Vienna Symphonic Library Forum
Forum Statistics

185,278 users have contributed to 42,390 threads and 255,479 posts.

In the past 24 hours, we have 2 new thread(s), 21 new post(s) and 52 new user(s).

  • VEPRO 5 CPU Dropouts

    I'm using the latest version of VEPRO 5 slaved to the latest version of Logic 10 on the same machine.  I'm having an issue where the CPU monitor in VEPRO will randomly drop to 0.0% causing a stutter in playback.  I've tried adjusting the multithreading settings, buffer settings, even reinstalled my entire OS including all new versions of Kontakt, Play, etc.  The problem happens when playing back an entire sequence or when soloing only one track.  It also happens when doing an offline bounce which is most puzzling.  There doesn't seem to be any logic to it.

    Does anyone have any ideas as to why this might be happening?

    Thanks.


  • can you give CPU and disk configuration ?


    MacBook Pro M3 MAX 128 GB 8TB - 2 x 48" - 1 x 27" screen --- Logic Pro --- Mir Pro 3D Dolby Atmos --- Most of the VI libs, a few Synch... libs --- Quite a few Kontakt libs --- CS80 fanatic
  • last edited
    last edited

    @Cyril said:

    can you give CPU and disk configuration ?

    Mac Pro 4,1 2.26 GHz 8-core

    28 GB RAM

    Mac OS 10.10.5 - Logic 10.2

    SSD for samples (for the most part.  Some samples still on a 7200RPM platter drive)


  • Are your Play and Kontakt instrument in VE pro or in Logic ?

    Are you using MIR ?

    Having a SSD inside your MAC PRO does not allow to have the full speed of the SSD

    Having some samples on HD is no good too

    Did you monitor CPU usage ? Are you using the 28 GB ? if not you can increase the pre-load buffers to use it

    Can you monitor the CPU and I/O disk with "Activity Monitor" when you have your drop outs ?


    MacBook Pro M3 MAX 128 GB 8TB - 2 x 48" - 1 x 27" screen --- Logic Pro --- Mir Pro 3D Dolby Atmos --- Most of the VI libs, a few Synch... libs --- Quite a few Kontakt libs --- CS80 fanatic
  • last edited
    last edited

    @Cyril said:

    Are your Play and Kontakt instrument in VE pro or in Logic ?

    Are you using MIR ?

    Having a SSD inside your MAC PRO does not allow to have the full speed of the SSD

    Having some samples on HD is no good too

    Did you monitor CPU usage ? Are you using the 28 GB ? if not you can increase the pre-load buffers to use it

    Can you monitor the CPU and I/O disk with "Activity Monitor" when you have your drop outs ?

    Hi, Cyril.  Thanks for responding.

    I don't use MIR.  I have 1 instance of Kontakt in Logic, but the rest are in different instances in VEPRO.  

    My entire machine's RAM footprint when using my full template is around 24GB when everything is loaded and working.  I'm afraid if I increase the buffers I'll run out of room pretty quickly.

    When I playback a sequence and monitor all 16 cores in Activity Monitor (it accounts for virtual cores as well), I don't see the dropouts or spikes that seem to be occuring in VEPRO.  That points to it being a VEPRO issue and not the system, although I don't know how accurate Activity Monitor is when monitoring CPU usage.

    The strangest part is the stutters in playback occur even when bouncing offline, which normally would eliminate any cracks or stutters.  This also points to something wrong with VEPRO.

    Regarding the HD issue, I was under the impression the Mac had four independent SATA channels that weren't slaved by the slowest link in the chain.  Is that not the case?  Would there still be a performance bottleneck even if I solo a track that is playing back samples located on an SSD?  So, if I remove the platter drives from my machine I would get an immediate performance increase?

    I'm grateful for your help.


  • After much hair-pulling, I seemed to have fixed the problem.

    I have no idea why, but changing the Logic X audio setting of "Process Buffer Range" to "Small" seemed to fix the problem (for the most part).  There are still occasional dropouts, but I think those can be attributed to everything being run on one machine, rather than some systematic problem.  

    There is very little information on the "Process Buffer Range", so I can't really offer any reason as to why this worked.

    Paul