Unwanted double actuations

Erae Touch
21

ok, so I finally got around to testing :slight_smile:

yes, I can also see the double notes… both on your template, and also a simple drumpad.

however, looking at the data streams its pretty obvious its lack of filtering…
(and if I look at your video @blipson , I can see its the same)

the ‘tell-tale’ sign is the time between the note off and (second) note on…

13:22:30.865	From Erae Touch	Note On	1	E3	15
13:22:30.874	From Erae Touch	Note Off	1	E3	0     <------
13:22:30.878	From Erae Touch	Note On	1	E3	21        <------ 
13:22:30.933	From Erae Touch	Note Off	1	E3	0

we can see that there is less that 4ms between that note off and the note on.
because we have so little pressure (see the velocities which are derived from pressure)
the sensors are ‘wobbling’ between two states - this is basically ‘normal’.
what we would expect is for these to be filtered.

note: the reality is this is probably less that 4ms, since the timing here is host time, so is impacted by things like usb transfer and polling times etc.

as mentioned above, this filtering is a bit of an ‘art’ on any surface, since the physical nature of the surfaces really determines how you filter.

(when Geert fine tuned the Linnstrument , I remember he published some really interesting graphics of how the linnstrument surface responded, and he used this to help tune things like velocity response0 \

basically it comes down to minimum sensor thresholds, and filtering out noise… and obviously in the case of note on/off… you have to be really careful, to avoid both redundant notes off and also false triggers.
pretty much it comes down to smooth the sensor values over time, and choosing that ‘time sensitivity’ is really critical…

unfortunately, it becomes a bit of a trade off , more smoothing = less false triggers, but slower response … and vice versa.

until then…

we are most likely to see this with :
quick light touches , so percussive hits - simply because its bouncing the surface around the threshold.

also Ive noticed you’ll see it more with multiple close touches… similar issue, its a surface, so close touches are never going to be fully ‘isolated’ (ive talked about bleed in previous posts)… thats an inherent in these kind of surfaces.

this has kind of made me realise why Ive not seen it…

most of the time, Im playing melodic lines using MPE, and I tend not to even use note on/off… instead I directly use pressure into a VCA (aka amplitude) , so when I press really softly a the threshold of the sensors I ‘know’ its going to fluctuate… so its not really unexpected, and I kind of accept that as where the ‘limit’ is of this technology… (at least with current firmware etc)

I think the times Ive played with it for percussion, I do tend to tap pretty hard… due to lack of my skills,
and also im not that familiar with where the threshold between a hit is.

overall, I think in this case, its need a bit of fine tuning on the filtering…

note: this is not related to others issues with different sensitivity across the board, that feels like a very different issue.

2 Likes
22

Yes, it’s not accidental double strikes during fast playing, but the Erae’s filtering configuration. Embodme responded to me that playing at the limit of detection (low, but desirable velocities) causes the mechanical/electrical rebound to generate double notes. They’re having me do further tests, then get back to them. I have to reiterate that I hope some velocity response parameters can be opened up for user (or beta user) tweaking so that I can work on dialing it in myself. The proof of concept, at least, would save me from hitting the panic button and asking to return for refund.

@BD2021, I don’t blame you for feeling this way because it’s exactly how I felt until they got back to me after multiple emails. I’m still skeptical as to how this is going to go, but I couldn’t put up with being roadblocked. Maybe asking to return for refund would move them to respond or improve their communications? On the other hand, in other situations, I’ve also cut my losses by selling just so that I could immediately wash my hands of a thing. It’s just that this Erae’s already gone a long way toward working out so that it still seems to have a chance of getting there.

1 Like
23

I think this (reliable) minimum detection point is always a discussion point on expressive controllers.
if you compare things like Sensel Morph, Joue, Linnstrument, K-Board, Roli, Soundplane they are all very different… and I’ve seen similar discussions on most of them, esp. early on after release !

