The summing bus myth

summing-bus-myth

Some people believe that there is a difference in sound among DAW systems which is caused by different approaches on summing the tracks in the mixer. Well, this is total bullshit. Digital summing sounds exactly the same in every DAW, bit for bit perfect. If a host does it differently, the host is broken. Summing is addition. It is the most simple form of audio processing.

If you do it differently or inaccurately you will be colouring the signal. There are people who say that some DAW systems sound more transparant or warmer than others thanks to “analog” sounding summing. Hans Zimmer for example. In an interview for Steinberg he had said:

Another thing is that Cubase sounds really good. And what I mean by that is that what you put into Cubase is what you get out. It’s not colored like other systems. I can tell straight away when somebody brings in a Logic track — I know exactly what it was recorded on.

Digital summing should be bit for bit accurate, if not, the host is broken. Summing in Logic, Cubase, Live, Reason, Pro Tools they all should sound the same. If not, something is wrong with the one that has a different sound. Steinberg should have corrected that note of Hans in my opinion. Or does Steinberg really believe that kind of marketing bullshit?

Try it yourself!

The easiest way to test summing is by putting channels out-of-phase against tracks which are in-phase. If you don’t hear a sound that means that the channels cancel each other out. This is good because it means: perfect phase, perfect summing. You can try this with 2 channels, 24 channels, or hell even with 320 channels! Any amount of even numbered channels will do. Put one half in phase, and the other half out-of-phase. Simple as pie.

Comments

12 responses to “The summing bus myth”

  1. Moritz Avatar
    Moritz

    I think there is also confusion of how people explain these differences and I include myself too.

    Your test is valid and yes it confirm that before mixing and processing all DAWs sound the same.

    A good test would be to mix the same stems or audio track in Logic, Cubase, Reason, Live etc with the internal effects of each software and than compare the final results.

    There will be differences here and there and that will show which DAW produces the best result overall.

    A Reason compressor will not sound like a Logic or Live compressor that is all down to how the algorithm has been better designed and emulated to replicate the behaviour of a compressor.

    For example I found a difference between how the Reason Bus Compressor sound compared to the Glue Compressor of Live- I found the Glue more accurate and responsive.

    1. Marco Raaphorst Avatar

      These are totally different things. Summing is about the process of combining channels, mixing channels. It is just a matter of clean calculation. But as soon as you start comparing EQs and compressors you will notice differences. So summing is 100% clean, the same.

      I like the Glue too and the Reason Bus Compressor. The Glue is more flexible but the Reason comp is just perfect too leave on on all tracks imo. These a great standard devices.

  2. Moritz Avatar
    Moritz

    Agreed

    Summing is 100% identical across all DAWs

    When you start using equalizers, delays, reverbs, compressors etc. each DAW will show its strengths and weaknesses.

    1. Marco Raaphorst Avatar

      Yes, then it gets way more subjective.

  3. Jason Shaw Avatar

    You are wrong. As you can see on this website, DAW’s vary wildly in frequency response. Sonar always sounded dull to me, which is why I stopped using it. Sure enough this website shows that Sonar has a lower response in the high end than most any other DAW. I know that you’ll say “but you can’t even hear those frequencies. True but NOT capturing them affects the interactions down the chain that we can hear.
    http://src.infinitewave.ca/

    1. Marco Raaphorst Avatar

      I only believe in a listening test. If I repitch a track a cent you won’t be able to hear it. And then ALL frequencies have changed. I’ve done many A/B tests in my life to stay sane 😎

    2. Ralph Avatar
      Ralph

      This is incorrect, the link refers to Sample rate Conversion, not the summing of signals which are totally different things. DAW’s do vary wildly in there sample rate conversion techniques.

      1. Marco Raaphorst Avatar

        There are small differences. In most cases extremely small, below human hearing.

  4. EmrahAlpat Avatar

    I’ve seen this discussion numerous time. Everyone claims to hear how sonar is dull or Cubase is transparent or whatever, until someone wants to do a blint A/B/C/D test. Then everyone loses their s**t.
    I am yet to see one single instance where people claiming to hear differences between cables, DAWs etc can identify which is which in a statistically significant way.

    1. Marco Raaphorst Avatar

      Same here. Listening tests, when done correctly, are the best. I have noticed many opinionated people won’t react as soon as I offer a listening test.

  5. Benjamin Golinvaux Avatar
    Benjamin Golinvaux

    I would say that this can be theoretically slightly untrue : there are algorithms that are slightly better when performing several floating point additions (with varying exponent). See https://www.wikiwand.com/en/Kahan_summation_algorithm

    1. Marco Raaphorst Avatar

      I don’t understand that piece of information, what’s the point? And I only believe in a listening test. Of course with floating point you can have an extremely small difference because of rounding errors. This can never ever be an issue in practice.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.