Marginal Differences in Mixdowns of Same Project

By chance I have picked up that if I export/mixdown from a project more than once, making no changes to the project, there are slight differences in the resulting wavs when you open up Wavelab and run the Global Analysis on them. For example, the Integrated LUFS might be the same but the peaks are slightly different as are the momentary LUFS max and mins. Differences in the RMS loudness measurements too. Also the DC offsets.

This project does have quite a lot of VST processing plugins on many channels (Steinberg and Waves Audio) and several VST instruments (Addictive Drums, Pianoteq and HALion 6) - also about 7 tracks of audio, guitars and bass guitar.

These differences are nothing I can ‘hear’ so not a problem, but I am curious and would like to understand. I am leaning towards it being random/timing fluctuations as the mix is rendered with all the plugins (particulary dynamics plugins maybe?) but I am only guessing.

If anyone has some knowledge on this I’d love to hear it - thanks.

Oh - Cubase Pro 12 latest update on W10 latest update and i7 latest gen with 64gb RAM

One thing that may play a role:
Stuff that uses a form of random(reverbs, choruses come to mind) may be slightly different on every mixdown pass, unless the coders/VST developers reseed the random function with the same value.

1 Like

If these instruments have multiple round robins, the same MIDI note at a given place in the project will play a different RR every time it is played. So the generated audio from the instrument will be different every time you render. Note that Addictive Drums use velocity randomization in addition to RR, so each run may have a different quiet/loud hits balance, hence the differences when analyzing the track.

1 Like

I’m guessing if you rendered all your vst instruments before exporting you would solve this.


Thanks Louis (and Zip) - both good points and you have reminded me that I do use the random function on both note velocity and note position on all my MIDI tracks (doh!) so that would result in slightly different peak/peak positions as well - as the difference are inaudible (well, to me) and the overall integrtated loudness unaffected I am fine, in fact a little bit of ‘random’ is to me a desirable facet (usually). That said I am not keen on any ‘mystery behaviour’ in my DAW and now feel reassured, many thanks.

And yes, rendering the intruments would add that as a constant - good point Bunny - but now I think I can see what’s going on I am fine with it as it is :grinning:

1 Like