|
Game audio has obviously come a long way from the 8-bit days. While the bleeps and bloops that made up NES game soundtracks may be what the non-gaming public associates with video games, today games often have fully orchestrated scores played by live musicians. Sound effects, too, have become far more complex, varied, and interesting. How many different footstep noises does the last game you've played have in it? Probably a lot more than you'd expect.
While game audio has improved immensely over the past couple decades, it's also becoming much more difficult to do consistently, and far more potential problems crop up. I'd like to list a few that I've encountered over the past few years, but before I do, let me say that my experience with recording and mixing audio comes primarily from a musical background and not a game audio background. Nevertheless, I think these potential flaws are ones that can be recognised and corrected in good game audio production.
Voices too low in the mix
This is something that I notice almost immediately, perhaps because of my musical background. In almost every genre of music (excepting some heavy metal or experimental acts) vocals are mixed significantly louder than any other aspect of the song. This is partly because vocals are often the centre of the song, but also because, unlike other instruments, vocals don't usually blend in as well and need to be clear.
I've noticed in a lot of games lately, though, that the voices are mixed far too low. This is especially problematic in games with a heavy emphasis on story. Mass Effect is probably the worst offender in this regard - sometimes the ambient crowd noise is louder than the main characters speaking!
Main speaking parts should always be clearly audible above every other sound, including music, foreground sound effects, and background ambience. That means they need to be loud. Unless there is a specific reason for the voices to sound muffled or distant (being underwater, for instance), they should be at the top of the mix. The player should never have to strain to hear the dialogue.
The sliders in the option menu don't affect the sounds they should
This one usually compounds the previous example. If there is a sound effects slider, turning it down should affect all non-dialogue, non-music sounds in the game. It should not just effect the menu noises.
I've encountered games where I think I've adjusted the audio mix to a more reasonable level, only to discover that only some of the sound effects are quieter and the rest are still drowning out the speech, or are still rumbling my sub-woofer. Which leads to . . .
Your explosions are too loud
I'm sorry, I'm sure you're very proud of how realistic and impressive and big and booming your explosions sound, but they're too loud. My neighbours do not want to hear them. Game audio should be mixed for living rooms, not movie theatres.
I've played plenty of games where the sounds seem to be well mixed and well balanced until something blows up, and then suddenly I need to turn my speakers down. But only until the explosions are done, because the appropriate volume on my speakers for explosions and every other sound in the game is not the same.
* * *
Game audio has come a long way, and is leagues better than it was even ten or twelve years ago. But while the quality of audio available to us has increased significantly, the mixing of that audio is often extremely uneven. A good game audio mix should make everything clear that needs to be clear, and keep all of the sounds at volumes relative to the rest of the mix.
|
Dialogue being too low in the mix is not totally unique to games. Try watching "Last of the Mohicans" (1992) without subtitles and see how much dialogue you can understand. It's probably better if you have a quality sound system, but entire scenes of the movie are just a bunch of low mumbles on your average laptop or SDTV.
The visual aspects of games have become more and more movie-like over the years; there are even movies embedded in almost all games (there's a reason they're called "cinematics"!). Game visual design and pacing are becoming more and more movie-like every day. Audio design moving down the same "let's do it like they do in movies" path is therefore no surprise, particularly given that movie producers area getting into games and bringing with them their "do it like WE do!" mentality.
But games are not movies.
Game developers need to internalize that. Never mind the ceaseless, pointless, and meaningless "games are/are not art" debates... this is about basic stuff. Games are played on 15" laptops, 24" monitors or 37" TVs across the room, not in movie theaters. Games are played through stereo headphones or inexpensive stereo speakers, not AC3 or 7.1 thousand-watt systems.
Audio designers, your game happens in living rooms or bedrooms, not theaters. Arguments about "but there are some people who DO have huge home theater systems!" will fall on explosion-deafened ears, because the vast majority of gamers do not play their games on such systems. (Yet another case of disconnect between developers and their customers, maybe?)
