Ihnoc
Active Member
Hello friends,
I have been experimenting with reducing the impact the Berin series instruments has on my RAM load. Note this has nothing to do with Sample Memory as these changes should improve RAM performance when purged. It does remove your ability to use samples from microphone positions you remove. I have written up a short explanation of process below:
https://alexjevincent.co.uk/optimising-berlin-series-instruments-memory-usage/
1. Removing Kontakt Zones by Group
In summary:
In my testing, I have managed to see the below results:
Horn 1 Multi Patch
I think it is important to highlight here that what I am changing here to reduce Object Memory is not the result of 'amature coders' or a lack of optimisation, it is simply how the samples are mapped. I cannot speak for the efficiency of CAPSULE specifically because these changes have nothing to do with it.
Finally while I have said it many times you should save a backup of your patches and save alternative versions of patches you are editing to do this.
2. Reducing TimeMachine memory usage
I have found that the TM Pro Voice memory and Object memory of patches increases if one of the TimeMachine articulations, like 16th or Triplet repetitions, is added to a multi. This is only pertinent to instruments that have TM Pro voice memory (Berlin Brass and its expansions, possibly Berlin Strings) In summary:
The voice limits increase because the repetitions adapt to a tempo, along with the number of voices per microphone, so the increase is tailored to a good experience. It is a shame it does not return to 1 when the repetitions are unloaded, so this might be a bug.
Many thanks,
Alex
I have been experimenting with reducing the impact the Berin series instruments has on my RAM load. Note this has nothing to do with Sample Memory as these changes should improve RAM performance when purged. It does remove your ability to use samples from microphone positions you remove. I have written up a short explanation of process below:
https://alexjevincent.co.uk/optimising-berlin-series-instruments-memory-usage/
1. Removing Kontakt Zones by Group
In summary:
- Save a ‘Light’ version of the patch. Note the Object Memory Used field in the ‘Expert -> Engine’ tab
- Use the Wrench to edit the instrument then use the ‘Expert -> Groups’ tab to search for microphone positions and then select them
- Open the Mapping Editor for the patch. Ensure ‘Selected Zones Only’ is active, then right click to ‘Select All Zones’, then right click ‘Delete Zones’
- Note the Object Memory Used field in the ‘Expert -> Engine’ tab and save the instrument again
In my testing, I have managed to see the below results:
Horn 1 Multi Patch
- ~305Mb Object Memory Used before removing zones
- ~168Mb Object Memory Used after removing zones for Close 2, AB, ORTF and Surround groups
- ~155Mb Object Memory Used before removing zones
- ~84Mb Object Memory Used after removing zones for Close 2, AB, ORTF and Surround groups
I think it is important to highlight here that what I am changing here to reduce Object Memory is not the result of 'amature coders' or a lack of optimisation, it is simply how the samples are mapped. I cannot speak for the efficiency of CAPSULE specifically because these changes have nothing to do with it.
Finally while I have said it many times you should save a backup of your patches and save alternative versions of patches you are editing to do this.
2. Reducing TimeMachine memory usage
I have found that the TM Pro Voice memory and Object memory of patches increases if one of the TimeMachine articulations, like 16th or Triplet repetitions, is added to a multi. This is only pertinent to instruments that have TM Pro voice memory (Berlin Brass and its expansions, possibly Berlin Strings) In summary:
- By default the Standard and HQ voice limits of TimeMachine Pro are set to 1 for Berlin Brass
- Adding one of the repetition articulations to a multi patch increases this to 32 and 8, which adds TM Pro voice memory and increases object memory
- Removing these articulations does not reduce the voice limits back to 1
- You can set the voice limits you like, by opening the Wrench, then Instrument Options and then the Voice Handling tab
- Doing this can reduce the impact of a multi by up to ~200MB
The voice limits increase because the repetitions adapt to a tempo, along with the number of voices per microphone, so the increase is tailored to a good experience. It is a shame it does not return to 1 when the repetitions are unloaded, so this might be a bug.
Many thanks,
Alex
Last edited: