For some time i'm trying to make a good ideea about how MP3 exactly works... I know already the ideas that make MP3 work.
Some time ago an ideea came in my mind: "if I can't understand other people's work why not to make my own audio copressor"... You know something? It worked!!! Today I have MY OWN SOUND COMPRESSION ALGORYTHM. I toked me about 2 years to reach this stage and thousands of experimentations, testing and listening hours...
I've called it Virtual Sound... I already have built up the third layer. First layer (VSF) is the basis for layer 2 and 3 (this has nothing to do with the three layers from MPEG) and now i'm having fun with layer 4.
layer 4 is in construction so I'll talk about layer 3 (VS3).
MPEG is encoding algorythm is based on a spectrum analisys of the audio signal and the quantification of the data for each sound component. The result after decompresion is a composition of the synthetised components.
Virtual Sound doesn't do it like MPEG. It aproximates the WAVE.
MP3 does it well at 128 Kbs.
VirtualSound does it well at 172Kbs, yet i've tried it also at 138Kbs and 115Kbs...
The analysis of the wave is extremely complicated when using MPEG, so real-time streaming requires CPU power. Today this is possible.
VirtualSound requires 10 to 20 times less CPU power at compression. Decompression is even easier. So it may have some future in the low cost audio systems.
More info ... later... The teacher is bugging me now... Bye....