[Bob Katz]

The more you know, the more you learn what you don’t know yet! I received this letter from a very knowledgeable designer of video games, and I felt obliged to discuss the depth of issues with him.

 

On 1/17/23 7:23 PM, Alex Previty wrote:

Hello Bob (at least, I hope this message makes its way to Bob)!

I work in audio for video games, and my current work is pretty involved in the overall “mastering” process of the pipeline. Historically, we have been monitoring around 79dB and have been utilizing traditional peak and loudness metering to make sure our content is sitting at a good level. However, these values seem to be hard to pin down and seem to be redefined on a per-project basis. To make things easier and keep them consistent, I have been looking into utilizing the K-System for monitoring!

We are “mastering” our audio to sound best in a traditional home environment, so working at K-20 and monitoring at 83dB will yield results that are too dynamic for most players. K-14 seems to be the best target for our content, but monitoring at 83dB will likely feel too loud and also be fatiguing over long listening sessions. Would you recommend reducing the monitoring level by an equal relative amount, resulting in listening at a level of 77dB to compensate for this?

In short, is it generally a good idea to make sure that “0” on the K-System meter measures at the same decibel level in the mix room (or even “feels” as loud), no matter if K-20, K-14 or K-12 is used? I understand that perceived loudness plays into this as well. I could understand not reducing the monitoring level if the sole reason for reducing the dynamic range is to get over a loud noise floor, however, lower dynamic range can also result in the need to listen to the content at a quieter monitoring level (i.e. when listening in an apartment at night). Because of that, keeping “0” the same in terms of in-room levels makes sense in my mind.

Thanks for reading. I look forward to your reply!

Alex

 

——–

I replied:

Hi, Alex. I don’t know if I ever responded to your message yet. I’ve been very busy.

 

 

Going by the numbers, the recommendation for video games by the AES would be -18 LUFS if it’s speech centric. -16 LUFS if it’s music centric.  But video games contain lots of sound effects and need the headroom, so I would recommend -18 LUFS or lower. And as I understand -18 is Sony’s recommendation.

 

 

As you know, the hardest part is that video games have a lot of random selection so how do you calculate the integrated LUFS when you cannot?

And that’s where having a calibrated, repeatable monitor gain in the same monitoring room becomes important. But even then, when you wake up in the morning and hit an isolated crash or boom, it seems too loud, but in the context of the surrounding audio, it will be fine. As the ear has to get used to the context.

That makes deciding on levels for video games 10 times harder than any other medium, including TV and motion pictures. I don’t envy you.

 

 

What I would do, and this is just a guess, is to decide on a monitor gain using a very well-recorded video game as a guide. And using a very well recorded blu-ray rilm as a guide. I suggest Star Wars Rogue one. Determine its monitor gain. I measured Rogue One from top to bottom and it was -27 LUFS or so. But there is a ton of dialogue and a lot of dyanmic range there. But I think that when you set your monitor gain to its perceptual center of gravity, everything falls into place, from dialogue to music to SFX. So you begin to become attuned to what they each should sound like.

 

Whatever you do, don’t totally rely on the integrated LUFS, because it does not reflect the perception, only the average. A very dynamic piece measuring lower in integrated level can often be played at the same monitor gain as a more compressed piece measuring higher in integrated level.

 

 

As for the K-System, I would start with 0 LU being -18 LUFS on the TC Clarity meter. And find a monitor gain that correlates with that using dynamic material as references, which you have normalized to -18 LUFS but keeping into account all that I said above, and not being stuck on holding to a -18 when it doesn’t make sense in the context of the above. If you have a boss who says it has to measure such and such, send them my letter.

 

Hope this helps,

 

 

 

Bob

 

P.S. Folks, this is only scratching the surface of the issues. The distance of the monitors from you, the engineer, and the headroom of the system, also affect your judgment of what monitor level you need to set.