Gamasutra: The Art & Business of Making Gamesspacer
View All     RSS
October 25, 2014
arrowPress Releases
October 25, 2014
PR Newswire
View All
View All     Submit Event

If you enjoy reading this site, you might also want to check out these UBM Tech sites:

Why Audio Guys Use "Nice" Speakers
by Mark Kilborn on 06/09/10 02:11:00 am   Expert Blogs   Featured Blogs

The following blog post, unless otherwise noted, was written by a member of Gamasutra’s community.
The thoughts and opinions expressed are those of the writer and not Gamasutra or its parent company.

There's a common misconception regarding audio people in the industry: we're making games to sound great on big, bad surround systems, like the ones in our offices, and we don't care about what the game will sound like in the "real world." This is simply wrong. After hearing it again yesterday, I've decided it's time to explain.

The "nice" audio setup you see in a lot of pro audio studios is not meant to be a nice system. It's meant to be an accurate, calibrated system. Most consumer sound systems are hyped, meaning they have uneven frequency response. This means bassier sounds may be boosted, higher sounds may be reduced, etc, all in the name of creating a more exciting sound. A visual analogy would be if the maker of a video monitor toyed with the color balance, brightness, contrast, etc before they shipped the thing out of the factory.

Problem #1, these frequency range boosts/cuts can't be changed. These are nuances that are inherent to the design of the speakers themselves. You may have an EQ on your home sound system, but if your speakers are consumer grade they will be coloring the sound to some extent.

Problem #2, all companies do this differently. Different speakers will have different response curves, whether they be in televisions, consumer 5.1 systems, etc.

Professional audio monitors (when audio people say monitors, we're talking about speakers) are designed to be as flat as possible. Since we can't predict what the situation will be on a consumer's system, the best approach is to get the flattest frequency response possible and work from that. I often see artists calibrating their monitor so they can "see what's REALLY there." We're doing the same thing.

Any audio guy with experience is going to test the game on as many different systems as possible. While we're working with a calibrated system, it's always a good idea to hear the game "in the wild." For example, in my studio I've got a calibrated 5.1 monitor system, a consumer 5.1 system (Creative Labs), a standard def television that I can pipe a stereo feed to, and a weak mono speaker so I can test the mix on that. I frequently check on the different systems to hear how things are coming together. I also check for differing bass management configurations and surround downmix scenarios (this refers to how a 5.1 signal is turned into stereo or mono). Finally, I can pipe the PC, PS3 or 360 build of our game to any of these destinations. It's a mess of wires, but it gives me the flexibility to immediately test a number of different listening scenarios.

A note about equipment: Mastering engineers, the people who make a living out of putting the finishing touches on professional music recordings, often spend tens or even hundreds of thousands of dollars on equipment, software, cabling and professional room tuning and calibration (yes people make a living doing this, they're called Acousticians) in order to create listening environments that allow them to output the most balanced recordings possible. The goal is to create optimal results on as many consumer systems as possible. The same, including money spent, can be said of audio engineers who work in the film and television worlds. Yet the average equipment spend for an audio guy in the games business is in the $10-20k range, according to stories I hear round the campfire. We are, in a sense, given nerf guns and asked to compete with high powered assault rifles.

Think I'm exaggerating? Find a pro audio engineer in the film world and get him to price out his rig for you. You'll be shocked to hear what he has to say.

So when assembling the audio and mix of the game, what kind of system should we be mixing for? The surround system? The stereo TV? It's hard to say, but there are a few things I think should be kept in mind:

First, television speakers aren't going to sound like a movie theater. It's simply impossible. I've been pushed by non-audio people to make our games sound bigger and badder, to make the explosions "knock me out of my chair." Yet the person making these demands is listening on the Aquos LCD that seems to be standard issue these days. Even Uncharted 2, in my opinion the best sounding game of 2009, loses a bit of its magic on television speakers.

Second, according to research numbers I saw recently from a major publisher (one of the top three, not the one for which I work), roughly 50% of consumers of AAA action and FPS console titles have their games consoles wired to 5.1 systems. Whether they actually have the speakers set up correctly or just piled in the corner, I can't say for sure, but they're familiar enough to say they've purchased a surround system and are using it.

So my philosophy, based on 5 years experience in games and an additional 5 in music and tv/film, is as follows:

Make it sound amazing on a 5.1 system. If half our customers have one, then give them the best show possible. They spent the money on the system, they care about sound, there's no reason to let them down.

Make it sound as good as possible on stereo, but don't compromise the 5.1 version to make minor gains in stereo. This would be the visual equivalent of significantly damaging the appearance of the game in 720p or 1080p to make 480i look a little better. If something is horribly wrong in stereo and surround has to suffer to fix it, do it. But only do harm to the surround mix when it really matters. Aim for excellence, not perfection.

