Nightdive achieved couldn’t have been done by someone else. In fact, arguably, the darkplaces engine is far better in terms of added graphical features. It has true model interpolation, unlike Kex. And sure it is unfair to compare a game engine that is the product of one fan working in their off-time, to an entire game studio that specialises in remasters, so I should perhaps shut up. As someone who worked on both darkplaces and vkquake and the rtx version of YamagiQuake, I should probably do the opposite.

There is a lot that is lost. We have taken something that had largely become a fan-made fan-supported work, and turned it coporate again. Despite the perception that corporate interference can make things better, it did not. We grew up under the perception that corporate has more resources, and all they lacked was true passion, but that is not the case. Corporate values and guidelines spoil the product, no matter if there was passion and resources.

What is interesting is that there is a way out of it. Ironwail a port that I admire quite a bit. It is a particularly interesting source port: it can take advantage of the few goodies that were introduced in the remaster, but is still nonetheless a fine regular vanilla experience. It takes advantage of the new content, an easy way to retrieve the complete vanilla assets, but not much else. Just use modern OpenGL to improve the frame-rate, add mouse look and allow the user to bind more keys to more functions and you’re golden.

What’s interesting is that Ironwail is made by a single programmer. This doesn’t necessarily needs to lead to a conclusion on your part that Nightdive programmers are incompetent. Just that the approach of taking an old game and putting it into a new engine is often more trouble than just fixing the old (and well made) game engine.

Black messa Link to heading

There are two games that were in the zeitgeist of the Half life community. They are both an interesting re-imagining of the situation at the Black Mesa research complex, with an interesting shift in perspective on the already known protagonist, that was overshadowed by technical limitations, misguided use of assets, a story which is only tolerable because of its link to the original Half Life, spotty voice acting largely due to poor direction, and occasionally obtuse level design, failing to update that which needed to be changed, and changing what was fine. While Black Mesa, being the less interesting of the two, is considered a “faithful adaptation”. Hunt Down the Freeman, is considered a failure, despite there being quite a few positives. The game attempts to push the boundaries of both the Source engine and the Half Life narrative.

So what would be a fair assessment in this case? Firstly, I must say that I differentiate between a free mod made by people who are just learning how to design games, from what was released as a fully paid game in 2020. Once you charge for your work, the standard to which you are held is different. What I would call one of the best mods of 2009, I would consider a mediocre game, that largely coasts on the works of prior giants.

Let’s begin by observing that games are an interactive medium. Black Mesa is what one would imagine playing through Half Life 1 maps with the Half Life 2 Gordon to be, if he were fighting Half Life 2 combine instead of the HECU, Half Life 2 headcrabs instead of Half Life 1 ones, antlion workers instead of bullsquid, and the HL2 vortigaunts voiced by Barney suddenly became hostile, and you had female scientists.

The problem is that Half Life 2 and Half Life 1 are radically different games. They are not more of the same, despite many of the macro elements of Half Life 1 carrying over to Half Life 2. This is something that the Half Life 2 team is acutely aware of, hence why a considerable amount of work now known as the Half Life 2 beta is cut. Half Life 2 is a game about war in a variety of locations. Half Life 1 is a science fiction novel of the time period done in an interactive medium. So the engine sensibilities of Half Life 2, do not map well onto Half Life 1.

First, the level design. Half Life is easy to navigate compared to many of its contemporaries, but a lot of lessons were learned during its production. Half Life 2 is much easier to parse, as it is much more attentive towards the player. It knows how to nudge the player in the right direction. HDTF is a game where this concept is just not recognised. Black Mesa makes some attempt to follow the same rules, but often falls short. Case in point, “Power Up” is a chapter where Black Mesa recognised the importance of directing the player’s attention. It fumbled the fact that no sane player would go towards the place of origin of Gargantua, and why Valve opened the door next to Gargantua instead. I am a fan of the moodier lighting, but not sold on the changes made to Residue Processing.

Why cut reasonably good puzzles in On a Rail, but keep the frustrating platforming and even more questionable signposting. They started with perfectly good levels and ended up with a clear case of Hunt-down-the-Freeman-itis, where you are lead on by a fleeing guard into a complete dead end. The only way to find the right way forward without using a guide is to get frustrated and search every inch, before backtracking all the way to the point of no return and noticing some peculiarity in the architecture. If like Hunt Down the Freeman, these same level designers had no template to start from, there’s a good chance that the end result would have been comparable.

