Re: [Traverso-devel] Traverso audio backend & custom audio app

[Niklas Klügel]

Hey!

> I'm not familiar with sequencer engines, and neither with layer- 
> based sample
> editing, would be great if you could elaborate a little on it :)

well, maybe it's  more useful if i explain how I work on my own music  
projects:
after recording samples by recording real  or synthesized  
instruments, I start rearranging
the basic song i created by editing most samples tonally as well as  
cutting the material up
to create different measures or applying those samples to a new  
tempo. it's like remixing
the song. i usually rely on heavy cut and paste editing down to 1/256  
where you can actually
change the waveforms of i.e. drum or synth samples. basically, that's
nothing fancy, i currently use ableton live for that. but since i  
want to move my whlo studio to
a linux based DAW, i am desperately in need of a tool where i can  
edit sound in such an exact way
and looking at the current recording applications for linux, none of  
them is _geared_ towards
such "sophisticated" editing. to give you an idea how this sounds:
home.in.tum.de/~kluegel/theships.mp3 (just drums, bass and a synth).

even on live the process of editing like this is awkward because of  
the sloppy way you can
move around in the arrangement view, but thats something different.
the main misconception with current audio applications (for my way of  
creating music)
is that they are designed to do everything in realtime - dont get me  
wrong, when i am jamming with friends,
doing a live gig or when i am
recording a guitar-lick that is supposed to go through an f/x chain,  
this is awesome -
but when you are doing more sophisticated editing you are forced to rely
on f/x envelopes and realtime-scopes only. there is no way to get  
good/relyable visual feedback.
and more importantly: you get out of processing power very easily.

i guess i am a bit extreme but otherwise some friends of mine work in  
a similar way on
their material, so i might not be the only one in need of such a way  
of audio editing.

regarding the layers:
my idea is to have a track consisting of several layers in a  
hierarchical dependency.
each layer is can either contain a sample or be can be the "sum" of  
all lower layers.
so, say you have a drum and a violin sample stacked, the sum of both  
samples would
be rendered into the upper layer -   in this case what you will hear  
as single/complete track.
layers can be stacked in an additive way but nothing would speak  
against subtracting two layers in the
time or - if you want to be fancy - frequency domain.
f/x should be added to each layer independently and be rendered/ 
consolidated to the internal sound buffers
buttom-up through the hierarchy of layers once. envelopes and  
segments created by cut/copy and paste
should be preserved. all modifications to the layers should be done  
on the rendered audio-buffer
on the same layer the modifications are done. this way, adding a  
sample-layer below an edited layer
will be modified in the same way the top hierachy is edited. e.g.:  
adding another sample
below a  sliced and rearranged drum-beat will result in both samples  
being sliced and rearranged
in the same maner the drum-beat was at the hierarchy of the origenal  
drum-beat-layer.

the concept isnt entirely new but i guess it is interesting from both  
merely academic sound-design
as well as practical music\production-views.

regarding your hint: i think i am able to add this functionality to  
traverso on my own
to a certain degree but i originally thought it would break too much  
with the way how you developers
intended traverso to be, even while preserving the contextural audio- 
editing idea.

thanks for your reply!

so long...
Niklas