Had my first play with Nautilus today and I must say - it’s very good!! Covers a lot of ground and sounds good while doing it. I look forward to exploring it further
One small issue I had was the curve used on the Mix knob - I couldn’t see any settings for this anywhere and wondered if it might be possible?
Basically at the moment it seems like the knob acts as a linear cross fader between the dry and wet signals. I found that this results in an overall perception of volume drop at 50%.
I don’t know enough about DSP to know the technicals but I feel like I don’t normally notice this problem. I was wondering if it’s as simple as the curve used on the mix? i.e. right now it sounds like ‘Intermediate’ on this chart, where I believe my preference would be ‘constant power’, maybe even the fade curve.
Maybe it was even something to do with the specific character of the sounds I was working with but thought I’d drop in and mention.
Your observation is correct!
This was a specific design choice.
We noticed that once there was a lot of processing involved, the constant power fade became very extreme at 50% because the wet signal was already so much louder than the dry.
With that being said, this would be a great feature to get control over on the flash drive using the Narwhal web app.
That way the user can configure it to be either linear or constant power depending on their desired application.
I think the mix curve option sounds vital, but I have related issues where fully wet is considerably quieter than full-dry and it becomes impossible to get good results when trying to fade in/out the wet signal. i wonder if there’s a way to have an automatic gain control for the wet signal that depends on the number of delay lines in use or level of chroma or something .
as i think about it, maybe the right mix curve option would fix this as well.
I’ve experienced this scenario also - the level of the effected audio seems very dependant on a bunch of factors - having gain control available on a shift function for the wet signal (with some kind of limiting I guess) would be really useful.
I used Nautilus in a jam the other day and the textures I was getting from it were great but I’d like to have been able to give them a bit more OOMF, it can be hard to mix it in right.
This probably due to ‘adding’ signals digitally in the sound software. In my spare time I program for the Nebulae v2 and although that is a different language, the problem is that signals do not ‘mix’ like on a mixing desk, at least not in CSound (also not with the more advanced chmix function).
Dipped and Constant Power curves could be done by calculating amplitude with a little logarithmic/exponential scale, but that usually ends in “Dipped Blue + Constant Power Blue” (or a combination with the Red lines), which doesn’t sound right. But for the Slow Fade/Slow Cut/Fast Cut, there is no clear algorithm you could use.
Transition I use mostly for balancing between different signal/sound sources (input and waves on the Neb), that seems to work best. It’s actually Intermediate but with a max halfway. For panning (which is fun in Stereo), I use a more Constant Power type but derived from a sine wave function which sounds very natural!
These are crossfader curves, the red/blue signifies channels 1/2, i.e. the two things you’re mixing - ‘constant power’ is the more natural sounding mixing curve used for crossfaders that want to avoid a ‘dip’ in volume at the midpoint - I beleive this is in part due to the exponential relationship of how we perceive volume. What makes you suggest it wouldn’t sound right for the Nautilus?
Maybe this is in relation to the way the sounds are combined - I’m a bit confused about how the way it mixes sounds would be different, i assume there’s a wet and a dry signal and the mix knob makes one louder and the other more quiet? It doesn’t seem to work like a send effect.
Yes I understand these are cross-fader curves. That’s how I ‘cross-fade’ in the Nebulae v2 instruments between different kind of sound sources. Left/right (panning), Input/Wave files (volume balancing), Reverb Dry/Wet etc. When programming them, you want to have the right curve for the right purpose. I just listen to them and try to figure out what sounds best. I had ‘Excel graphs’ like this to help visualize me ‘why’ stuff didn’t sound right.
What it showed and I meant is that when you use a ‘simple’ (log or exp) mathematical function calculating these cross-fade curves, you can not easily use the ‘inverted’ result for the ‘other curve’. Lets say in Dipped, the result of the inverted curve of ‘blue’ is not ‘red’, but it’s the blue curve from the Constant Power. So when your cross-fade is like: “Dipped Blue curve + Constant Power Blue curve”, that combination does not sound right, not only on the Nautilus or Nebulae, but it will never sound right I guess.
