Kamikaze
Chad, thanks for the break down of the content. I'm not that happy with the Programs supplied with the clavnent, but the SFZ files and samples seem good, and the SFZ multisample format makes creating a new program easy. Interesting to see how how the SFZ files break down. I've been using light, but I may have a play with the full and change CC1 to another available option

 
Are you looking for Lite with the Mute on CC option?  I can make a couple .sfz variations of these for you.  What memory footprint would you want to stay within? 
 
 
 

ShellstaX
Hi Chad,
 
Thanks for the quality samples Chad.
 
To assist wider understanding wrt some of the posts above regarding size can you/anyone confirm/clarify my rudimental understanding ...
 
The samples are 96K x 24bit x 1channel = 2304kbps BitRate
All the Clavinet samples  are ~1.5GB (on disk)
All the PianetT samples are ~650MB (on disk)
 
When calling an sfz are it's memory requirements determined by what it has definitions for /calls to (as opposed to necessarily loading all instrument samples - but might typically be one full keyboard / mic arrangement, velocity layers and releases)? 

 
This will just depend on the samples that are called as <region>s within the .sfz files.  Any sample that is defined as a <region> will be loaded into RAM by Rapture Pro, regardless if it's input control parameters don't have it trigger until certain values are met.  So even if a group of samples aren't supposed to be triggered until MIDI CC1 has a value of 64 or greater, they will still be loaded. 
 
I think this answers what you are asking? 
 

So the resource requirements for disk and memory are a direct result of the quality of the samples and the component requirement of the sfz definition. (?)

 
Yep, based on the file size of the samples.
 
 

In the case of 'Gothic Plunkton' (2025MB), it uses:
Mellotron Strings Tape - 25MB
Vint Saw Bitred - 8KB
Clavinet Full No Muted - 1.56GB
GoldTops GTN Powerchords - 57.2MB
 
That still doesn't add up to 2GB - more like 1.6GB ???
Regardless - still quite a memory draw and of course that'd likely be moreso if the other samples were 96/24.
I'm still running an 8GB machine :)  (which has coped well for my requirements thus far). 

 
I'll have to look at that program, but there is an option for Loading Mode which per Element, if using 32bit loading mode it essentially doubles the RAM usage.  Maybe some of the Elements are using the higher memory loading mode.  You can switch it and re-save as a new program to test.
 

 
I noticed  also, seemingly moreso for the lower notes, the duration is longer (/size bigger). Is that due to lower notes having a longer sustain?

 
Exactly, you'll see this with all sampled string instruments, plucked and hammered.