The original Half Life is a game that doesn’t exactly take its premise too seriously. Gordon is a glorified lab rat. He’s probably not too bright because most of the career theoreticians that I’ve met can’t pull off half the acrobatic manoeuvres that Gordon can. In America, Gordon being able to proficiently reload an MP5 is probably not too much of a stretch, but the amount of weapons that he carries is beyond the realm of plausibility. It’s not a super-realistic game, and so the gun-play veers much closer to fun. Crowbar Collective seem not to have gotten the memo.

Obviously they could have tried a few of the things that valve themselves experimented with: they could have tried introducing a three-four weapon limit, or one weapon per-category, (which Hunt Down the Freeman did, but not too successfully). They could have experimented with making the game more systemic, so that satchels and mines are more of a necessity (which is kinda what the original Half Life was all about). They could have embraced the goofy nature of the game, and perhaps make the game a bit more fun, with quick swapping weapons, more responsive gun-play, and maybe remembering that the Spas-12 is primarily an automatic shotgun. Instead, you get a game, where explosives are less reliable due to the clutter in the levels, the guns are downgraded, you get iron-sights, but only for the 357 magnum, you lug the entire arsenal with you, despite using most of that arsenal being suicide for most of the rest of the game, and things like Rocket launchers being useless for most of the game.

Even the justification is only on paper. There were people who asked if iron-sights could be implemented for all guns (and I admit it’s probably a controversial topic, but my personal opinion is that it would make the gun-play a bit more immersive), they stated that they didn’t want to make later guns like the MP5 overshadow the earlier guns like the Glock. If you played the game, you probably know that the MP5 is many things but not less accurate than the random spread slow firing rate weapon that is only marginally better than the crowbar at killing headcrabs, if you haven’t nailed the timing. This wasn’t the case in the original Half Life, both because the damage output for the 9mm was respectable, and also because you couldn’t one-hit-kill headcrabs. The fundamental misunderstanding here is that in Half Life, a 90s shooter, the main enemies, the HECU grunts, couldn’t shoot and move at the same time. This has profound implications, because it made it plausible that one could shoot a grunt at a range where the Glock’s increased accuracy gave it enough of an advantage to take down grunts ammo-efficiently.

Instead, in Black Mesa’s balance, the Glock is a useless weapon. The MP5 overshadows everything else, because it is both accurate (given you understand the recoil doesn’t add random spread), has abundant ammunition, and is the only weapon that isn’t complete suicide to use against grunts. The shotgun doesn’t reliably kill HECU, doesn’t reliably kill vorts, and has the firing rate of a real PhD in Physics using a real pump action shotgun, in a game balanced around the MP5. And you’re telling me, that the reason why you only have iron-sights for one gun, is because you think it would ruin the balance of your game?

What’s interesting is that Hunt Down the Freeman is actually not that terrible in that respect. They have an inventory system, and you get one gun per category. In the newest update the gun you get at the beginning of the level, is usually the right tool for the job. And they tried something. They didn’t do it well, but at least they tried. Black Mesa simply needed to follow the old encounter design, and they managed to make surgical adjustments that made the balance worse. This isn’t exactly’ crowbar’s fault, they took the combine, and adapted them to HECU. The only thing similar between the marines and the combine soldiers is that they are bipedal and use standard humanoid weaponry. They couldn’t be further apart from each other. Making them work, would veer the weapon balancing much closer to the Half Life 2 balancing, and to some extent that had been attempted, Half Life 2 does have a useless pistol. But it was not done well, because the shotgun is worse, the SMG is much better, and the energy weapons don’t exactly wipe the floor with the rest of the guns, even when you have the ammo for them.

And then there is the tone deafness. On its own the soundtrack for Black Mesa is a wonderful piece of art. I listen to it, when I don’t play the game. But in context of where it is used, it doesn’t fit at all. Crowbar collective fundamentally misunderstands Kelly Bailey’s soundtrack, it’s not meant to evoke any particular emotion, it is a texture that is ambiguous and doesn’t dictate what you should feel. Adrenaline horror suits a triumphant Gordon pulling off cool stunts just as well as a scared Gordon being shot at from mortars cowering in the corner. Sirens in the distance is an ambient cosmic track. Space Ocean fits the feeling of emptiness that the early Gordon that just came out of the test chamber feels. And it suits the exploration well. Drums and Riffs doesn’t exactly convey anything beyond Gordon is moving somewhere. The only part that is remotely similar to what Joel Nielsen has done is Diabolical Adrenaline Guitar. This leads to the music being largely in the background and encounter design dictating the mood. At times, it’s possible to forget that Half Life does actually have a soundtrack, and that parts of what seemed to be the ambient atmosphere are actually music. The few times the Music does get your adrenaline pumping, so to speak, are the situations where you are on the surface, and facing impossible odds in a warzone.

