Home |
Search |
Today's Posts |
#1
Posted to comp.compression,rec.audio.tech,rec.audio.misc,rec.audio.opinion,rec.audio.pro
|
|||
|
|||
The future of audio compression?
What do you guys think? Will there be a successor to the current state-
of-the-art AAC codec, or do people not give a damn about compression no more since FLACs and high-bitrate MP3s nowadays download in less than a minute even in ****ty 3rd world countries? On the other hand, 5.1 audio consumes a lot of space, but that's why we got AAC with all its cool extensions that retains the same quality at the same bitrates that MP3 requires for stereo files. And bandwidth keeps getting better and better at a slow but steady pace, so will there be any demand for transparency at bitrates lower than 128 kbps? AAC was built 'cuz mobile audio players had space limited to like 32 megs back then, so I'm guessing there would need to be a better justification for effort on a new format than "well because nowadays the best comp can store up to 400,000 songs, and I wanna store 400,001." There are justifications for new VIDEO formats for obvious reasons. But if there was to be a new format, what would it be? MP5 would probably be the most suitable name, though it's possible that homofags from ISO might throw in some bull**** like "SAAC" (Super advanced audio coding) as they did with AAC when it should've been called MP4 and ONLY MP4. Anyway, would it just be an extension like MP4 was from MP3 with only tweaks to the same general technique? I'm thinking no, 'cuz AAC is really struggling to make its advantage over MP3 DEAD OBVIOUS instead of "subjective" with add-ons like SBR and Parametric Stereo. It looks to me the frourier and MDCT techniques have been exhausted by now and won't get better, so MP5 will likely need a completely redefined method that will beat the **** out of every existing format. The thing that comes to my mind is some kind of smart, procedural coding that looks at high-level elements of the audio (being able to seperate different instrument and voice staffs, summarize their characteristics and replace the audio with side information about how they flow.) Say you got a recorded 3 hour speech by Obama, the codec stores information about the traits of his voice in the header, and converts the actual words into text. Upon playback, the decoder converts the text into audio with voice characteristics as specified in the header. At the lowest bitrates, his speech would have his voice, but sound really robotic and lack intonation, e.g. if he laughs, he will laugh the same way every time (negating the length) because it's stored in the file as "hahahaha" and not enough info in the header to convey the original tone since we're using a low bitrate. Way more permissible loss than the audio sounding like a rhino shat in the mic like MP3 sounds at 32 kbps. And with music, same thing, any loss will not be bit-by-bit low-level loss like current encoders but difference on the exact tone of the notes, the intensity of the cymbals and drums, etc. much like a sequenced MIDI version of a song. I wonder what the new default bitrate would be? I predict 16 kbps to be advertized as comparable quality to 128 kbps MP3 (MP3 will never die), and 32-64 being the new transparency point in audio. Finally HQ streaming be possible on dial-up, or am I just tripping balls? What about you mofos? What do y'all think? |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
The Future of Audio | Pro Audio | |||
Future of digital audio | Pro Audio |