Helix 3.7: The Freeman Update

[James Freeman]

I think it's one refinement away from being perfect.
The speed and acceleration are good now, it's the slow range that still need tweaking.

EDIT:
To be more clear, the slow range cab be less sensitive and the ramp/acceleration can be more gradual, in my personal opinion.

 

[SwirlyMaple]

I think it's one refinement away from being perfect.
The speed and acceleration are good now, it's the slow range that still need tweaking.

As long as it's optional and we can still keep the current implementation, then I conditionally agree.

 

[James Freeman]

The tweak I'm suggesting will not change the feel, speed or acceleration of the current implementation, I'm talking specifically about the one click sensitivity.
Making things switchable is redundant.

 

[Orvillain]

I actually agree RE: ballistics. I don't really like the change they introduced.

 

[James Freeman]

I think each of us who doesn't like something about the encoder ballistics needs to specifically exactly what he doesn't like, otherwise we have opposing opinions that are not very helpful.

For me the new potentiometer-like speed and acceleration on 0-10 or 0-100% values like amps and pedals feels very nice, it's just the 0.1 steps are overly sensitive and "jump the mark" even when I turn very slowly, or when I press the encoder for defaults it also turns.
With the EQ block it is very noticeable, both the dB and Hz values are very sensitive.
As I said in my initial post in that matter, I avoid tweaking on the hardware as much as possible after 3.50.

 

[Orvillain]

I think each of us who doesn't like something about the encoder ballistics needs to specifically exactly what he doesn't like, otherwise we have opposing opinions that are not very helpful.

For me the new potentiometer-like speed and acceleration on 0-10 or 0-100% values like amps and pedals feels very nice, it's just the 0.1 steps are overly sensitive and "jump the mark" even when I turn very slowly, or when I press the encoder for defaults it also turns.
With the EQ block it is very noticeable, both the dB and Hz values are very sensitive.
As I said in my initial post in that matter, I avoid tweaking on the hardware as much as possible after 3.50.

I hate the fact that it now takes multiple spins to get through the range, and it quite often overjumps after a few spins. In the past, I was able to just quickly dial things in, and even if I couldn't get SUPPPPPPERRRRR accurate down to the 0.1 level.... I could at least easily ballpark it.

Everything now feels way too sensitive and as a result takes a lot longer to get into the zone I want. The sonic differences between 0.1 and 0.2 in many cases are just not apparent enough to warrant this kind of fidelity in my view.

 

[James Freeman]

I hate the fact that it now takes multiple spins to get through the range, and it quite often overjumps after a few spins.

Are you talking about Delay Time?
This parameter also changed after 3.50 but it's very different than the 0-10 parameters.

 

[Orvillain]

Are you talking about Delay Time?
This parameter also changed after 3.50 but it's very different than the 0-10 parameters.

No, it is pretty much across the board for me. It has been a long time pet peeve of mine to be fair. I remember it being a lot better during the 3.2 days.

 

[the-trooper]

I hate the fact that it now takes multiple spins to get through the range,

Did you downgrade? Sounds like you downgraded.

 

[Orvillain]

I'm being an idiot and just misremembering. lmao. Ignore my comments, talking out of my ass. I just turned it on to sanity check myself, and actually I'm wrong. You can easily go from one end of the range to the other. What you can't do is get a nice smooth finer control anymore, and THAT is the problem. So I *think* I'm in agreement with James.

I confess I hadn't turned my Helix on for a month or two, so just forgot what my issue with it was, only remembering that I had an issue.

It is a killer device, no question. I love it, but there are a few irksome things.

 

[Alex Kenivel]

I agree about ballistics. Wide sweeps are okay but the finer movements can be tricky. I got to the point where I can get on with it. What kills me is min/max controller assigns taking WAY too long to get across the range and often glitch-lagging really bad.

 

[Sascha Franck]

It had this before, though. And it worked so poorly that's probably why you didn't notice it

Oh, I'm perfectly aware it had it. Just that it was useless. As said, Zoom has this done quite well.