Contrast this with how music is used in Black Mesa. Blast Pit is sad music that plays when you burn down blind tentacle monster. You should be moving slowly and quietly, instead the music pushes you to move fast. Questionable ethics makes you feel like a human tank. That is the most surefire way to get killed and reset. Forget about freeman is used roughly as well as Nuclear Mission Jam, but has a decidedly faster rhythm for a decidedly slower encounter. Surface Tension 4 is just baffling, to be quite honest. It’d fit the scenes of a warzone, because then at least it would convey that you are, in fact, witnessing your own kind being decimated by an alien force. Instead the track is used for a section where the player is likely to want to think and move slowly. Apprehension perhaps fits somewhat. Though sadly, its best part simply wouldn’t fit the timing. The music is not universally a bad fit, but it needed a lot more polish to fit well. If the game were slightly different, sure, all of these tracks would have fit, especially if parts of them were looped, parts were cut out, and parts were made part of the ambient sound. As it stands, there is a considerable shortfall between what is and what could have been and what is.

Now do I hate this game in general? No. It’s a fine game. Not perfect, derivative and unpolished, but still something that I wanted to play on my Steam Deck, and something I occasionally replay. However, there is a worrying trend of people stating that Black Mesa is the ultimate way of experiencing the Half Life story. Do I really need to elaborate on why that assertion is preposterous?

Even if the game didn’t have glaring tonal and gameplay problems, the fact of the matter is, the game is not the original Half Life. The best way to experience that game, is to experience the exact way it was designed. Cartooned graphics, implausible premise, MP5s that have 50 round magazines, and theoretical physicists that are fitter than marines are less dissonant cognitively speaking, if the implausibility follows a certain common background level, rather than having spikes in some few areas. Plus a few of these things are forgivable, given the time period. Half Life came out at a when “par for the course” was very different. Games were considered a far less serious medium. So while some occasional problems are forgivable for Half Life in the 90s, the perspective is different in the 2020s.

Crowbar collective is largely aware of this issue, which is likely why they severely shortened “On a Rail”, and completely restructured the border world chapters. The fact that most scientists have the same voice, and the one exception, might as well say “Stanley was a highly trained professional so he didn’t need to hear all this”, is kind of a problem. Exasperating this problem is that the voice actor for Barney and the G-Man, does not have the range of Mike Shapiro. The choice to have multiple characters still be voiced by him in Half Life 2 was also accompanied by the citizens being voiced by someone else, and Alyx and Judith and Breen and Eli being voiced by people who don’t play other parts. There was even a father Grigori who wasn’t even meant to be an American. And it would make sense for Black Mesa to have more than just Americans, wouldn’t it.

So where does that leave us? Black Mesa serves only one purpose. It is an alternative take that serves to remind us what made the original iconic. It is not competitive with the original Half Life; it is not meant to be. What I can clearly say that it is a game that should exist, and it may have been worse off had it not had the guidance of the original Half Life’s script and level cohesion. I believe that the team behind Hunt Down the Freeman was equally talented, but since they did not have the same creative direction as Black Mesa, they were at a disadvantage, and their failings are more obvious. So in effect, a game that tried to do more, and tried to be its own thing, was punished for it.

What is even more frustrating is the question of how much was added by Crowbar collective? The end product is a 90% good game, 80% of which is just Valve’s ideas, repackaged. Does Crowbar deserve credit for the difference between Half Life: Source and Black Mesa, or the difference between Black Mesa and Half Life? These are not questions I believe have definitive answers, but questioning these premises will lead you to question the premise of remasters in general.

Homeworld 2 Link to heading

You might think that if Homeworld 2 was a step down from Homeworld 1, that the remaster would have been kinder to the misshapen sequel? After all, Homeworld 1 had only technical issues to fix, and the remaster’s approach was a step towards the “fixes” from the sequel, and that largely is the reason it’s worse. Well, they managed to maime the second game just as much.