Part of the audio acceptance test really needs to be listening to the game audio through the gamut of environments, and as Adam mentioned, checking for dialog intelligibility and acceptable dynamic range. Listen through cheap headphones. Listen through nice headphones. Listen through the built-in craptastic speakers in a laptop. Listen through the built-in craptastic speakers in an LCD TV. Listen through $50 computer speakers in a small room (the size and furnishings and relative position of the listener to the monitor/TV/speakers MATTERS). Listen through a $300 "home theater in a box". And have the listeners be people who aren't part of the audio team, or who haven't subconsciously memorized the dialog in a given test scene. Ask the receptionist whether it sounds OK and if they can understand the dialog, or if the explosions are too loud (or not impactful enough).
Oh yeah, listen through that $4000 Bang and Olufsen system your CEO was kind enough to equip your in-office gaming room with too, if you have it. But don't pretend that this has any relationship to the real world, as defined by the people who will be buying your game.
As for myself, I spent a large chunk of change on getting what I thought would be the ultimate gaming PC. Best video card, lots of RAM, best monitor that I could fit in my desk space...which leads me to describing my desk space. Smallish. I don't have that large of a space where I keep my PC, it's a very small room, so in my binge of purchasing the most awesome equipment, I didn't reconsider my decision to also invest in the most thunderous subwoofer I could wedge into my generous budget.
Huge mistake.
Maybe if I was rigging the sound system for a coliseum, convention floor, or the Superbowl, I would've gotten my hands on the correct equipment. But this subwoofer would create an audio tsunami of destructive carnage when the PC would boot up. Even with the sound really low, the woofer would drone out every sound in the room (and quarter mile radius) rendering my sound system completely worthless.
I ended up returning it, and settling for a Boston Accoustics system for thirty bucks...and I still need to adjust that woofer to lower levels sometimes.
Point being, audio is definitely a big issue in games, and something people aren't putting that much emphasis into. There really isn't a point in selling a game "because it has neat audio." It's worth it to invest nice graphics into a game, worth investing in solid animation, great gameplay, and awesome character design. Those things sell a game.
Audio doesn't sell a game (music/rhythm games excluded, obviously). But it can certainly kill it (music/rhythm games HEFTILY included). Yet while fault sometimes lies on the developer, the user's home theater system could frequently end up being the culprit.
I absolutely agree with your general observations. It is interesting that all of your 'game audio mistakes' come under the broad category of 'the mix'. This is something that can only be done well by the developer when all the music, dialogue and effects are present at the very end of production and the game can be played through in its entirety. That window of opportunity is increasingly fleeting for a variety of reasons.
It is an area I’ve been evangelizing to spend more time on in production to get the sound just right prior to release - and it is a huge challenge, even for developers with a lot of resources dedicated to audio. But it is arguably one of the most important steps in production, after all, what is the point of spending all that money on producing AAA dialogue, music & FX if they cannot be heard clearly at the crucial moment in the game.
I am firmly of the opinion that a separate mix needs to be produced for high-end (7.1 / 5.1 surround) and also for standard speaker configurations (Stereo TV speakers etc), in this way the high end mix need not be compromised by tweaking for the stereo down mix, and likewise, the stereo mix will be better tailored (often with cleaner dialogue) to the stereo / TV speaker environment. This is something I think we'll see more of in over the next year or so and it is something that will really increase the quality for gamers across the board, no matter what system they listen on.
In the end, it comes down to the developer having the time and resources to sit down and mix the game properly and usually the only way to have a good shot at this is to plan for a post-production audio period from the very earliest stages of development, which is very rare. I think you hit the nail on the head when you said it is becoming more difficult to do game audio [mixing] well; complexity and amount of assets have certainly increased and there is more pressure on audio at the last minute, so it is easy to forget all about a mix pass, as long as the game 'makes' sound its probably the best some developers can expect.
There are however, a great many games that have superb mixes and make the most of dynamic level adjustment to immerse the player further into the game play and even go far beyond simply carving out enough space to hear dialogue, their mixes really seduce the players into the game (Deadspace springs directly to mind). So I do feel there is also a sense of optimism in this area.
Cheers
r
There are plenty of things you can do sonically with your music and SFX and even dialog to make sure that the dialog sticks out above the other assets. It's amazing what you can just do with some simple EQ and a compressor.