Therefor I think, because of Nebulae programming experience, the Intermediate curve is a tradeoff, sound mixing wise and ‘trouble some’ curve calculating wise. It’s a balance between Dipped and Constant Power (probably by design what @andrewikenberry says) and as a plus, it’s the most easiest to implement… but of course, I’m just a enthusiast, not a professional (sound) programmer.
It’s the preferred curve for crossfaders on DJ mixers and volume faders in general use an exponential curve, so manually mixing between two tracks on any desk will be performing the equivelant of a constant power fader. I’m no audio engineer though I’m prepared to be humbled.
I think you’re technically describing a fade curve, which is an ease-in-out relationship (shown on the chart above). These tend to result in the midpoint being perceptably louder, with both sources being at full volume.
The solution to both of these problems is the constant power curve, that’s where the ‘constant’ comes from in its name, as it’s designed to give a level volume throughout its range - ideally it sounds natural (although that depends a lot on the specifics it seems!)
I can’t say I’ve ever come across an effect that has a volume drop at 50% wet/dry mix, it’s stranger still as the design reason for this is that the wet signal was too loud:
The wet and dry signals should be level matched in some way anyway, ideally, so this seems like an odd way to resolve that problem - stranger still I can’t relate to it, as posted above the wet signal if anything is too quiet.
Some gain improvements needed across the board I guess but the current curve doesn’t work whatever the issue is
Thanks for your insight in the thread, I appreciate the way things are coded etc. can have a big impact on what’s possible!
What I was trying to describe is not in the chart… I was trying to describe curve chart like this, where the top and bottom curve in an inverted ‘copy’ when using a log or exp type curve:
Volume wise this doesn’t make sense when using pots with the perception that ‘in the middle’ at 12 o’clock would be an equal amount of volume. In the chart you see that the two curves do not cross in the middle (convenient red dot left over from poor photo-shopping), they cross at about 65% which would be at 1-2 o’clock of the pot (or 10-11 o’clock). That’s what I meant with “doesn’t sound right”, when the volume is ‘balanced’ when the pot is at 2 o’clock, and not 12 o’clock from what you would expect.
Let the Qubit dev’s decide what implementation is best , at least I was not able to apply the Dipped and Constant power ones for a pot. I use Intermediate, Transition and one based on a sine function that sounds really cool in 3D rotation of the stereo field with a ‘artificial’ way of creating near and far sounds.
Ah ok, interesting! Thanks for explaining! I feel that in this case that would exacerbate the problem though - the point in which you have both sources mixed evenly would still be too quiet (as you still only have 50% of each source) but you’ve gained an exponential relationship on the knob - it’s fixing the wrong problem no? haha
This chart possibly explains my point better than I can:
Note how an intermediate curve (which has the same intersection as your curve, just earlier) has a drop in volume at the midpoint. Taking 50% of two sounds doesn’t give the same output as 100% of one, due to that exponential relationship of how we perceive volume.
So your curve’s A+B power level would be a slightly exagerated, skewed version of the intermediate curve:
I’ve had the Nautilus for a bit now and am also not loving the wet/dry behavior. Would it be possible to add “dry volume” and “wet volume” to the attenuators via Narwhal so those behave like standard volume faders? Not sure how that would/should then interact w the regular mix knob, but spose that could just be set centered, or perhaps available for an additional reassignment via Narwhal?
I think the solution would be having a “send” function as a configurable option for the Attenuverters. Dry/Wet could work the same, but you could dial in the “send” level for the wet signal, and even push it beyond the default levels.
This could be the counterpart to the Input Level function found by holding Tap and turning the Dispersal Attenuverter. This also let’s you keep the second Attenuverter for assignment/use.
Though @andrewikenberry’s initial suggestion for fade control via Narwhal is a great option, I think this has some merit for consideration as well!
(on second thought, having Tap/Feedback Attenuverter determine the Send Level might also be a viable option, and would free up both Attenuverter assignments, though having to dive under a layer of front panel functionality might not be ideal for most users)