Homeworld 1, is an epic-scale tale of zionism. You can equally play as the Kushan and the Taiidan. The factions are interchangeable, with perhaps the gameplay being the main and only differentiating factor, the Taiidan are weaker when in head-to-head combat. A means of creating a difficulty curve for replayability. It is a new-game+, that dovetails nicely into the narrative; violence begets violence, and after overthrowing the Taiidan, the Kushan became exactly like them. The triumph and a sort of poetic justic: exile of the Taiidan to the Kharak system, now plays out in reverse, with you as the former bad guy, receiving the short end of the stick. This is pure genius.

Homeworld 2, eschews that narrative, and is much more of a traditional hero’s journey. As such, one wanted to have as much creative freedom as is conceivable. It is implied that the Pride of Hiigara is a newer ship, built by a culture that though inspired by the Kushan, had a lot of Taiidan influence. The Vaygr could have equally descended from the Kushan as the Taiidan, though the iconic wing-shaped mothership and the fact that the Kushan were the intended original protagonist implied that while they are a mix of both, the narrative of the first game shouldn’t be taken as gospel.

As such, the entirety of the Hiigaran fleet is re-cast. And that, perhaps, is the greatest advantage conferred onto Homeworld 2, it eschews comparison to the first game as much as possible. It would have strayed even further if it could, perhaps by inventing completely new factions. Though technically speaking, the Hiigarans are implied to be a very different people to the Kushan. Note that the engine exhaust, the style of the ships, the colour of the ion-cannon beams, just as much as the choice to continue using the ion-cannon technology, is in-line with the Hiigarans having intermingled with the Taiidan. It is also possible that the Vaygr are a descendant of the Taiidan in the same way, absorbing much of the Kushan lore. Both, believe in Sajuk, which to paints a historical face on both: they are turkic. And much like the Seljuk Turks, they are a bloody tirant that took over a vast space, and assimilated much of the local culture, while leaving a turkic imprint. The new voices suit this extremely well.

The Sajuk Khar (which if you paid attention already hints at Sajuk being a ship). Now it does make sense that since canonically, Karan S’Jet is again the fleet command, re-casting her was probably done because of budgetary reasons, and indeed that was the original intention. It makes sense to re-record the lines with the actress that portrayed her in the first game from Heidi Ernest as opposed to Jennifer Dawne Graveness, that did the work for Homeworld 2. Jennifer Graveness did an amazing job, but it is, for lack of a better word, expected. It reduces the cognitive dissonance. It makes sense, particularly given that Heidi Ernest has aged. She is the more experienced fleet command.

The choice to replace the fleet intelligence voice is less understandable to me. There is an in-universe reason why Eli Gabay would be supersed Michael Sunczyk. They are different people who have the same post. Say what you will about whom you prefer, but the voice lines suit Eli Gabay much more than they do Michael Sunczyk. That’s just the consequence of the change in writing between the two games. Yet, for some inexplicable reason, Gabay is out, and Sunczyk is in. I may be in the minority of people that prefer Gabay’s delivery. Particularly, this line Bring Sajuuk to bear, sounds decidedly better if delivered by a calm, but quick fleet intelligence, as opposed to what we see in the remaster. Sunczyk is a competent actor, but he doesn’t have the range necessary for that phrase to sound even remotely convincing. To be fair, this missions isn’t a particular high-point for the franchise. The line itself was out of place, even in Homeworld 2. The reasonable thing to do here would have been to adapt the screenplay to the fleet intelligence character that Sunczyk played. But in an attempt to fix what can only be described as a blunder in the original, was compounded by the second.

Now, obviously this is nothing to say of the gameplay. There are differences, and listing all of the minor changes for good or ill would take me forever. What I will say is that the duology of Homeworld 1 and Homeworld 2 are the finest examples of remasters either changing too much or changing too little. Having kept Homeworld 1 as the baseline, and adapting its engine to a larger scale conflict in Homeworld 2, keeping the voice cast as it was in Homeworld 1, and in-keeping with the design intentions of Homeworld 1, would have made Homeworld 2 a much better experience. Similarly, just creating a third game instead of trying to remaster the first two would have worked equally as well, eventually that is what has happened, but the game titled Hardware Shipbreakers should have been released before the remastered collection.

System Shock Link to heading

This is an interesting review. Of all the games that are on this list, System Shock is not one from my past. I have actually played through the remaster first.

