The risk with poorly designed loudness processing systems is that programmes which naturally have a wide dynamic range can sound compressed and thus lose the impact of well-designed and modulated sound
While loudness control has been a topic of great importance for a long while, it is only recently that standards bodies have come together to develop a means of measuring perceived loudness, as opposed to absolute levels as measured by the traditional VU or PPM metering.Jean-Marc d’Anjou from Miranda takes a look at the new standards.
ITU recommendations BS.1770 and BS.1771 define both the ways of measuring loudness and the levels that should be regarded as acceptable, in the short term and in the long term, ensuring that the viewer’s goal – to be able to watch a programme and its interstitial commercials without reaching for the remote – is met.
The ITU recommendations are now being enshrined in practical standards for broadcasters. EBU recommended practice R 128 is based on them, as is the Calm (Commercial Advertisement Loudness Mitigation) Act in the USA, and US broadcasters are rapidly updating their workflows to meet the December 13, 2012 deadline for compliance.
It is important to note that, as the name of the US legislation hints, the measurements and corrections should apply across the entirety of a broadcaster’s output, not just within a piece of content. If the loudness movement is to be seen as a success, programmes, commercials and trailers should all feel to the viewer as being at the same volume.
The tremendous challenge this presents for broadcasters, and one that is completely transparent to both legislators and viewers, is how to manage the large quantity of content segments, from a growing number of content contributors, without impacting the efficiency of the workflow or the quality of the broadcast. New technology now being introduced to the market can solve this problem through the use of segment-aware loudness processing techniques that create an optimal workflow, guarantee compliance and ensure that viewers won’t be forced to reach for their remote controls due to spikes and dips in programming volume.
Traditional loudness control can fall short
There are a number of established techniques used during production to ensure a good audio mix. The Dolby DialNorm (dialogue normalisation) system, part of the Dolby E and Dolby Digital encoding, is widely used. It incorporates metadata, which ensures that the DialNorm value — generally set to the average loudness level of a segment — is maintained through the playout process. Ultimately this metadata is carried through to the viewer’s set-top box, ensuring the average loudness level is preserved, between channels as well as across segments in a channel’s output.
This elegant solution, while preserving audio characteristics for the viewer, presents a significant problem. Broadcasters have found it difficult to preserve the DialNorm metadata throughout their playout chains, which typically include a large number of devices from multiple vendors. This is further complicated because broadcasters do not typically have full production control over all the content they broadcast, and not all of the content will have good DialNorm metadata. Live programmes, including news and sports, are also a challenge, because typically DialNorm is set as a post-production process, usually by analysing the complete programme.
For most broadcasters, using a real-time loudness processor, installed just after master control and ahead of encoding for transmission, is the simplest and most practical way to address the loudness requirement, and overcome the issue caused by wild feeds. With a set-and-forget configuration, any variation in the loudness across different feeds will automatically be corrected using the real-time processor. The potential problem? What if the segment has already been processed for loudness during production?
This unnecessary multi-processing of audio content can result in a reduction of audio dynamics and other audio degradation. Segment-aware loudness monitoring and processing techniques, such as those found in Miranda’s Intelligent Automatic Loudness Correction system, enable broadcasters to overcome this challenge.
Marrying file-based, real-time processing and automation playlists
The CALM Act, and other potential legislation, requires not only loudness correction of broadcasted content, but it also requires logging and reporting by content segment (programmes, commercials, promos, etc). A segment-aware loudness monitoring solution, then, must be directly linked to a station’s automation system to properly identify all the segments in the playlist. This integration enables a loudness monitoring and logging system to easily be reset for each segment. This segment-aware technique is often overlooked but is absolutely required to comply with most legislation.
In an ideal world all content would arrive at the broadcaster with good loudness performance, and this would be confirmed as part of the ingest QA process.
With the majority of content and most commercials arriving as files, the use of file-based processing is an obvious step. This involves determining the average loudness for the file and then moving the audio up or down to meet the target loudness. The big advantage of this being that it maintains the dynamic range of the content. If a piece of content is flagged in the automation system as having been processed, either during production or during ingest using this file-based processing, the loudness processor can be switched off dynamically. This may include switching off the realtime loudness processing for “trusted” content — material which has been produced in-house or comprehensively tested at the QA stage, and which might suffer from double processing.
The risk with poorly designed loudness processing systems is that programmes which naturally have a wide dynamic range can sound compressed and thus lose the impact of well-designed and modulated sound. If too much reliance is placed on fast compressors and limiters then audio pumping can become audible to viewers and are likely to result in more, if different, complaints.
Processing and logging
As I’ve discussed, file-based processing is ideal for content that arrives in advance and for content that is purposely produced with a wide dynamic range for dramatic effect, such as feature films. File-based processing can’t be used, though, for late-arriving files or for live content. In these instances, a real-time processor must be used. So, a well prepared broadcaster will use a system that combines the two different approaches.
Most broadcasters will adopt a “safety net” approach, whereby the loudness monitoring and correction is at the last stage of the playout chain, just ahead of the transmission encoder. Some, though, will use the technology at the point of ingest, so that all signals within the plant are considered controlled, meeting the house loudness standards.
While the safety net should ensure that output loudness standards are maintained, it is very important to log loudness values for all segments. This will ensure that any subsequent complaints can be easily addressed and provide an audit trail of compliance with loudness regulations. Only a segment-aware system can provide the level of detail and certainty required to effectively handle this. Through the Miranda Intelligent ALC system, processors record the average loudness per segment, enabling easy creation of detailed reports on the channel output by piece of content. Most importantly, though, the Intelligent ALC will govern the interaction between loudness processing elements and the content at hand — turning the appropriate elements on to ensure proper audio levels, and turning them off to avoid processing files that have already been processed. This is a key piece of the puzzle — effective loudness processing means not only complying with the legislation, it means ensuring the highest quality viewing, and listening, experience for your viewers.
Loudness has been a difficult concept to define and to maintain, which is perhaps why the industry took so long to come to grips with it. Because of the dangers of over-processing sound it is clear that an end-to-end solution is required.
Content producers and broadcasters have to work together to determine loudness norms and limits, and ensure they are strictly respected with the minimum of processing and audio artefacts. Ultimately, this is the solution to a problem which is of great significance to viewers. If a viewer is forced to pick up the remote control to adjust the volume during a transmission, that viewer might just as easily change channels. Good loudness control keeps audiences happy.