What comes after AAC?
Reply #21 – 2014-03-26 12:16:30
If I had to put my money on any successor for 2-channel sound, then there would be two candidates: - Ogg Vorbis. Because Spotify uses it. - HE-AAC. Because of DAB+. In a practical sense, sure. But I was thinking more on the conceptual level, etc.Do they really? And if/when, does it have to be a lossy? In a few years, your movie collection will fit in your hand. Yes, well sure. The idea is to look for better codecs for research's sake. If we look at it from a high level I think we can divide the reasoning behind codec improvements in three groups, quality vs bitrate, processing power requirements and features.Quality vs. bitrates is a straightforward one. Somewhere in the 90's probably the first practical lossy technologies were able to produce transparency. After that point the developments were focused on reaching transparency at an ever lower level. From 320 kbps then to around 96 kbps now. The goal here could be to further lower this level.Processing power requirements are focussed on making smaller or cheaper devices possible and extending battery life on portable devices. The goals here could be to further reduce power requirements and are theoretically endless until they use an infinitely small (but never zero) amount of power.Features is the most flexible one. I am ignoring metadata features as they not strictly part of the codec. Newer codec features we have seen include increasing support for more than 2 channels. The question is where that will end. Is more than 48 channels (as AAC supports) necessary? You could argue that extremely high (for music) frequencies are a feature. However, AAC supports up to 96 kHz per channel. Is higher necessary? Low delay is a feature but generally quite good thanks to the special Low Delay version "AAC-LD". Even lower delay could be a goal but you reach a phase of diminishing returns where further improvements are increasingly less worth it. For any of these variables, the improvements need to be big enough for many stakeholders (consumers, professionals, industry etc.) to replace the incumbent codec for something else. Probably at big cost. To be honest, I can only think of the power requirements variable to be important enough to warrant a change. Then again, I might be overlooking something.