Airwindows Purest Gain/Console and the Ardour Faders

hello everyone,
as i understand in this video

DAWs downsample audio for changing the gain in Fader Operations. That effects the audio slightly, this greatly the sound of airwindows’ console encoder/decoder pair.

I’m just wondering, if that’s the case with Ardour as well. i doesn’t matter all too much, as Ardour supports post Fader Plugins, but still interesting to know.


Ardour does not use block-size processing for faders, but applies a gain factor to each individual sample, furthermore the gain-stage is dezippered (using a time-constant 1/(25Hz) to prevent audible amplitude modulation). Processing also happens in-place (zero copy).

However parts of what Chris said is however not entirely correct.

Stacking up plugins is not an issue (unless the plugin itself causes problems).
This is because float error propagation per operation is linear in value, not exponential.

The additive error is roughly equal to the round-off error from a single operation, multiplied by the total number of operations

from Smith, Steven W. “The scientist and engineer’s guide to digital signal processing.” Chapter 4: Number Precision

We have been there before when discussing biased-rounding that the airwindows “dark” and “undither” plugins use here on this forum. Perhaps it is relevant in other DAWs. – But don’t take my word for it. Do a null-test, ideally plot the spectrum on the residual. I’m surprised that Chris has not done this in the video.

00:14:00 … 00:15:30 in the video is also sloppy. Most DAWs do not fade using the block-size. @AdmiralBumbleBee has measured it: DAW v. Daw - Part 1: Automation and Fades and DAW v DAW - Part 3: Fader Automation Speed

Logic Pro which Chris shows in the video does this correctly, the fade is block-size independent and spot on.

That being said a “slow fade” plugin that he advertises at the end of the video is useful in some contexts.