[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Date Index][Thread Index][Author Index]

RE: Early bird loopers with Live?

> Can you two explain this to me? I almost sent a reply to Jeffs post
> myself, using the same exact phrase that Andy just used, that EDP
> makes loops, not retriggered samples,
Yes the architectures are quite different, especially in the way loop
memory is organized.  It's hard to explain with few words but the EDP
is more "loop oriented" and Mobius is more "sample oriented".  Memory
in the EDP is like a long piece of tape you're constantly recording and
feeding back onto.  Mobius memory is more like little pieces of tape
that are spliced together and overlayed with complicated cross fading
to make it sound like a continuous multilayered loop.
There are advantages and disadvantages with both approaches.  The EDP
memory architecture makes the beloved Loop Windowing easy.  You just
move the playback head and loop point farther back on the "tape".
This is also the reason the "rounding" we've been talking about works
so naturally.  It's really quite brilliant. 
Mobius memory architecture makes things like Shuffle, Divide, and
other "instant" functions easy.  I doesn't have to make changes
gradually as the loop plays, it can stop and completely reorganize the
tape splices then continue with that as the basis for a new loop.
Loop Windowing is difficult because of the way Mobius layers work,
they're more like static loop samples, you can undo/redo to move among
them freely but creating a window that spans several layers and
sliding the window back and forth is complicated.
This architecture of course influences how I think about looping
functions.  I tend to think of a cycle as a sample with a definite
start and end rather than a span of time on a continuum.  This is the
heart of the problem with the unquantized multiply seen in figure 4.5
on page 5-29.
The "extra cycle" problem only exists if you have quantization turned
off or set to SubCycle and you time things just right.  With
Quantization set to Cycle or Loop, my incorrect interpretation just
happens to match what the EDP would do so you don't notice it.  I tend
to be a "bar" oriented looper so multiply is almost always quantized
to a cycle boundary and I don't notice the extra cycle problem.
This will be fixed someday.  Unfortunately I've become involved
in an online game that takes a lot of time :-)