At a press conference during the NAB Show in Las Vegas, Linear Acoustic President Tim Carroll sounded an alert: In the rush to comply with the CALM Act, broadcasters must be wary of solutions that ensure compliance while reducing audio quality for the consumer.
“My fear is that people will move toward an easy answer (to loudness control),” he said. “There seems to be a lack of discussion on the preservation of quality, and that is a disturbing trend. I notice a spike in claims that there is equipment that can magically fix loudness problems and make metering happy, and that scares me because a happy meter is not always a happy customer.”
Carroll went on to explain that blind adherence to the CALM Act can be dangerous to content, resulting in inferior audio quality that, ironically, meets the letter of the law. This is especially problematic when irreversible changes are made in the digital domain, permanently changing the character of the sound. To address this problem, Linear Acoustic is introducing new products based on its CARBON technology, a hybrid system based on the concept of reversible (bypassable) audio processing and for which the company won a technical Emmy Award last year.
“The goal is to get it right at the beginning,” Carroll said. “And if you can’t, you have to do it in a way that can be reversed, so managing loudness doesn’t mean ruining content.”
The new Linear Acoustics products introduced at NAB to achieve this goal are the LQ-4000 loudness quality multimeter and AERO.calm (Coded Audio Loudness Manager).
The LQ-4000 is a high-density loudness meter that can decode PCM, Dolby Digital, Dolby Pulse, Dolby Plus, Dolby E and MPEG. Its I/O supports ND/SD-SDI as well as MPEG and IP transport streams. It is one of the first products to offer real-time Dolby Dialogue Intelligence, which melds dialog gating with the ITU-R BS.1770 specification.
“Dialogue Intelligence is an important technology because loudness metering is related to what we pay attention to when watching TV,” Carroll said. “What we are generally following is dialog. So, the letter of the law reads that your dialog loudness shall be correct. Dialogue Intelligence produces consistent measurement results, which is very important.”
The AERO.calm audio processor controls both loudness and dynamic range, without affecting the original content. This is accomplished through a hybrid approach of multiband audio analysis and metadata control that allows the user to choose between reversible and permanent audio processing.
All this support’s Carroll’s primary message: Adherence to the law should not be blind. There is already ample evidence that some loudness-measurement systems can yield inconsistent results, and that others achieve compliance at the occasional expense of sound quality.
“If you take a piece of content and measure it, it should be like measuring a 9V battery — always the same result. But with a loudness meter, that is not always the case. If you have Dialogue Intelligence on it, you will have the same result every single time — no matter where you measure it, and no matter whose product you use to measure it. That’s really important.”
Carroll concluded his presentation with some advice on proper loudness control:
- Meter appropriately and often. The goal is to get loudness right as early in the process as possible.
- Correct in the file domain if at all possible. This allows users to look at the entire program and make a decision, rather than being surprised and having to act aggressively on the content.
- Avoid irreversible processing and decode/re-encode schemes because they damage content.
These last two points are related. Loudness control working in the IP transport stream domain first decodes the audio, makes changes and finally re-encodes it.
“This type of change is irreversible,” Carroll said. “Not only is there quality loss from decoding and re-encoding, but there are permanent changes to the audio as well. So yes, it will be smooth and consistent. Yes, it will satisfy a meter. But, it can sound terrible. Making the same changes in the compressed domain means we can make a measurement and we can send data in the stream. That will give it the same sound, but not touch the audio.”
He concluded, “Remember that presenting quality content is something that’s going to keep viewers, and that’s going to keep revenues flowing. This is a silly little saying, but it’s important: When it comes to loudness control think globally, but act reversibly.”