7/20/2023 0 Comments Ardour 5.12 64 bit built![]() Reaper - Automation snapping, but moving to precise values is fiddly.Bitwig - Automation snapping and moving to specific values is precise.Ableton Live - Automation snapping and moving to specific values is precise.Pro Tools - Automation snapping and moving to specific values is precise.In order of how easy it was to setup these tests: ![]() That means 800 samples to the left of the desired cutoff and 800 to the right. ( one exception to this)Įvery image is showing a 1600 sample window. There are two images side-to-side: Automation from 0dB to -Inf on the left, and automation from -Inf to 0dB on the right.Įach block on the x-axis is 50 samples, or about 1 milliseconds (1.041666666…ms exactly). I wrote a short program to render the wave files at the desired points: 28,800 samples and 57,600 samples. I’ve ordered my results from best to worst with commentary. This concept will become even more important when investigating other automation functionality. Simple updating each sample’s value every sample yields “better” results, but requires that the automation signal is filtered to avoid any aliasing in the modulation signal. What you’re likely seeing below in the not-awesome results is residuals from that era: automation being chunked, then smoothed. It’s 2019, sample accurate volume automation isn’t an issue. Fixing this adds another layer of cpu intensive operations or a carefully tuned filter along with sample-accurate value updates.)Ĭreating an interpolated automation signal that will provide clean automation at a sample-accurate rate is not easy, and for many years it was somewhat CPU intensive. This can result in a ‘straight’ automation signal that is rendered as a wobbly sine. The downside is that aliasing is possible, and so are signals that are exactly fs/2. This is the (same concept that all of your digital-to-analog converters use). That’s why filtering yields a smooth signal from a “stepped” one. (Remember that a ‘stairstep’ signal can be represented by a sum of harmonics, if you filter the harmonics then you are left with the nice curvy fundamental sine wave. I know of a few products that use a simple low-pass filter, which also can yield a noisy and unstable automation signal. Most products use some form of interpolation, from in-house algorithms, to polynomial to linear. There’s two ways to smooth the value: “manual” interpolation, and low-pass filtering. For many years the standard was to update automation every once in a while, then smooth that value. Not all DAWs have done sample-accurate updates for automation data.
0 Comments
Leave a Reply. |