I admit this was the one thing I was nervous about with the Kickstarter, and not getting to try the Erae before purchase… we didn’t know how sensitive it would be (or how close touches can be) , and having seen such variations on other controllers… it was a concern.

once I got it, and tried it, I was pretty happy…
for sure, it’s not the same as a Continuum or Eigenharps, but these cost a lot more - and have quite different approaches (which is why they are so much more $)

so the tech is what it is, and built to a certain price point, and I hope Embodme can bring the most out of it.
also the Erae has a pretty diverse usage, from percussion, melodic to generic controller (x/y , sliders) , its going to be better as some of these than others.

at the end of the day though, it comes down to a simple question for users…
does it do what you want/need it to ? is it useful to you?

of course, this assumes its working ‘within’ spec, and that you do not have a faulty unit.

I think its also a bit tough early on in dev, as we don’t quite know what ‘niggles’ can be ironed out, and what are fundamental to the hardware, so I think thats where its great that you (@blipson) are working with Embodme in this area.

I do feel a pressure (and velocity) curve will help.
what I find currently is the detection point feels a bit ‘on/off’, you need a certain pressure, then it feels like “bang” its full on… it feels a bit linear, over a quite a short travel distance.
overall, I think it needs a exp/x2 curve, such that we need to press harder to get the higher pressure/velocity values.
whilst that might feel you have to play harder, I think it’ll give it a much more ‘progressive’ feel.

its also likely this needs to be user tuneable, as I suspect the curve need to play percussion , melodic lines … is likely to be quite different. (also things like using for modulation is possibly different again)

24

Mine is definitely defective. The notes of inconsistent velocity is across the board with some areas in the middle of the board. In the middle of the board with weaker velocity detection would mean a defect in my experience. Even on the edges a difference between 65 and 105 would still be defective. If the difference was like 15 sure, but not 40. That’s just ridiculous and these weak spots near the middle of the instrument is a sure sign to me that this needs to go back.

Hopefully there is a solution because for me this instrument essentially unplayable. The variations are just TOO great.

25

Yes, it does feel that way. Still, I use the minimum velocity productively by pressing until actuation (in rhythm), which reliably gives velocities under 20 (as I recall), and so it’s acceptable; I’d again say “very good” compared to Zendrum (which doesn’t respond to press, only strike) and far better than Sensel Morph, whose minimum was like 40 before their last update to try to achieve a better velocity curve.

The lack of low velocities was the dealbreaker with the Morph, then the update just gives too choppy a curve. I think it’s simply a lack of processing power in the Morph. A mark II with an upgraded processor would provide a better continuum of velocities and a larger playing surface within tolerable latency, but apparently Sensel’s business is not about musical instruments anymore. The Morph seemed like a proof of concept toy with the serious device to come later, but alas it’s not to be. I do love it’s simplicity, though.

@BD2021 you’re definitely experiencing in a much more extreme form the small and tolerable drawbacks I’ve accepted (less sensitivity at the edges, velocity variations addressable via power cycle). I hope Embodme processes an exchange for you in a professional manner that keeps you on board long enough to see if the Erae at it’s best can work for you.

ADDED: I spent a lot of time setting up my LinnStrument as a percussion controller. Roger Linn told me it would pretty much be useless as a tapping fingerdrumming controller, but it’s not bad. It’s got it’s own feel: the low range of velocities have to be actuated by pressing with progressively greater velocities, then you morph into striking. The problem was that Roger was adamant that velocities under 20 are musically useless, but my Zendrum gives velocities 3-6 reliably, and BFD, with its 64 stereo velocity layers and Roland’s TD-50 with it’s continuum of synthesized velocity response are wonderful to use over the full dynamic range, which the LinnStrument can’t do. Also, the LinnStrument’s utter inflexibility with lighting schemes makes kits of so many percussion pieces (which is why a fingerdrummer buys a LinnStrument) cognitively aggravating. Roger finally relented and added custom user lighting after enough people clamored for it as a musical necessity, not just a cool feature for nerds, but I haven’t taken it out of mothballs for drumming since it drove me nuts about a year ago. I can see giving it another try after working with the Erae a while.