I believe Rob may be on to something regarding separate mixes - this already happens for features, getting a film mix, as well as a TV mix. The reason this would be a good idea, is because you lose the ability to add drama, power, and finesse through dynamic range when you do a TV style mix. It is more compressed, less volume between loudest and softest parts generally, it requires eq'd and mixed and with multi-band limiters slammed dialogue so the dialogue always sits well above everything else in the mix. This is all so that in an open plan living room, with kitchen behind the couch, the son can still hear the movie playing on TV through its crappy speaker over Mum noisily preparing dinner in the kitchen. I have over-dramatized slightly, but this already occurs to a certain extent, the film is mixed so the most important thing can be heard, the dialogue. This is a compromise - it recognises that dialogue must be heard on crappy systems, so it mixes everything else very conservatively.
This is not how I want to hear my games, and it is not how I want to mix the audio for my games. I understand that it is necessary, but it is not something that I would want to promote. If people are prepared to spend on graphics cards every year or two, cpu's every two to three, but wont invest in decent speakers, they are denying themselves the holistic game experience. It is a disingenuous position to make a stand, and one I can't understand. As a sound designer it makes me sad that no one wants to hear my sounds as good as they can be. I understand the need for a version that will sound good on crap, but is there no-one that wants to hear their games sound amazing too?
As for the idea topic of having a good sound system, I personally do. I listen to just about everything in 5.1. And often, when playing games in that mode, I find that the dialogue is too low to be heard clearly.
But I think it's also important to recognise that most people playing games do not, and maybe never will, own high end systems. The portion of gamers who buys a new graphics card every year or two is actually pretty low. I buy a new computer roughly every 5 years, personally, and I know people using computers older than mine. Plus people play console games on a wide variety of set-ups. Most people still have SDTVs, despite the impression among the more "hardcore" gamers that HDTVs with 5.1 sound systems are common. Most people still play games on pretty basic technology.
I certainly don't wish to make excuses about bad mixes - a bad mix is a bad mix, it should never happen. If you can't hear the dialogue on a mid level 5.1 system, then something has gone wrong somewhere. It could have been that layers of mix snapshots fought against each other, ducking the dialogue too low, or a few key lines were set with different mastering effect settings, making them quieter thatn all the other dialogue, so they no longer sit well in the mix. It could have been a lot of things, and there are usually safety nets such as QA, sound team playthroughs, etc that would catch these sorts of things. When games get released with issues such as these, someone somewhere has made a rather large error.
I still say I think Rob's idea is a good one, make two mixes, one for nice 5.1 - 7.1 systems at reasonable levels, and a compressed TV style mix for the rest. It would work. It wouldn't be too much of a compromise, and would mean no one would miss important narrative details due to unintelligible dialogue.
My major point, probably badly said, lost in amongst my own noise (ha ha, bad sound joke) is that to be really expressive with sound, we use dynamics, the difference between loud and soft sound. If I were to do this in a game, the soft, subtle, emotive nuances of my work would be lost if it were listened to down low. It would also mean nothing loud would pop out, because the loudest part would be the dialogue. Everything would be mixed so it would be loud at all times, so that when the audio is reduced by people playing at night, or in a noisy atmosphere, they could still hear everything. This is what we already do to a certain extent, but I would not want to do it even more. The possible solution to this is, like Rob said, is have two different mixes. The I could do my cool, kick arse dynamic sound, and a boring hear all things at all times sound mix.
But people, for the most part, upgrade your audio path! Sound in games should be getting better and better. If you show you care it will. If you show you don't, no one will spend money and time making it better. It could be amazing, sometimes it already is, and in the future I sure hope it gets better and better.
I tried to find an email for you but no dice, so apologies about the OT reply. I did just want to take a moment to thank you to your response to the GLAAD news piece, though. I get so tired of the defensive derailing and denial I see in the community sometimes, and your response was such a breath of fresh air and... such a relief, honestly. So thank you.
I think the issues with sound really has to be placed at the foot of the hardware maker. Sure they can convince you that one more speaker will make all the difference, But I only have 2 ears, and the person speaking to me only has one mouth, yet im still able to locate the source spacialy. I only use a basic stereo setup with a sub, and I still can hear the small nuances in the sound. It may just be me, I do have sensitive hearing, and do my best to keep it that way.