[LAU] Petri-Foo preview release: 0.0.2

James Morris jwm.art.net at gmail.com
Sat Jul 2 16:16:28 UTC 2011


On 2 July 2011 16:45, James Morris <jwm.art.net at gmail.com> wrote:
> On 2 July 2011 16:20, Renato <rennabh at gmail.com> wrote:
>> On Sat, 2 Jul 2011 11:21:07 +0100
>> James Morris <jwm.art.net at gmail.com> wrote:
>>
>>> > There's the hackish way, where both samples will always play, but
>>> > one can be quiet while the other is loud. You just need to apply the
>>> > velocity sensing to the amplitude in the correct way. But note that
>>> > there's no threshold value, it's a smooth rampage of amplitudes.
>>> > Only at the extremities will one sample be heard and not the other;
>>> > mid points will cause both to be heard to varying degrees.
>>>
>>> great example of why i need users to report things which don't work.
>>>
>>
>> you mean that the method you were describing of applying velocity to the
>> amplitude doesn't work? I ask because I wouldn't know how to test it, I
>> haven't exactly understood how I should do it
>
> Hi,
>
> No not the method of applying the settings in a specific manner, but
> the part of the program that deals with velocity sensing. I expected
> to be able to use negative velocity sensing values to invert the
> meaning of low and high velocities. Trouble is, the slider is not set
> to allow negative values, and I need to check that the code could
> handle negative values also.
>
> The method would have been to have one patch with normal velocity
> sensing, and the other patch with inverted velocity sensing.

Except the method probably still wouldn't be satisfactory because of
the problems around inverted velocity.... causing.... inverted
velocity!

Which lead me to realize a better method would be to add a velocity
range for the patch to trigger on. There already exists a key range, a
velocity range would be a natural pairing of this.

note though, key tracking maps only keys within the key range, so the
lowest key in the key range provides 0.0, and the highest key in the
key range provides 1.0 (rather than the lowest and highest keys on the
keyboard (ie 127 midi 'keys'). velocity sensing then would have the
same behaviour, ie, it doesn't map to the min (0) and max (127)
velocities, rather it maps to min and max values in the velocity
range.

the problem though, is that petri-foo still doesn't allow layering of
samples within a patch. so say you setup five patches with five
samples to be in five overlapping velocity ranges. that's five
different patches.  you don't want five patches, you want to achieve
the illusion of one patch - ie you probably want to achieve the
illusion of a single musical instrument as realistically as possible.
so if you adjust one patch, you probably also need to adjust the other
four also which could rapidly get out of hand.

i really don't see or desire for petri-foo to be a sampler to be used
for realism. you'd be much better served by linux sampler if realism
is what you're looking to create.

cheers,
james.




>
>
> Thanks,
> James.
>
>
>>
>> renato
>>
>


More information about the Linux-audio-user mailing list