And here’s my overall impression with the game: it’s average. Not painfully average, not below average, just average. It is a faithful remake of a boring game, whose main influence was in ideas that were better implemented in later games.

I don’t mean to sound rude, but System Shock 1, is a historical curiosity. It inspired many games, true, but on its own it is decidedly baroque. The mechanics of having the story be told via audio logs, for example, is a common feature that can be traced back to System Shock 1, but the fact that your audio logs are the only source of information on where to go next, and often contain acting so bad, that it’d be better if they were purely textual.

As another example, cybernetic enhancements as a tool to gate off sections of the current level, incentivising limited and purposeful back-tracking, as well as an in-universe mechanic for death are all features that debuted in the original System Shock. And while the original game was a little obtuse, it could be forgiven that that game is older than I am by a good year. Its rough edges are forgivable. But with all of that, it is a game that did almost as well as it should have, it didn’t exactly fail, but it wasn’t topping any charts.

At the same time, the remake is in a precarious position. If it changes too little, it will be criticised for the archaic game mechanics, even if those are improved. If it changes too much, is it really a System Shock remake anymore?

The position that Nightdive found themselves in, is difficult and largely a self-imposed limitation. Nightdive could have equally chosen to create their own game. They could have spun off from the original, given that securing the rights to a remake would have been just as involved as getting a licence to do an original story. They could have built an immersive sim in the spirit of System Shock 2 and Deus Ex, which were an undeniable refinement of the original System Shock formula. What they have decided to do instead is to remake a game that could still run, in a manner in which, I, as someone who has never experienced the original System Shock, considered never experiencing it at all. The only reason why I overcame that desire to quit, was precisely because of previous experience with your remasters.

My exact thought process went like: “OK, I’ve seen what they did with Blood, so the original game can’t be as bad as this”. And I was right. The original System Shock 1 is a game that I actually managed to finish. While I will say that playing this game was quite a headache, given the accessibility problems, and the fact that it is an old game with known technical limitations on newer systems, all I needed really was this wonderful source port: https://github.com/Interrupt/systemshock, and getting the game data into proper file structure. The enhanced edition, also made by nightdive, is not as bad as the others, so I will confess, it is a good way to experience that game too. The remaster, though…

So the Nosedive remaster, adds nothing to my experience of the game. It is a nice option to have for other people, and I don’t want to spoil their fun, but for me personally, it is a wolf in sheep’s clothing. It is a way of extracting further profit, without really changing much about the underlying product. It would be difficult to update System Shock 1, in a manner in which the changes would be so minimal as to please the old guard, while simultaneously updating a rather dated experience. Why bother with a remaster then? Well, I can only speculate. it is my humble opinion that the answer is blatantly obvious: if nightdive made their own game, even if that game were superior to System Shock 1 in every way, it would have less of a legendary status. Further, when remaking System Shock 1, because of the sheer age of the original, the newer game would be held to a much higher standard.

The main problem is, perhaps, that System Shock 1, came out at a time when video games were still an experimental medium. One had to contend with polygon count limitations, nobody really had a convenient system for designing levels. The “crazy NeXT machines” are laughably weak, and compared to the work that they were meant to do, underpowered by today’s standards. Games were simple, because it was hard to make anything complex. As such, you as the player had to fill in the blanks quite a bit.

We had to use our imagination in games from that epoch and for a long time afterwards. Consequently, there was economy of information: every object in a level served a purpose, not unlike Chekhov’s gun. You had a couple of props to convey the purpose of a location, like barrels stuffed with meat in Blood, or seats in the adult theatre in Duke Nukem… or a couple of potted plants in Wolf 3d. Elements meant for interaction had less competition on screen, and levels were more readable. This design ethos perists today: Half Life 2, for example, uses a different model for doors that are meant to be opened, to doors that are set dressing meant to convey the purpose of a structure in a level.

In System Shock 1, the lack of clutter guided the player towards solutions to the level progression. The reason I refrain from calling them puzzles is because they are straightforward navigational obstacles. You’re meant to read the logs, and figure out where to go, even if brute-force exploration can do the trick. There’s one instance of the player being punished for not paying attention, but overall, simply walking forward, exploring new decks, and pressing buttons meant progress.

