The Bitstream Blog :: Posts

Ever Larger Acronyms

Hello, happy reader…For this entry, I’d like to veer ‘round to audio again for a time. Specifically, MPEG, audio and lossless coding. The Fraunhofer page for HD-AAC mentions that, “Lossless audio coding schemes offer audio data compression, with the decoded audio data being bit by bit identical to the original audio data.” You should already be familiar with some common forms of lossless coding; MLP or Meridian Lossless Packing for DVD-A, or maybe Monkey’s Audio, FLAC or LPAC. :: LPAC is interesting because an improved version of the LPAC algorithm was chosen as the reference model for MPEG-4 Audio Lossless Coding (MPEG-4 ALS) back in 2004. One of the aspect of this MPEG extension that I find interesting, considering what I think is the potential for downloaded entertainment to eclipse many current distribution models, is that very good quality audio (176/24) can be multiplexed into an MP4 file. It also postpones the day when my clients would have to upgrade their storage systems when they run out of space. :: OK,  MPEG-4 ALS is certainly interesting. It shares many structural aspects with MLP. Since MPEG-4 was designed not only to present a composition of individual audio, video and “parametric ” or synthetic data objects, but also to be scalable. The Hyperdictionary defines scalability as “How well a solution to some problem will work when the size of the problem increases.” To me, it means that, as the QoS (Quality of Service) or available data rate declines, the decoder can still provide a reasonable rendition of the intended composition. This so called “bit-rate adaptation” is quite useful. :: The MPEG folks wanted to take lossless coding one step further and make it scalable, whereby the same data can created in the contribution or origination stage, and pass all the way through a typical production process to the distribution stage, without re-encoding to and from a mezzanine format. With MPEG-4 SLS, another extension to the MPEG-4 standard, you can provide lossy audio with an AAC “core” to “near-lossless” to literally lossless, all within the same, familiar codec. It also has a guaranteed 2:1 compression ratio in lossless mode that makes things much easier from a systems design standpoint. As with other members of the MPEG family, the SLS codec places most of the computational burden on the encoder, though the scalability feature makes the decoder non-trivial. :: One file, many playback environments each with differing requirements of data rate and quality. What got me started on this rant is what I see has a rather bass ackwards development in the consumer space, that of the algorithmic “improvement” of lossy-encoded audio. This is exemplified by Bel Canto‘s use of Sonic Focus‘ secret sauce to fix up MP3s on playback. One could argue that the consumer already has a bunch of often badly beaten MP3 an any help one could give would be an improvement. Fair enough. However, maybe more CE manufacturers should be looking at MPEG-4 SLS, especially those offering “jukebox” products meant to be the home repository for all things entertainment…TTFN  :: PS – If anyone reading this knows how to define a persistent paragraph break in Wordpress, like a </br> in HTML, I would really love to know about it!


Comments are closed.