If anything can be done to salvage mono, give it a shot, but accept that the majority of people aren't using an RF adapter with their 360 or PS3 and that mono, by its very nature, is going to sound poor.

Finally, if you're working on a platform with relatively fixed sound reproduction (handhelds), design your content on a professional system so you can make the surgical adjustments you need with clarity. But the final signoff should always happen on the handheld hardware. Always.

I'm sure I'll hear as much agreement as disagreement regarding my feelings on which platform should be targeted. Audio people love to argue and we all have different opinions on the matter. But don't think we're using big speakers because we just want to hear things on a big, loud system. That's not the case at all.

Related Jobs

Crystal Dynamics
Crystal Dynamics — Redwood City, California, United States

Audio Lead
The Odd Gentlemen
The Odd Gentlemen — Los Angeles, California, United States

Sound Designer
Digital Extremes
Digital Extremes — London, Ontario, Canada

Sound Designer


David Clair
profile image
Great article, and i think you are totally right with your perspective on "Make it sound as good as possible on stereo, but don't compromise the 5.1 version to make minor gains in stereo." etc. etc.

The majority of customers that are concerned with sound will most likely have a 5.1 (or other surround PL or PLII) setup of some kind/quality.

Tim Haywood
profile image
"Yet the average equipment spend for an audio guy in the games business is in the $10-20k range, according to stories I hear round the campfire. We are, in a sense, given nerf guns and asked to compete with high powered assault rifles."

I just wrote a very long rant about underfunding, then thought better of publishing it on a public website.

I do agree with you Mark, underfunding in an issue at the moment, and for those teams in that situation the only way to address it, is to keep having patience and work with the bosses and gain their trust, and then put together paperwork, with practical examples for a proposal about how things can be improved. I recently did this to securing funding for an internal recording studio, and its already paying for itself. I think the same processes of examples and proposals can be demonstrated on this issue, and I would suggest using the film industry as the touch stone for the ideas and motivations as to why we in the game industry need to use professional sound mixers.

Mark Kilborn
profile image
Tim, I've been fortunate at my current job on this. We're pretty well equipped, and my superiors respond very well to requests for items that are reasonable and well explained. But I've experienced the opposite in the past and it was very unpleasant, and I know plenty of audio guys are being held to film quality standards but not given the resources to achieve that. What's nice is that some studios are investing in their sound, and the results are undeniable. Their games are winning awards for sound and pushing the quality bar ever higher.

By pointing to the film industry, as you said, as well as at some of the better equipped game audio facilities, we've got a great argument for why we need expensive equipment. There's also the argument that we're the only part of the game that truly exists in three dimensions, although I sometimes get the evil eye for quoting that :)

Brendan Blewett
profile image
Eff that noise. Audio exists in FOUR DIMENSIONS.

Simon Ludgate
profile image
Hey Mark, I don't know much about sound, but reading your article piqued my interest in trying to better understand the sounds produced in games. I'm one of those gamers who uses 5.1 surround, but also has to plug in headphones to avoid disturbing family members late at night. My 5.1 system has a headphone jack and seems to "remix" the surround signal to stereo when they're plugged in, but would I be better off setting my PS3 to stereo rather than surround for better quality / more accurate sounds?

Mark Kilborn
profile image
Probably not. While I can't say for sure what your system is doing when you plug in headphones, my guess is that it's doing a stereo downmix of the surround content. When surround is downmixed to stereo, the content from the left surround speaker is mixed into the left front speaker, the content from the right surround speaker is mixed into the right front speaker, the content from the center speaker is mixed into both the left and right speakers and the subwoofer content (which we call the LFE channel) is thrown out. There's math behind how much of each is mixed into the left and right but I'll spare you.

This is the standard approach to downmixing, and most hardware behaves in this way. I'm guessing that's what's happening when you plug in headphones.

Now, if you were to set your PS3 to stereo mode, it would do the same thing. In stereo mode, the low level audio engine of most games will be working with the console's audio hardware to downmix the surround content to stereo in the same way I described above. If they're not, I'd say that's a tech failure on the developer's fault.

So it's two paths to the same thing really, except in the latter scenario you'd have to manually change back and forth in the PS3 settings menu. Sounds like your current setup is the better of the two.