26

the Sensel Morph was always more a ‘product demonstration’ than a product…

they were pretty clear from day #1 that their business strategy was to license the technology to computer manufactures… for things like an advanced trackpad experience. this is why, the initial overlays were pretty varied, from mousepad, video overlays as well as a bunch of musical applications.
I remember asking them during crowd funding about larger sizes, and they said it was technically possible… but even then it felt unlikely they’d ever actually do it.
(I was hoping the professional graphic tablet market might drive them)

I think the music side worked quite well for them… but ultimately, they see a much bigger market than music tech…so really we are just a demo market, showing to investors etc that there is user demand there…

so unfortunately, I doubt we will ever see different form factors for it…
all that said, obviously someone as Sensel likes the music side, as I don’t think otherwise the Buchla overlay would have been made… as this was quite ‘niche’ to have decided on.

Im surprised about that…
fundamentally, its not really about what velocities are useful or not… really its about maximising the use of the 7 bit range of midi over your sensor/surface range. i.e. velocity = 10, does not equate to a dB value, its freely interpreted.
this is why some drum pads you have to hit reasonably hard to get low-ish velocity values, compared to say a melodic instrument.

(slight issue here, is many things like synths assume in their patches that low velocity values are very quiet… so whilst you can alter the preset, thats a bit of a pain if you have to do for every preset!)

the problem for velocity, is few people can reliably hit more than a bunch of velocity values.
so we have 127 steps, but I doubt many can hit reliably more than about 5 to 10 steps (or course some people might!) (*)

so possibly what Roger meant was, when you hit something lightly , your going to get 1-20, you can’t reliably get 5 vs 15.

as such, this is not bad, since its nice to have a bit of variation, and its true of physical instruments too.

this comes back to velocity curves, since using an exp expands the range of those lower velocity values, and so makes them easier to hit, but at the expense of higher values.
… how useful that is depends on application.

outside of percussion of course, sensor curves are more useful, since something like pressure is a continuous movement, so a player really can utilise more of those steps.

in fairness, I don’t think the Linnstrument was really designed as finger drumming instrument, so whilst it might suit some people… I can see why Roger would be reluctant to go too far down that route, if he believed it wasn’t well suited to it.
I respect that integrity, rather than sometimes when we see controllers promising they will be great at everything, and not really doing anything very well…
“jack of all trade, master of none” is not what I want in a musician instrument…

(*) a musical score is often documented with only 7 or so values , mf, f , ff , ff, fff, pf, fp… perhaps percussion scores use more?

27

5-10 reliable levels of dynamics is actually pretty skillful, generally speaking. I have to practice regularly to maintain 6-7 on a Zendrum on my best day. That means 6-7 ranges, of course, not exact values so that “mistakes” within a range become microvariations of velocity that are part of a humanized performance. As for the exact velocity numbers, I should clarify I’m always thinking in specific terms of how BFD responds to percussion controllers.

2 Likes
28

This whole thread has gone quiet over Xmas.

I had my Erae replaced. But this one is going back too. Again there are areas that have greatly reduced sensitivity and when I press Alt + Home + Clef button, I see the calibration map and there are some really bright dots - right where the insensitive areas are.

Maybe there will be a new calibration system setup or something, but this one is also unusable to me. I tried it out with Pianoteq and was disappointed. I work it out by testing individual notes on the keygrid, and notice on a few notes that it isn’t getting good results. I don’t know what to say, but I’m beyond upset. I hope that this was just a stroke of poor luck.

I wonder if anyone has pressed that combo and seen their calibration map, what it looks like. Any bright spots? It would be nice if there was a way to calibrate manually but apparently that option was removed. So now we either wait for a firmware update, or a software update that can take care of this, and that is if it’s not a hardware problem we’re dealing with.

Hopefully the next replacement will be much better. I even asked them to test the next one if they intend to send another replacement. Perhaps I just got incredibly unlucky and there was a batch of them that had this problem and that most of the units were fine. Both of my Erae Touch are defective. I’m almost in tears. I wonder why the replacement wasn’t tested first before it got sent out. Shipping across the globe is expensive, and a simple test might have saved them the trouble of having to send out yet another replacement.

I’m just using my creative thinking about how to solve this problem. I think it’s a hardware problem and I think these two units I have are unfixable. They’re both going back anyhow.

