- Contains three parts: beginning, middle, end.
Beginning: contains the natural attack of the sound
Middle: this is the looping part, you can loop this as many times as possible
End: contains the natural release of the sound
This format would use some special algorithms that creates seamless transitions between these parts
(I am smart, eh?
- uses some kind of compression technology similar to REX2 that can minimize the size of the loops without affecting the sound quality
- contains a whole ocatve of loops at different pitch. Each part (beginning,middle,end) can be independently changed to a different pitch. Each repeating middle part can also have a different pitch. There would have to be some special algorithm that creates nice transitions between parts that have a different pitch.
I call this "SUPER LOOP" format.
Here is a little drawing that lets you imagine what this Super Loop format would look like inside of a sequencer:

B - beginning
M - middle
E - end
+1 - part at one semitone higher
-3 - part at 3 semitones lower
Figure 1 shows the original Super Loop imported into a project. There is no looping here, just a natural attack, middle, and natural end.
Figure 2 shows the same Super Loop being looped 5 times, changing pitch the second time and the fourth time it loops.
To go from Figure 1 to Figure 2, all you would do is just drag the vertical line between the middle part and the end part.
Also, there should be an option to hide (cancel sound output) for the beginning part and the end part.
That's about it for now. Please feel free to add or subtract.
EDIT: Consider the waveform as only the top-half of a waveform hehe, I don't know why I drew it this way.