Ted Brown
profile image
I learned this lesson on a game I became a "surprise audio designer" on. The audio director for the studio got me some "balanced" headphones that I thought were crap at first, because it didn't sound as good as my consumer headphones. But then I found out that my headphones boosted the bass! So after balancing levels with the "pro" headphones, it still sounded great in my personal ones, and on a 5.1 system cranked extremely loud. The headphones I were given were Sennheiser HD-280's, I believe... any audiophiles want to chime in on good balanced gear for amateur audio guys that suddenly have to "go pro"? =)

Rune Ploug
profile image
Yes as a guy who is going indy and still working my last months for Sennheiser I believe I can be of a bit of assistance. Mainly ask your question Ted at where all we headphone audiophiles live. Denon's high-end closed like the 2000 model and Sennheiser's HD280 or 380 model of pro series are good choices for a balanced systems. There is many other solutions. For sound card get something like uber's of ASUS PCI-express audiophile grade soundcards since these reasonable amplification and the stuff if you are going cheap or a dedicated external pro solution if you have more than a few hundred bucks.

There is many brands of good reference speakers including Klein Hummel but I can't remember which are cheapest for a decent speak:

Dave Endresak
profile image
Good article.

One thing I didn't see mentioned that can be pretty important (and limiting) is the actual physical environment when using external speakers. Specifically, the various elements in the physical environment can impact audio playback, and likewise with changes to the physical elements even if a system performs well prior to such changes. This becomes a major limitation for most people because consumers usually have to work with their living space and and cannot play games in an ideal (for audio, anyway) physical environment that would offer them the audio experience they ultimately desire.

I guess it's sort of analogous to the sacrifices made during mixing except that the physical environment (usually) isn't anywhere near as flexible as the digital signals being recorded.

Mark Kilborn
profile image
Dave, saving that subject for another post. It's a big one :)

Keith Fuller
profile image
Enough with the hifalutin audio claptrap. Just make it sound good on my Korean War-era CRT, Mark. Is that so much to ask?

Christopher Key
profile image
"A visual analogy would be if the maker of a video monitor toyed with the color balance, brightness, contrast, etc before they shipped the thing out of the factory."

And, in fact, TV manufacturers do this. It's very common for the preset settings on a TV to have the red component turned way up. It's called "red torch". The reason is that when you have your TVs in a bank in a store, it catches your eye versus a TV that's calibrated correctly. Unfortunately this is so common now that people will sometimes think a correctly calibrated set is 'wrong', looking muted somehow.

Christopher Braithwaite
profile image
Why can't games have a sound mix optimized for both stereo and surround? Why should either setup have to suffer for the other? Developers deal with different aspect ratios for televisions and different hardware configurations for consoles and PCs all the time, why should audio be any different?

Mark Kilborn
profile image
Good question Chris.

Ideally, they would have a mix optimized for both stereo and surround. In application, however, that can be difficult to achieve for a number of reasons. Let me first say that my philosophy doesn't encourage letting one suffer in any major way for the other. I believe a high quality stereo and surround mix can be achieved with current tech and tools. But there are situations where you're left with a decision that will adversely affect one and favor the other, and while it won't be a big enough issue to ruin either, it's something that might be a bit noticeable.

Achieving everything you want in both stereo and surround, with zero compromise, would require significant technical and creative effort. You would require an audio engine that's aware of whether it's being played back in stereo or surround, and you would have to create special rules for how the audio engine behaves in each scenario. Should things pan and spatialize differently? Also bear in mind that the 360 is capable of outputting stereo AND surround simultaneously. How do you deal with that? Would custom content bloat the memory footprint?

Using the example of an explosion, you can count on the presence of a subwoofer in a 5.1 system. You can't in stereo, so how do you compensate for that and still deliver the same power? Through very creative EQ you can create a punchy explosion on a stereo system. But making those changes to the asset may create an explosion that's overpowered in 5.1 and sounds very inconsistent with the rest of your mix. So you can create two assets if you're unwilling to compromise, but multiply that by every asset in the game that might require some special attention and you can see how the work would pile up. A lot of audio departments feel understaffed as it is, so that's an unrealistic expectation to place on them.

Going back to your example of developers dealing with different aspect ratios for TVs and hardware configurations, they're doing the same thing we are: compromising. They might make the text a bit bigger than they want because they're afraid of the text being unreadable (see Dead Rising). They might find that some texture they've created looks great in 720p or 1080p but looks mushy in 480i, so they make adjustments.

We always have to compromise. It doesn't mean that the stereo mix sounds awful and the surround sounds great though. That, to me, indicates a failure on the part of the audio staff on a title. But it might mean that the guns in the surround version might sound a little better than the stereo version, and an audio guy might consciously allow that to happen because he feels the creative gain in surround is worth the compromise. It's all about judgment.

John Byrd
profile image

Wylie Garvin
profile image
I totally agree with Mark's piece. I'm not an audio guy, but I've seen first-hand that 5.1 can make a huge difference, but you definitely need a proper environment to do the mixing in. My vote is to just mix in 5.1 so that that sounds spectacular. If the 5.1 mix is good, a stereo downmix will probably sound good too--at least, nobody with stereo speakers is going to be complaining about the sound.

