EXS 24 Logic Sampler Users Group

Nick Batzdorf wrote: 
>
>   From: Murray McDowall <murraymc@...>
>
> >In recent tests on a P4 2.4 PC with the EXS24 (and also with the freeware
> Sfz
> >sf2 player) and the free Maestro grand piano (gig format converted to SF2
> and
> >EXS24 using CDExtract) I am getting over 60 sustaining voices (pedal down,
> >every note playing a different audio file, no repitching of samples) from a
> >Seagate Barracuda ATA 5 120 gig drive without clicks or pops.
>
> That's totally malodorous. I've never used the EXS24 on a PC, but I'm 
> running six times that many voices on a dual gig G4. If your 60 
> voices are stereo, I'm still getting three times as many.
>
> People complain about the present version of GigaStudio being long in 
> the tooth and only getting 160 voices. Something is rotten in 
> Australia.


Sir! How_dare_you besmirch the honour of my country. Australia demands
satisfaction! 
Expect whistling sounds overhead imminently.

Seriously -- we had better clarify what a voice is in this context. I am
talking stereo voices and different files for every voice. All these streaming
samplers cache the files already used so if you have an instrument which pitch
shifts the same sample to create three notes then only one file is streamed. 

I very much doubt you are getting a genuine 180 stereo voices -- each involving
a unique continous unlooped file streaming up from a single 7200 ATA drive.

The test I was reporting was performed with long unlooped samples (eg 30 second
long piano notes) and with an instrument which has a one note: one or more
samples mapping. Program some very big chords or play a gliss on the white
notes on your 88 note controller with the pedal down -- that will give you 52
voices and then add some black notes to go further.

You cannot simply count up total numbers of voices for parts that stop and
start, replay the same note, share the same file for more than one note or have
notes which use looped sections.

Following these rules I am having no trouble saturating the EXS24 limit of 64
stereo voices for a single instance with all voices sounding continuously for
several bars or until all the samples finish with no clicks and no information
dropped (BTW you need to open the Virtual Memory panel to see whether any
information was dropped -- the EXS24 will just stop playing some notes if it
can't keep up with the disk access and there will usually be no audible clicks
when this is happening).

Adding further EXS instances I can get more voices but there seems to be a
limit of around 80 voices from this Seagate Barracyuda ATA 5 -- a drive which
is certainly not the fastest available  either in seek time or sustained read
speed. I would expect get better performance again with 64kB clusters rather
than the 16 kb currently. 

With Kontakt, the EXS24 or evenb the freebie SFZ  the buffer settings for
DFD/virtual memory make a big difference to the number of voices you can
successfully stream continously. 

Whether or not the instrument is selected in the Mixer/Arrange page affects the
size of the audio buffer Logic uses for that instrument so there is another
variable. Your soundcard's ASIO/Core Audio  buffer size is yet another
variable.

Regards,
Murray