Tronyn interviews John Romero

Tronyn, currently engaged in an Epic Quest through vast wilderness, sent the following parchment via carrier falcon:



Ten years ago, for Quake’s ten year anniversary, you did a couple of interviews: one with The Escapist ( ), and another for the 2006 Quake Expo (Archived at  ). One looked at game design issues in general and your career, and the other focused on the history of early Quake. I want to focus instead on level design specifically. I also want to keep the interview short (five questions) as I know you are working on a new FPS project, Blackroom.


Question 1: Atmosphere

Errant Signal made a video about Quake which contains some astute commentary on the game’s level design.

He calls Quake “a game about texture and tone… a mood piece,” and says that like Reznor’s music, Quake captures the sense of “feeling spurned by a hostile reality only to seek refuge in a sort of disconnected depressive state.” I get the impression that the aesthetics of the game, the element of psychological immersion and of getting an actual reaction or response from the player through the environment – that this is an element that always mattered a lot to you. How did you set out creating atmosphere? Do you think atmosphere can create an emotional response in players, as opposed to using storyline? How important do you think atmosphere is to level design?

ROMERO: For Quake, from the beginning, the atmosphere of the game was to be unsettling and violent. I didn’t want the player’s view to stop moving even when they stopped their avatar’s movement, kind of like slight breathing all the time, keeping the 320×200 pixels always looking different, just a little bit. That feature is still in the game but is only used during the intermission camera and is exaggerated there. I definitely believe the environment and atmosphere are all about telling the player where they are, what the tone of the gameplay will be, and how they should feel. The atmosphere, in level design, is incredibly important to conveying the feeling that the designer is trying to impress. To help create the atmosphere in my levels, first we needed to have our color palette set to muted, dark colors. All the art was created with this palette. Then, while I was making the levels I would have the indirect incandescent lights in my office on very dim and I’d play epic-sounding Queensryche songs from The Warning (No Sanctuary and Roads to Madness were faves), and other metal that sounded similarly epic (Marching Into Oblivion album by Darkstar). As I was making my levels they had to match the songs that were playing. I was forcing the feeling by controlling my creative environment.


Question 2: Sense of Place
I just want to quote Errant Signal’s commentary one last time, as he refers to Quake’s levels as “abstract constructs designed to facilitate play” and “Random cyber-goth corridors,” as opposed to levels based on real world locations. I’m not sure if this is accurate: some of the levels in Doom and Quake (and later, Daikatana) absolutely have identifiable models in the real world: castles, cathedrals or graveyards, although some of the levels that are dungeons, caverns, or military bases might be a bit “Random” or “cyber-goth.” Do you think it’s better to build levels which represent a realistic or contemporary type of setting, or do you think levels with a vaguer or less easily identifiable sense of place allow for more creative possibilities?

ROMERO: My preferred type of level design is Abstract. Realistic design is not as interesting in my opinion. I would rather explore a place that is unexpected, unconventional, and highly creative. Abstract level design permits all of this. Realistic locational design puts so many restrictions on what you can do that I don’t ever go there. I have no problem designing a level around a known archetype, and it’s really fun to see where I can take it, but the idea of the archetype and incorporating known elements are the only restrictions I like on me.


Question 3: Layouts

In the beginning, FPS level design actually came from top-down arcade games, where the levels were depicted from above as two dimensional “maps.” An early id FPS, Catacomb 3d, plays like its 2d top-down predecessors, except for taking place from the first person perspective, and even Doom still retained the 2d map. Yet a lot of FPS level design these days, perhaps owing to the influence of the console market, doesn’t seem to take advantage of the possibilities in 3d level design; there are modern games with levels whose layouts could be recreated in Doom. In contrast, you and your fellow artists were eager to exploit new possibilities as the technology became available; in the Quake Expo interview, you mentioned having to wait for the technology to catch up to the type of levels you wanted to create (i.e. three dimensional). It seems like a common theme of your career, particularly early on, of having to wait for technology to catch up to a vision or idea you have you in your head. Will your future projects draw on the style of level design that you’ve established, or will you move on to trying something totally different? What do you think the next frontier of level design should be?