I don't have 5.1 at home and didn't realize how much of a difference it makes until right near the end of [a recent AAA title]. We had a little sound room next to our production floor, that was soundproofed and had a 5.1 system in there. The game sounded *amazing* in there... except for a couple of cutscene movies which we had done in stereo instead of 5.1. Now I realize how big of a mistake that was, and if anyone asks my opinion about this on future titles I will tell them, do everything in 5.1.

Another thing is for all videos in the game, make sure you have the dialogue, music, ambience etc. in separate tracks that can be tweaked at runtime. This helps with localization (one dialogue track per language, but they can all share the same music and ambience tracks), it helps with downmixing from 5.1 to stereo (for example, you can treat dialogue and music tracks differently in the downmix), and with last-minute levelling changes. If we get new video content from upstairs a few days before shipping, and one of the videos is now too loud compared to others or compared to the game, I want an easy way to adjust it.

Richard Marzo
profile image
Mark says: "There's also the argument that we're the only part of the game that truly exists in three dimensions, although I sometimes get the evil eye for quoting that :)"

This is very interesting. How does this sound?: electronically generated sound, since it always comes from a single point, is in fact always one-dimensional. More accurately, the speaker is a point or collection of points generating sound in some three dimensional space, first the speaker housing itself, and then the room or other architectural construct (or even outside).

The actual surface(s) of the speaker(monitor) which physically generates the sound is of course three-dimensional, just as the screen of a visual display is, but the surface(s) acts in a way which is topologically equivalent to a point (or set of points, usually only a handful). If you were able to shrink the surface down and still generate the same sound, within the same casing/room/etc., the effect would be the same. A set of speakers would again be a set of 2, 6, etc. "sets of points", generating their effects in a basically random three-dimensional space.

2-d displays on the other hand, behave a little differently. Since 3-d displays are a kind of false 3-d, though a really good one, much of what is said about 2-d applies to "3-d" displays as well. Though resolution will vary, the screen creates a flat picture (layered pictures for 3-d), which significantly loses all of its qualities as it is shrunk down.

A holodeck (sorry) would generate light and sound in true three(four, sigh) dimensions.

You know, after re-reading my own comment, I am not so convinced of its validity :) I feel more confused after writing and trying to edit it.

Let me know what you think.

Btw, excellent article, I learned much. The gulf between film and game audio mirrors that between film and game graphics in some ways. Films often offer higher fidelity, whereas games provide the variety that control gives to players. Game audio is more in the player is directly involved in creating it.

Daniel Armstrong
profile image
Awesome article and thankyou for getting this into the "Producers" eyes and ears!!! I was actually thinking yesterday if there was any data with regards to Gamer's audio setups. If anyone can point me to any other studies that would be super.

I can't even begin to explain how important a calibrated system and a good listening environment is to create "mixes" that translate to wide range of different playback devices. I really feel that if your doing it right in the first place then less work will be needed at the end. I work with a pro 5.1 system with bass management, a consumer kit and also headphones and I can now really trust my own judgements with regards to playback translation on other devices with these options. No more need for NS10's and terrible ear fatigue hehe!!

We really, really need to listen to the film/tv guys who have been doing this for years, we can learn so much from them, considering we are creating media that playback on the same equipment. Just with less expensive gear. Anyone fancy getting me some ATC's, a floating room and a Neve Gemini??

I did really like what the Bad Company guys did with the ability to select your playback devices thus having seperate mixes of the game.

Lets us not get into reference levels and dynamic range etc etc hahaahah!!!!

vinny abraham
profile image

In films, when you down mix to stereo, the LFE information is down mixed with the left and right (front channels). The subwoofer content is never lost. If one has a full range loudspeaker for the fronts, they reproduce the LFE content really well. How is this different with games?

In game mixes, is bass management in the hands of the sound designer or with the end user ?

About HI fidelity sound in games,

I have two sound systems in my home. A mediocre (consumer grade) sounding rig for gaming and a hi-end custom made rig for music. Most games sound very good on the consumer grade system. If I listen to the same sound tracks on my hi-resolution system, it actually sounds un-listenable. I can hear all the compromises done to the mix to make it sound good on consumer grade audio systems. I am now wondering, who should be complaining ..:-))

Kevin Fredericks
profile image
great writeup! I think you are totally right that developers are just beginning to understand how much tech goes in to sound design, mostly due to the long history of film scoring and audio production.

Here's a looper: I'm writing a score for an iPhone game. I have played around with a lot of different sound treatments, but I've discovered that the recordings from my RadioShack dictation microphone is beating out the signal from my AKG when it's brought down to hand-size. It is a lot like flash games that just compress and output a fully orchestrated recording vs. the ones that use .mod style or chiptuned music.