System Shock the remake, has the same intention; you don’t have a dotted line to follow, or an objectives menu. Your journal is a collection of the information that you’ve gathered, not a TODO-list. The HUD exists for the player character, as well as the player. But these design elements are anachronistic in an Unreal Engine 4 game, that came out after Mankind Divided. Sure, it is immersion-breaking to have a pop-up that states: your player character knows the combination to that lock, here it is for you, to role-play as if you remembered it. This has been the case in modern imsims. The remaster doesn’t exactly have this; it wants you to remember the combination, or to look it up in the unsorted, ungodly pile of logs that you’ve seen or heard. The belief is that the original System Shock didn’t have this, because it would be more immersive to sift through logs than it is to have a magical popup. Back in the day, it would be, because even having combination locks that worked, was itself an achievement and a hallmark of great programmers working on the game. Today – obnoxious design decision is probably the best way to summarise. The player is more likely to create the missing TODO list and record the key combinations externally. You could argue that that is immersion, and perhaps for you it may be.

Later imsims have had a system wherein some code locks had randomised codes, so that you actually had to honestly engage with them, but at the same time, you had options. In Dishonored 2, I had a lot of fun figuring out Jindosh’s lock. My Corvo Attano, was a scholarly man that preferred the work of the mind to the work of the blade. I could have solved that puzzle with prolog in a couple of seconds, but that was as much of a choice as brute forcing the lock combination, which I could do too. But that was one of very few combinations of that sort, in a game with painterly environments, that one wanted to explore. Randomising some combination locks in a game as ugly as the System Shock remake, is asking for the player to optimise away the fun.

System Shock 1 pioneered a game genre. It has had an environment that was trendy (at the time), with a novel premise, though based in literary works, where it is less fresh, the idea was new, and a great fit for the setting of a computer game. The audio logs were new, expensive, and vital for the immersion, because it was uncommon for games that came out in that era to have a significant amount of voiced characters. The villain was a wonderfully acted delightfully devilish goddess, who intervened quite frequently, and was a constant presence throughout the game, compared to its contemporaries. Taking out cameras to hinder her was a wonderful idea and a fresh gameplay element. It is old, but given the constant reminder of its age, you see it as an early work of the renaissance – not as good as the later ones, but historically significant, thus interesting.

The System Shock remake is an annoying nostalgia bait. Its setting is done to death, and amongst its many peers, this Unreal Engine 4 game distinguishes itself by being ugly, and poorly optimised. The premise: AI takes over station is as stale as humanly possible in the early 2020’s when LLM craze and AI panic were already tiresome. The Frankenstein complex was considered a literary cliche by someone who died in 1992, so your story is at a great disadvantage, even if it weren’t overrepresented in computer games. Voiced characters are the norm, audio logs are used when voiced interactive non-player characters would have been too difficult to implement, and are thus considered the cheap option. This is not helped by the poor voice acting that may be a slight improvement over what the original System Shock had. The villain is wonderfully acted, delightfully devilish transparent attempt at creating the illusion of omnipresence. Shodan can bark all she wants, the scripted events are transparently scripted, and you know for a fact that she can’t do much. Unlike e.g. Left 4 Dead, there is no intensity director, no guiding force to keep the player on edge at all times, just an annoying AI with delusions of grandeur.

The remake, simply exposes how much the formula had been refined since its inception. System Shock was influential, it spawned many offspring and its design elements can be found in disparate genres. Doom 3 is arguably a System Shock-like. Bioshock, Prey and Deus Ex are all direct descendants. Dead Space is a 3-rd person System Shock, with refined tools (kinesis and stasis are given the much needed depth, which System Shock’s cyberware lacks); deeper, much more terrifying villain, much improved atmospherel; sound design that has so far not been suprassed; believable acting in audio logs, so much so that one can point to some instances as examples of what voice acting should be. These games elevate the building blocks first shown in the original, much like the paintings of the renaissance elevate the ancient Greek art.

During the long years that Nightdive was trying to remake System Shock, it could have produced yet another imsim. They chose instead to take the road most travelled, and ride on the wave of nostalgia. What I find most frustrating is that they largely succeeded. Nostalgia works. Many critics will concede that the remake has a lot of good and many new things, but in fine print only, mention that what is new is not good, and what is good is not new. Yet the consensus is that as in the proverb, this dwarf resting on the shoulders of giants, nonetheless sees further. If this game were not a System Shock remake, but a standalone experience, legally distinct from System Shock, it would have been lambasted.