ROMERO: One way that I helped to enforce our progressive level design development was to come up with rules to follow. At the beginning during Wolfenstein 3D development I was having vague thoughts about what made a level fun. Then, when DOOM development was upon us, Tom and I were making levels that looked like Wolfenstein. I had to focus to create a completely different way of designing levels that really stretched the engine, and created a rule: no location in a level can look like it could be made in Wolfenstein’s tech. When we began working on Quake, the same rule applied: no location in a level can look like it could be made in DOOM’s tech. I told the designers, “Look UP. If there’s nothing interesting up there, put something up there.” It didn’t take long for everyone to get it and run with it.  My new FPS, BLACKROOM, is my first FPS in 16 years. I will be exploring level design in a new way. It has to accommodate intense gunplay and skillful movement, but it also needs to do NEW things that are unexpected and really interesting. The game’s fiction enables me to do whatever I can imagine so my mind really is the limit. Well, also the player’s capacity to deal with it is the real limiting factor. I will only know during playtesting if what I’m going to do is something that players will find fun and interesting.


Question 4: Routes and Navigation
Some of the later Doom maps, particularly those in Thy Flesh Consumed such as e4m2: “Perfect Hatred,” really pushed the gameplay possibilities – requiring the player to “jump” across lava pits, including one area with a steep staircase surrounded by a lava abyss which the player must ascend while a cyberdemon (!!) can at least attempt to fire at him or her. Similar scenarios occurred throughout this very challenging episode, and I assume you felt comfortable designing levels like this (I’ve heard you designed “Perfect Hatred” in a single night!) because by this time you knew there were hardcore Doom players who could take whatever you threw at them. However, many Quake maps also challenged players with traps, moving parts, and objects players must manipulate. Nothing was explained, players were just trusted to figure it out (or die or quit). Another great example of this was the Quake secret level, e1m8: “Ziggurat Vertigo,” which introduced the player to a low gravity environment where the journey was straight up. Other levels made significant use of water or, like your e2m2: “The Ogre Citadel,” gave the
player multiple route choices. Obviously, it can be frustrating to get lost in a game, but Quake trusted players to figure things out; this greatly contrasts with modern games, as a video called “If Quake Was Done Today” notes.

Do you think there is still room for challenging FPS level design that forces players to figure things out on their own?


ROMERO: Absolutely. And I plan on designing my new game with the same expectations players had when they began playing Quake and Half-Life 2. Players are smart and find handholding offensive. Putting players in interesting situations is an important part of keeping players continually excited about your game. Keeping a player from getting lost is partially up to good level design, but my new game will have more ways of helping the player stay on track.


Question 5: The Implied World

Based on your work on episodes like Knee Deep in the Dead, Thy Flesh Consumed and The Realm of Black Magic, you strike me as a master of the episodic format – of envisioning a series of thematically connected environments that, taken as a whole, add up to a coherent vision of some other world – whether it’s a rotting medieval wasteland or a Martian moonbase. A lot of your designs also hint intuitively at a setting more expansive than what the player actually sees. In the first level of Doom, you immediately get the armour at the top of a stairway and see out the window onto the mountains of the Martian moon, which fade into the misty distance. Tolkien actually used the same metaphor of seeing an inaccessible mountain in the distance as a way of giving imaginary worlds the illusion of depth. When designing these levels, did you have any overall idea of an “implied world” in mind (like the win-screen map for Knee Deep in the Dead, which shows the levels as different structures nestled in a crater) or were the levels just strung together based on how difficult they were (easiest first, hardest last)? How did you go about setting up episodes?
ROMERO: Definitely, the entire idea of having a sky in DOOM was to imply you were in an interesting location that there was more out there to explore, if only you could get out there. The plan was for the player to travel through the UAC base to get to the location of the portal, so I created levels that increased in difficulty and length as the player learned more about the game. For episodes 2 and 3, Sandy Petersen decided how he was going to create those levels as some were based on Tom Hall’s scraps, and Sandy knew back then that difficulty ramping was important but that the player already went through episode one so he didn’t have to go too lightly on them. The interesting thing is that back in 1993, everything we were doing was new and even having a few enemies in front of you was kind of crazy and difficult. Today it’s not even challenging but back then it was the first time we saw it. The difficulty was appropriate for 1993.


  1. Yeah that was a nice read (the questions as well as the replies), and Romero is a cool dude for participating. Funny that he still remembers the specific albums he mapped to. 🙂

Comments are closed.