Hi guys!

I recently wondered what's the de facto industry standard in the game industry for dealing with audio headroom.

I'm pretty sure that every showcase-application, like uploading it on soundcloud/website, demands - often quite heavy - limiting/compressing, but what about the actual in-game use?

I recently viewed the audio files of Deus Ex: Human Revolution, and they are quite low in volume, not even peaking at -0.1 db, but significantly lower. In other games, it seems like the maximum loudness gets somehow in the code overwritten, so that quite low audio files are played at a louder volume, but that might just be my imagination, as I don't have access to all soundtracks. I also don't know how combinations of audio and sfx are handled, there isn't a API intern limiter, is it? But if this isn't the case, than one has to really be careful with the audio levels, I guess. I mean, the developer could always reduce signal input, but since this is a developer forum, I thought I might ask?

I just wondered: Has anyone a clue at what level to peak? EBU recommendation?

Regards

Sebastian

Hi Sebastian,

There is no specific standard, though Sony and Microsoft have been working such a standard for console games (http://gameaudiopodcast.com/ASWG-R001.pdf).

One challenge is that the audio environments for games differ tremendously: PC, iPhone, Xbox.... all have very different requirements. For example, in a console game, you may want large dynamic range, because the gamer is probably playing in a decent room, and had minimal distractions. However, for an iPhone game, it's quite likely they're playing in a noisy environment (train, subway, standing in line at Starbucks, etc...), so you wouldn't necessarily want a big dynamic range.

Note that the level of the game's output is different from the levels of individual audio files themselves. Game audio engines have the capability of changing the volume of each sound based on what's going on in the game.-- the game "mixes itself" based on what's going on in the game--at one moment, there might be barely anything happening, and at another there may be literally hundreds of concurrent wav files being played and mixed together..

GIve the pdf above a read as a starting point. But in general, this is a 'hot' topic in game audio.

Thank you, Brian! That pretty much covers my question.

My to go to route is now organizing every audio file as uncompressed wav, peaking at -0.1 db, and reducing the dynamic range if needed, but rendering it for web with a light/mid compression. It would be even better to consider the dynamic in the mixing stage, but I guess that only makes sense if you're producing for one (higher budget) game.

Also thanks for the insights on how game audio engines handle the volume, that bit of information was really helpful!

Regards

Sebastian

Like bschmidt1962 said above me, there isn't really a specific standard. One of the most important thing when dealing with overall loudness in game music is mastering all of the music at a consistent level across all of the tracks, as well as all of the sound effects. When the programmers go into the engine and set specific audio tracks on levels & sound effects on animations & actions, they will actually set the playback level (Usually a variable in the game settings or something along those lines). It creates less work for them and a faster production overall if each of the audio files are set consistently on your part.