 

[ragingplatypi]

While I agree with most that the ballistics are great for wider sweeps and more difficult for fine tuning (especially with the press and turn for snapshots), I never gave it much thought because I’ve put up with so many particularities with dialing in music gear for years. Whether it’s a Boss rack unit, Casio keyboard, or a Fender Hot Rod Deluxe (first version volume knob people, c’mon!), I’ve danced some pretty weird dances.

 

[SwirlyMaple]

While I agree with most that the ballistics are great for wider sweeps and more difficult for fine tuning (especially with the press and turn for snapshots), I never gave it much thought because I’ve put up with so many particularities with dialing in music gear for years. Whether it’s a Boss rack unit, Casio keyboard, or a Fender Hot Rod Deluxe (first version volume knob people, c’mon!), I’ve danced some pretty weird dances.

 

[SwirlyMaple]

What kills me is min/max controller assigns taking WAY too long to get across the range and often glitch-lagging really bad.

That's where it gets tricky, unfortunately. A nice fast sweep across a large numerical range is at odds with low-speed single digit precision, and there's also a limitation sometimes with pulsed digital encoders where spinning them too fast causes pulses to be missed and dropped, causing the algorithm to see a slower speed when you twist the knob faster. (Sorta like how aliasing happens with A/D conversion... sorta.)

You can optimize it with code that has some threshold for the acceleration factors, and scales the response based on size of the total range, but there's a limit to how much resolution you can get from the encoder and thus also a limit on the precision with which you can calculate the acceleration and total number of rotations.

Another complicating factor is that you have to sample multiple pulses before you can calculate the speed and acceleration (a la numerical integration or at least a slope approximation, which probably isn't accurate enough to "feel" good), and this has to be balanced with the latency experienced by the user in turning the knob. Sample too many points and you might have a really accurate response, but it won't happen til a second after the user turned the knob, which will cause them to turn too far and over-correct. Sample too few, and you get an error and the response doesn't quite match the rotation. This is probably where the "glitch-lagging" is coming from -- it's probably capturing a bunch of points and then computing the movement afterward.

My worry is that they've already optimized this as well as they can within the limitations of the hardware, and further attempts to improve it may make things worse.

 

[James Freeman]

My worry is that they've already optimized this as well as they can within the limitations of the hardware, and further attempts to improve it may make things worse.

Yeah, that's possible.

The 0-10 encoder rotation takes exactly 360 deg of the encoder's rotation and it doesn't seem to use acceleration, so when you move the encoder back and forth it will return to the same 0-10 value, it feels like using a regular pot on a pedal or an amp which is great.
That probably means there's no way to improve the 0.1 step sensitivity without affecting the 0-10 rotation feel.

On a guitar amp/pedal the distance between 9:00 and 10:00 is something like 5mm depending on the radius of the knob, and obviously on a real amp/pedal I would not obsess over 0.1 steps between 2 and 3, but on a modeler where I can see and adjust down to this resolution I WOULD.
Maybe it's all in my head but when I want the value on 5.0 (not 5.1 or 4.9) I want to be able to do that without testing my patience.

Anyway, I will let this one go, Line 6 are probably not very happy about this "episode".

 

[SwirlyMaple]

Anyway, I will let this one go, Line 6 are probably not very happy about this "episode".

If you keep complaining about the encoders, JCM Santa might put you on the naughty list.

 

[James Freeman]

I always feel like they're about to nuke my Helix remotely.
Next time I plug the USB I'm in for a surprise.

 

[Orvillain]

They could do with some sort of "hold this thing and then turn the knob, and you'll get properly fine micro adjustments" ... I just don't know what you'd use.

 

[Sascha Franck]

My worry is that they've already optimized this as well as they can within the limitations of the hardware, and further attempts to improve it may make things worse.

Not too sure. As said, Zoom has this covered a lot better than the HX family, IMO at least. Now, sure, they use clicking encoders and there's less parameter resolution, at least the latter should make things somewhat easier, but as long as just "auto-accelerate" is concerned, the different parameter resolution isn't too relevant. It's really just boiling down to "how fast do you need to turn the knob to trigger acceleration and how much will it accelerate?". In an ideal world, you'd possibly have 3 settings to chose from.