By Tarcisio Longobardi
In the past, the most common way to measure the “loudness” of an audio signal was to represent amplitude variations over a certain period of time through a VU meter, a peak meter or a waveform. These tools, however, don’t give us an accurate estimate of how humans will perceive loudness. As a result, it is possible for two audio sources with identical peak values to be perceived as having very different overall loudness.
For example, movie and television show audio usually has a relatively wide dynamic range — characters can go from a whisper to a shout — as opposed to commercials where the audio material is compressed. Consequently, commercials often sound much louder than content — despite both being within specified decibel peak levels. Audiences experience loudness jumps between programs and commercials, and those inconsistencies are the cause of much frustration and complaint.
To find a solution to this problem, a new kind of measurement that attempts to quantify our perception of loudness has been introduced.
The United Nations specialized agency for information and communication technologies (ITU) introduced LKFS (described on ITU-R BS.1770), which means “Loudness K-weighted relative to Full Scale.” It’s a scale for audio measurement where a K-weighted filter (a filter used to emphasize frequencies that humans are more sensitive to) is applied to audio material to obtain weighted measurements that try to estimate how loud a human listener will perceive a given piece of audio.
The European Broadcast Union (EBU) uses the term LUFS, which stands for “Loudness Units Full Scale.” Despite the different names, LFKS and LUFS are identical. Both terms describe the same phenomenon and, just like LKFS, one unit of LUFS is equal to one dB.
Since 2012 many European countries have adopted EBU R128, which is a set of rules that set maximum levels of audio signals during broadcast based on K-weighted loudness normalization. In 2012 the US Congress approved the Commercial Advertisement Loudness Mitigation (CALM) Act. The Act sets rules similar to EBU R128, requiring commercials to have the same average volume as the programs they accompany.
RTW Continuous Loudness Control (CLC) assures full EBU R128 compliance in a streamlined way, allowing the user to adjust program material to a target loudness value, as well as to an adjustable TruePeak value, with or without correcting the original loudness range.
CLC Works either as a standalone application or as a plug-in inside a digital audio workstation. I have used the standalone AAX version in Pro Tools 11, and the VST3 version in Reaper 5.18 to test the software.
CLC has a very simple user interface. The display area is divided into sections: The “Metering” section is a fully compliant EBU meter. The left side shows the measured loudness numeric values of the input audio; the right side the values of the relative processed signal. In the “Processing” section, the values and their dynamical processing are displayed on graphs. The bar graphs in the middle display the values of the current increase or decrease of loudness and the current percentage reduction of the loudness range.
The circle in the middle shows whether or not the brick wall limiter is engaged and how much of the signal is limited. Finally, at the bottom there are a bypass button, a button for resetting the device and another for accessing the setup menus. The interface has a plain, straightforward look, which makes it easily readable. However, there’s a lot of unused space, which makes it unnecessarily big — taking up valuable screen real estate.
While testing CLC, I found it has a very simple yet functional workflow. Once the processing mode (dynamic, semi-dynamic, static) and target loudness are chosen, maximum LRA e true peak limits are set, and the processor works automatically — granting us loudness normalization of the audio signal. I didn’t have to tweak it any more unless I wanted to change the target values.
During my tests, I was impressed with CLC’s ability to handle loudness variations in a very inaudible way — there was no pumping or brick-walling.
CLC says it accomplishes its dynamic correction by using data acquired by performing realtime analysis of the audio content to make predictions of the future signal progress. I have to admit I was skeptical in the beginning, but the software was surprisingly able to perform dynamic corrections in a very transparent way, even in the first seconds of content by using a technique that combines a look-ahead algorithm with statistical data. Dynamics compression is used only when very abrupt changes in dynamics occur, and adjustments triggered by abrupt volume increases are hardly audible.
CLC was able to recognize natural loudness increases caused by loud passages as a natural part of signal dynamic and thus leave it mostly unaltered. Low-dynamics passages pass through unaltered as well.
Thus CLC works well in its “default mode,” but it is also has presets that suit different kinds of programs (e.g. news and discussion, movies and sports). The presets change different parameters of the user interface and other invisible characteristics of the processor. I found that choosing the right preset is an important part of the workflow because this heavily affects the way the software handles the same situation.
Although these presets are effective, it could be preferable to have more control over the dynamic processing. Once the preset is chosen it is impossible to make any changes. It would be useful to have a plug-in with the possibility to fine-tune parameters and to know exactly what the software does to handle a specific situation.
While CLC is designed for a realtime workflow, it also features an interesting offline operation mode, called “file mode,” in the standalone application. If you use the file mode, the loudness of audio signals coming from a file can be processed. The loaded audio file will be analyzed and processed, according to your settings, and stored as a new audio file. The complete analysis before processing of the audio file will allow a more precise result with regards to the target value.
Oh, and the CLC was just as easy to use in 5.1 as it was in stereo.
CLC has a simple workflow and is extremely user friendly. Its advanced algorithm sounds good, and will help engineers make their programs CALM-compliant in a simple and efficient process. However, its “set and forget” workflow doesn’t allow the user fine control.
Supported platforms are: Windows 7, 8, 10: VST2.4, VST3, RTAS, AAX and standalone; Mac OS X: VST3, RTAS, AAX Native 64, AU and standalone. System requirements are: dual-core 2.5Ghz processor, 4GB RAM, 200MB free hard disk space, iLok USB smart key and iLok account, Internet connection required for activation process. Sample rates are: 44.1kHz, 48kHz, 88.2kHz, 96kHz.
Tarcisio Longobardi is a sound engineer at Silver Sound Studios in New York City.