This article is a guest post, written by Bernard Rodrigue of middleware vendors Audiokinetic, discusses loudness measurement over a range of game audio specific scenarios, such as mobile platforms versus home consoles, and family friendly versus AAA war titles.
The EBU R-128 loudness recommendations and ITU-R BS.1770-3 specifications include three different kinds of measurements:
- Loudness Level (at different time scopes)
- Loudness Range (LRA)
- True Peak Level
The loudness level itself is the foundation of the EBU/ITU standards as it provides a universal way of measuring the perceptual loudness of a digital signal. The true peak level gives an approximation of the continuous signal level in the analogue world. The loudness range gives a statistical analysis of the loudness values over a period of time and provides a measure of the dynamic range.
For this article, we will go deeper in exploring what can be done with the loudness range, and see how it can be used more specifically for games.
Mixing with the Loudness Range
The loudness range is measured by capturing the loudness level over a certain period of time, normally corresponding to a specific segment of a game (level, cut scene, menu). The algorithm eliminates all silence periods (-70 LUFS), removes the very quiet periods (<10 percentile) and removes the extremely loud periods (>95 percentile). The computation of loudness range is then based on the statistical distribution of measured loudness. Therefore, a short but very loud event would not affect the loudness range of a longer part.
Every EBU/ITU compliant loudness meter should be capable of measuring the loudness range. A low loudness range (5 LU to 10 LU) indicates that there is not much difference between quiet and loud moments of a segment. A high loudness range (15 LU to 20 LU) means there is a great difference between quiet and loud moments.
The EBU recommendations do not specify a maximum or minimum loudness range. They do, however suggest that a loudness range over 20 LU is probably excessive. Now depending on the targeted audience, the targeted medium (mobile, console) and the type of game (family, shooter, puzzle, etc…), different loudness ranges might be preferable.
Here are three different game scenarios:
- Puzzle game for a mobile device
- Family/Party game using motion capture (Ex: Kinect)
- War game on console
Those three games are very different in the way they need to be mixed. More specifically, they target different devices and there are meant to be played at different locations.
We will focus mainly on two different aspects:
- The ambient noise present while playing
- The type of speakers and how loud they are
There is a chance that games for mobile devices will be played on public transportation. Cars, buses and trains are noisy environments, even with headphones. If the very quiet moments are not loud enough, they have a risk of not being heard, because they would be masked by the ambient noise. Also, games on mobile devices can be played using the device’s internal speaker, which has a weak frequency response and can distort easily. In general, mobile games should be mixed with a lower loudness range, so the listener is able to hear all of the content, even during the quieter moments.
Family and party games also face ambient noise issues. They are often played while there are many people in the room talking and moving. There is a good chance the game will be played on television speakers, which can be of average quality. This type of game can be mixed while targeting a medium loudness range.
A first person war game has a better chance of being played in ideal listening conditions, either on a high quality sound system or through headphones in a quiet room. By their nature, a high dynamic range is required in these games because deafening gun shots and bombs can occur within a moment of quiet footsteps or conversations. When mixing, a high loudness range is preferable for this genre of game.
Reading and correctly interpreting the loudness range measurements becomes valuable once the targeted output device and the type of game are identified.
Adjusting the Loudness Range
When measuring the loudness range for a game segment, someone could be faced with values that are too low or too high.
When facing an excessive loudness range, one of the first things that can be done is to monitor and record the momentary and short-term loudness levels for the analysed segment. By identifying moments where the levels are too high or too low, volume adjustments can be made to specific sounds to lower the dynamic range.
Due to the unpredictable and variable nature of video games, it might be hard to identify which sounds are in the low and high regions. Another way to reduce the loudness range is to use a compressor effect on the signal.
The EBU R-128 recommendation suggests the following settings for the compressor:
- A low threshold (< -40 dBFS)
- A moderate compression ratio (1:1.2 – 1:1.5)
- A long release-time (> 1 s)
A side chaining approach can also be used to increase the ambiance volume. If important sounds are set up to duck the ambient sounds, the mix clarity is maintained by detaching the front elements in the mix, while decreasing the dynamic range.
On the other hand, when facing a dynamic range that is too low, more volume variation needs to be introduced in to the mix. Here are a few examples of things that can be done:
- Verifying the limiter settings on the master bus, as it might compress the mix too much.
- Lowering the volume of the background music or ambiance, so the perceived volume for quiet moments is lower than busy moments.
- Increasing the volume of very important moments in game.
- When using an HDR (High Dynamic Range) technology, diminishing the HDR attenuation ratio or increasing the HDR threshold at which attenuation occurs.
Customizable Dynamic Range Settings
It is possible to target a specific loudness range based on certain factors. However, the selected target might not always be the best choice for the game audience. More and more games offer a setting to make the dynamic range customizable by the end user. This is an interesting idea.
To support this feature, the dynamic range needs to be lowered dynamically to ensure the quietest moments can still be heard. As seen in the previous section, there are multiple approaches to achieve that. Someone could, for example, use a compressor to change the dynamic range at run time.
Monitoring the Room Level on Mobile Devices
Most mobile devices now come equipped with a built-in microphone. An interesting idea would be to use the microphone to monitor the current noise level in the room. Some could use that level to dynamically apply compression on the game signal, in order to adapt the dynamic range of the game to the noise level of the room.
A similar approach is commonly found on mobile devices, where the operating system measure the amount of light coming into the built-in camera to adapt the intensity of the mobile device’s display in real-time.
To implement such feature in the audio domain, the system might need:
- Filtering on the microphone input to eliminate the noise of the fingers tapping the screen on the touch-sensitive device.
- Averaging the input values over a period of time to attenuate the noise fluctuations.
- Applying output phase cancellation to the input signal when using the device built-in speakers.
The loudness range is a good measurement tool that can help taking decisions while mixing by giving a better reading of the mix.
However, carefulness is required when working in perfect hearing conditions, like the mixing room. All those subtle sounds present in the monitor speakers might not be heard in real life. Also keep in mind the loudness range is only a global indicator of the volume variations. It rejects rare and loud events (loudest 5 percentile), so there could be occasional but ridiculously loud sounds (like gun shots) that would not affect the loudness range.
More about Audiokinetic
At the time of the writing, Audiokinetic is currently finishing the implementation of a loudness meter compliant with EBU R-128 and ITU-R BS.1770-3. It will complement a HDR suite helping game developers to bring their game one step further in the dynamic mixing world.