Synth CPU Usage Comparison

Discussion in 'Samplers, Synthesizers' started by Ankit, Sep 22, 2021.

  1. Ankit

    Ankit Rock Star

    Joined:
    Oct 17, 2015
    Messages:
    835
    Likes Received:
    470
    Location:
    Horsehead Nebula
    [​IMG]


    Note - This is just and comparison of cpu usage. It is not about which synth is better.

    Midi input: Basic Chord
    1 Oscillator
    1 Filter (Default)
    Distortion (Default)
    Delay (Default)
    Reverb (Default)

    CPU usage Source: Performance Meter in Reaper (Track)

    Plugin Version: Latest Available
    OS: Windows 10 LTSC
    DAW: Reaper x64
    Processor: Core i3 9100f

    What I did:

    Played a chord in a loop

    Used 1 Oscillator
    Set unison voices to max
    1 Filter
    1 Distortion
    1 Delay
    1 Reverb

    [​IMG]

    Obviously it is not a fair test. Just a quick one to give an idea. CPU usage varies with oscillator modes, filter types, effect algorithms, etc.
    I am on Windows.
    CPU - Core i3 9100f
    DAW: Bitwig Studio.
    Windows Task Manager: CPU usage is from the plugin host process of Bitwig Studio, which is a separate process & we get exactly what a plugin is costing CPU. Unlike Ableton, Bitwig hosts plugins in a separate process.
     
    Last edited: Sep 23, 2021
    • Like Like x 7
    • Interesting Interesting x 5
    • Love it! Love it! x 2
    • Useful Useful x 1
    • List
  2.  
  3. Barry T

    Barry T Ultrasonic

    Joined:
    Jan 20, 2019
    Messages:
    87
    Likes Received:
    37
    Very interesting - what DAW did you use?

    Could be cool to try Avenger too, it is rumored to be rather heavy on the CPU.
     
  4. RobertoCavally

    RobertoCavally Producer

    Joined:
    Jan 20, 2021
    Messages:
    168
    Likes Received:
    111
    Always an interesting relative comparison. Which DAW?

    Anyway, I have a similar result with Serum (you probably mean TasK Manager), 2-3%. But when I duplicate the track 10 times it goes to 5% "only". And that is in Ableton that, as far as I can usually read here, is not super multicore optimized (but cores do kick in..)

    Can you make a test with 10 instances of e.g. Serum, all armed/played at the same time?
     
  5. Ankit

    Ankit Rock Star

    Joined:
    Oct 17, 2015
    Messages:
    835
    Likes Received:
    470
    Location:
    Horsehead Nebula
    I used Bitwig Studio. CPU usage is from the plugin host process of Bitwig Studio, which is a separate process from DAW & we get exactly what a plugin is costing CPU.

    I don't have it. :/

    CPU - Core i3 9100f
     
    Last edited: Sep 22, 2021
  6. Ankit

    Ankit Rock Star

    Joined:
    Oct 17, 2015
    Messages:
    835
    Likes Received:
    470
    Location:
    Horsehead Nebula
    That's interesting. Sure, I will update.
     
    • Interesting Interesting x 1
    • List
  7. ArticStorm

    ArticStorm Moderator

    Joined:
    Jun 7, 2011
    Messages:
    5,276
    Likes Received:
    2,073
    Location:
    AudioSexPro
    What about Massive 1.5.8?

    Also which versions were used for the comparision?

    Sylenth1 might be on top, but for me it sounds ass. I have maybe also used it over the years to much. 13 years are maybe far to long and often.
     
    • Agree Agree x 1
    • Interesting Interesting x 1
    • List
  8. RobertoCavally

    RobertoCavally Producer

    Joined:
    Jan 20, 2021
    Messages:
    168
    Likes Received:
    111
    Wow.. that's the worst offender on my system - 7%. But again, 10 instances and the percentage barely rises to ~9%. Seams it spreads well across cores..

    Meanwhile, Diva has entered the chat ;)
     
    • Like Like x 1
    • Useful Useful x 1
    • List
  9. ArticStorm

    ArticStorm Moderator

    Joined:
    Jun 7, 2011
    Messages:
    5,276
    Likes Received:
    2,073
    Location:
    AudioSexPro
    I am not talking about Massive X. I am talking about Massive. I run it always in Eco mode, its perfect, phat sound for a fraction of CPU usage. And just sounds better than Sylenth1.

    DIVA will just rip the ass off, like usually, but u-he did such a great job on emulating things. And emulations just cost CPU usage.
     
  10. Ankit

    Ankit Rock Star

    Joined:
    Oct 17, 2015
    Messages:
    835
    Likes Received:
    470
    Location:
    Horsehead Nebula
    Just installed it. On 7 unison voices, it was averaging around 5% in eco mode.
    [​IMG]

    I completely forgot about Massive. I will update it.
     
    Last edited: Sep 23, 2021
  11. Ankit

    Ankit Rock Star

    Joined:
    Oct 17, 2015
    Messages:
    835
    Likes Received:
    470
    Location:
    Horsehead Nebula
    Updated the comparison. This time using Reaper's Performance meter. 10x performance column also added.
     
    • Like Like x 1
    • Love it! Love it! x 1
    • List
  12. junh1024

    junh1024 Rock Star

    Joined:
    Jul 28, 2011
    Messages:
    1,345
    Likes Received:
    400
    • Like Like x 1
    • Funny Funny x 1
    • List
  13. odod

    odod Platinum Record

    Joined:
    Jun 5, 2011
    Messages:
    589
    Likes Received:
    249
    try Spire with Bellatrix audio patches .. always heat up my CPU :(
     
  14. ArticStorm

    ArticStorm Moderator

    Joined:
    Jun 7, 2011
    Messages:
    5,276
    Likes Received:
    2,073
    Location:
    AudioSexPro
    Wow i didnt know Massive is still so cpu heavy compared to sylenth1.

    Now lets try with MX? Shall we?!

    thanks for the tests.
     
  15. Ankit

    Ankit Rock Star

    Joined:
    Oct 17, 2015
    Messages:
    835
    Likes Received:
    470
    Location:
    Horsehead Nebula
    I want to do this comparison on different DAWs & I want to include LFO modulation in the test to make this test a little more reliable. It will take some time.
     
  16. ArticStorm

    ArticStorm Moderator

    Joined:
    Jun 7, 2011
    Messages:
    5,276
    Likes Received:
    2,073
    Location:
    AudioSexPro
    Sure not a big deal. Thought you also look for suggestions on Synths. In case you might have missed some good ones.
     
  17. clipper

    clipper Kapellmeister

    Joined:
    Aug 1, 2019
    Messages:
    84
    Likes Received:
    54
    Location:
    Málaga.
    First of all, thank you very much for sharing your results with us all. Really interesting table.

    Could it be possible to add some more? Such as Diva, Largo, LuSH-101, Omnisphere, Pigments, Repro-5, Spire, Super 8, TAL Bassline-101, TAL J-8.
     
  18. Xupito

    Xupito Audiosexual

    Joined:
    Jan 21, 2012
    Messages:
    5,295
    Likes Received:
    2,875
    Location:
    Europe
    Berkeley BSD Unix version pleaseeeee
     
  19. Ankit

    Ankit Rock Star

    Joined:
    Oct 17, 2015
    Messages:
    835
    Likes Received:
    470
    Location:
    Horsehead Nebula
    I think it will not be fair to compare analog emulation's cpu usage. They are all very different.

    I will add Massive & Massive X in next update which I am hoping to post soon. I will also add omnisphere, cypher, strobe & Icarus v1.
     
Loading...
Loading...