What is "aggressive" compression technology?
How to identify a aggressive compression algo, especially in speech
coding domain, which is my interst. Here, the compression refer to some audio data compression algo,such as CELP,PCM and so on.


Anyway, i am talking the concept of "aggressive " used here.


Any comments would be appreciated!

HyeeWang

I would call an algorithm "aggressive" if it, despite severe objectively measurable distortion (THD, delta information lost etc. ) introduced in the encoding process, still carries most of the relevant information. In case of speech, you may drop MOST of the information contained in a PCM waveform and still be able to understand the content (of course, you need to know what exactly to keep and what is the suitable form of storing such information).
The most aggressive speech compression I can imagine would be transcription of the spoken word into text (and even using minimal coding on that), then using voice synthesizer upon playback. If only this was so easy to achieve at usable failure rate (misrecognized words, unnaturally sounding synthesis and other problems).