The “faithfulness” of the remake is but a shield. Why is the game ugly? They “preserved” the art style of the original. Why is the plot so dry? They “preserved” the plot, and were afraid of mishandling the characters. Why is the voice acting so bad in most of these logs? They wanted to keep it authentic to the original. Why is the gameplay so clunky? The original was. Why is the music so miss and miss? Development hell. John Perros is a fine musician. I quite like his earlier works, even those that are attached to the earlier versions of the remake. I can only guess that the self-inflicted problems associated to remakes, created an unfavourable atmosphere for creativity.

It didn’t have to be a remaster! But making a remaster paid off. We bought the game. I bought the game. I didn’t even bother to ask for a refund. I just published a review which was somewhat negative, and was awarded the clown “trophy”.

The reason why remakes such as this one keep happening, is because we’re all sold on the idea that modern technologies can improve a pivotal work of art. To any reasonable human, the question of remaking the Mona Lisa with modern technologies should sound preposterous. It would also seem natural that there are a multitude of fan-art which features or references the Mona Lisa, with varying degrees of success. Yet for some reason with games, the situation is different. Only one studio is allowed to make a remake, but you must not dare judge it by modern day standards, because it is an older work, and dare not judge it by the old standards either, because the work is new. Well which is it?!

The best way to experience System Shock 1, is by playing something which preserves the charm of the original; it could be GPLv3 source port, which I would prefer, or it could be the Enhanced edition, because it is a convenient package. But to me it is dubious, whether someone playing through the System Shock remake, has really experienced System Shock.

Homeworld 1 Link to heading

This is probably the most straightforward case of a publisher wanting to cash in on DMCA exclusivity. The game would have been remade, and has been. Consider that it is now possible to play that game from within a browser. The reason why you can do that, is because Relic has released the source code for the original Homeworld in a fashion, in which it is possible to create derivative works for non-commercial purposes. That basically means that if the game grows completely incompatible with your OS, or CPU architecture, you could simply recompile it. It means that I could play through the original Homeworld 1, with original assets and only remove the problems which cause me headache, namely the resolutions problems. The rest of the game engine remains completely intact, and given that Homeworld 1 has perhaps the best rendition of the game engine, with the best gameplay mechanics, is probably the optimal solution. Obviously a reverse-engineered port is also possible, and given that LLMs do not produce (allegedly) derivative works, even if they were trained on code that includes the original source code, it is fairly reasonable to state that this game engine, could, and probably would become the basis for a large-scale space 3d strategy game.

Instead, we got a by-the-numbers rebastardisation. The dialogue was re-recorded, and that meant that the subtle distortion effects and compression artifacts (that in my opinion enhanced the experience) were gone. Some of the original actors were replaced, but they found similar-enough sounding voices. The visuals got an uplift, and so, instead of relying on your imagination to fill in the blanks, you have modern technologies to ensure that your game couldn’t be played on a weaker device. But hey, now the tiny ion-cannon beam has a subtle (as in you’d have to squint to see it), effect on the ship’s hull. There’s a bunch of motion blur and bloom and volumetric lighting, that isn’t exactly volumetric-enough to fool you, and admittedly, the depth of field effect is good.

But the game’s mechanics have been reduced to almost nothing. There’s no fuel mechanic on fighters. There’s a difference in how the motherships produce fleets, even though technically, the rest of the game’s balancing is off. The line-of-sight simulation-based projectiles are replaced with randomised coin-toss based mechanics.

How is this an improvement?

Well, the idea was to take the best of both worlds, to take up the mechanics that were introduced in Homeworld 2, and they took the less egregious ones (I would imagine the fanbase would be in uproar if the squadron mechanic was ported too). The assumption is that if the team behind Homeworld 2 made this “improvement”, it must have been a good thing. The reality is much different, and much more grim. Homeworld 2 was an experiment, that largely killed the franchise. If anything it isn’t even well-regarded amongst fans, because it did far too many things far too differently, and not necessarily better.

What could you expect from Gearbox software?

Conclusion Link to heading

I find that remakes are largely unnecessary.

Creating games set in established universes is meant to cash in on nostalgia, and if the creators are given much creative licence, these sorts of games can be good proving grounds for new ideas. Black Mesa could have been one of a myriad of mods that work on top of the Source engine. Other remasters could have similarly done something more authentic. More new.

With very few exceptions, passionate fans, whose nostalgia keeps the games alive, do a better job at preserving and improving upon the experience. Remasters cost money and often do a worse job.