While using certain sampler modules, I noticed the module’s output didn’t quite drop to zero when the sample wasn’t playing. Initial output (before playing anything) was always zero, but the problem showed up after playing at least part of a sample. I could see it in the meters in the MindMeld Mixmaster; sending the output to the Nysthi Multivoltimetro allowed me to measure the results.
Out of the modules I’ve tested so far, I’ve noticed this with the Nysthi Quadsimpler and Complex Simpler and the Bidoo Ouaive, and to a lesser extent with the Nysthi Simpliciter. Haven’t yet tested samplers from other developers. The output is often very low, but I’ve seen it at least as high as 1.9 volts. I find it weird that a module’s audio output would be anything other than zero when it’s paused or stopped and no sound is coming out. Looks like this would qualify as DC offset, which I try to avoid. (Just in case the OS may be relevant: I’m running VCV on Linux – Ubuntu Studio, specifically).
Has anybody else noticed this? (Haven’t managed to find a post a bout this.) Is this common with samplers? Should I simply send all sampler output to a DC blocking module such as the Nysthi 4DCB and not worry any further?
Sure, most samplers will stays at their same (DC) value if they pause them. Yes, if you don’t want that put a filter after them.
4dcb is your friend for taking out DC offset but I haven’t tried it on constant values. Also run samplers through a vca modulated by an envelope.
Since the solution has been offered above, I thought I’d also mention that, in the physical world you can use a capacitor for this kind of DC blocking, and it’s what manufacturers call ‘AC coupled’, meaning that DC or, in other words, steady state voltage offsets, are blocked. Capacitors do this by their nature and are very handy in coupling together circuits where you only want the change of voltage to get communicated, and you want to block any of the non-ideal elements of amplifiers or things that might settle down to a DC offset rather than a perfect 0v when no signal is present.
The bigger the capacitor, the lower you can get your roll-off, so when you look at crossover circuits down in the bass range, you’ll see big capacitors (it takes physical size to get a large farad value) being used to pass slow voltage changes, but still block the even slower ones.
Conversely, also in the physical world, inductors are used to block high frequencies, and they will pass DC (steady state or slowly changing voltages) with no problem. So capacitors and inductors are like yin and yang for passing frequency content. And in common parlance you’ll hear about dc-blocking, ac-blocking, and dc-coupled, and ac-coupled, and these are all just extensions of trying to solve the same problem that you’re seeing with your paused samplers.
In the software world, 4dcb is going to be some kind of filter that gradually comes on to block steady state values, it’s like a little machine that looks at the incoming number and says to itself that if the number hasn’t changed in a while, I’m going to gradually add an offset and subtract it until it reaches zero. And if the signal changes faster than the code that wants to subtract it to zero, then the signal will pass through.
nysthi ::complexsimpler, there is a contextual menu switch to bring to 0 the value if playheads are stopped
(but I think I never released in 1.1.6 … so it’s only in v2)
Thanks all! I may adopt @chaircrusher’s VCA approach.
@synthi: good to know, thanks!
you could put a “flow” by AS in between the sampler and the mixer and cut the signal off completely when its not playing.
If you are looking for a basic sampler without DC offset I recommend RadioMusic, it can also go through a folder of samples by CV which is very useful for browsing and playing different drums. I use this if I don’t need anything special like Simpliciter or visual like MusicalBox from NYSTHI.