So, these “calibration maps” which show bright dots, means that there’s a problem with the hardware? Or are these dots “re-calibratable”, and if so, why can’t this be done? Or are the bright dots really defective sensors that can’t be “re-calibrated” to a workable level of usability?

29

I share the same kind of concern and hopes since my second unit has shown the same amount of irregularities, thus “better” placed.
My problem is different that yours, the huge variations of pressure sensitivity and range, but I suppose that the causes are the same, as the technechnobear has explained.
Mine units also show the same king of bright dots and the manual calibration in the previous version didn’t improve it at all.

I am afraid that the fact that these problems show only for some users doesn’t mean that it is not more or less spreaded on all units but that most of users don’t need such level of precision … which is good for Embodme but can make us worried a little …

Actually my AT stays in its sleeve waiting for a software update that will allow at least to adjust manually what the sensors are sending, but I am sure that firmware updates will also help to ironise them.
So good luck with your third one!

30

Which indicates a hardware fault. This is unacceptable for a musician who wants to write decent musical parts so I can incorporate them into my mixes.

What your experience with manual calibration highlights is that the sensors are faulty in some way. What it means is that the problem is, essentially, unfixable. I’m seriously considering a refund if the third unit has the same problems because time is of the essence and I can’t be spending all my time worrying instead of performing.

31

I understand …
For me I like to give them the time to deal with this since it might be a complex task and other similar devices have shown improvements after months or even years!
So I hope they are not unfixable.
The more important factor is how a small company can manage researches and support over the time, and what happened with Sensel or actually Artiphon with the Orba has been very deceiving, but Embodme seems to be of another kind?
The fact that they make exchanges shows a good will.

32

ok, just to be clear… the calibration view does not show sensor hot spots!
it shows you the last N touches ( I can’t remember how many off-hand) … even though they are now not active - not what the erae thinks still has pressure on.

the ‘idea’ is you already know when a note is stuck (or such like) so this shows you where the touches were so you can ‘desensitise’ them… so its not giving you any information about ‘hot spots’ - simply what has been touched recently.

I totally agree, this is confusing… and only became aware of this when I spoke to Embodme on Skype a while back… I did explain that I found this confusing.

I did wonder at the time if this was ‘just me’ , as its working very different to the Soundplane, with the Soundplane it shows ‘residual pressure’, so you know the sensors are showing some kind of noise… so I assumed the Erae was doing the same.

what I would recommend if you have concerns, is to see if you can have a support call with @embodme over Skype or something.

I think the critical thing (esp, if replacement units feel no better) is to talk with them to see if there is a mismatch in expectations / experience.
I think a frank and open discussion with them , will also reveal if the issues you are seeing is something that is a hardware issue, or something that possibly resolved in firmware.

as I said, I had a call a while back with them, and found them very open, and very welcome to feedback. honestly, in a time with automated support systems its refreshing to be able to talk to someone knowledgable.

perhaps this will reveal that its ‘not for you’ , but think its worth giving it a go… and frankly probably a bit less frustrating they trying different units in the hope it’ll fix the issue.

33

What you explain about the calibration view is very strange
When I have tried with the first unit to improve the sensors with the calibration method Embodme had told me, the “map” was showing always about the same spots, after extinctions, automatic calibration Playin or not didn’t change.
With the second unit the spots are placed in different areas, but they also match more or less were are the “weak” cells.

34

Apparently mine got damaged in shipping. Slight damage but it renders the instrument unusable for me as a musician.

I’m not sure what to do or say about this, because it is just so frustrating. I’ll need a new instrument because this one is unusable in its current state. I am sure they can be repaired if it is a hardware problem, but there will always be units that need repair for whatever reason.

Shipping is expensive no matter how you look at it. I don’t understand how these are potentially being damaged during the transit. The packaging looks sturdy and strong, so I don’t understand how this could happen.

35

Has this discussion gotten a little confusing? Sometimes it’s not clear if we’re talking about:

  1. manual calibration, Alt/-

or

  1. detection map, F-Key/Home/Alt

On the detection map, the brighter pixels indicate lower sensitivity, right?

36

Yeah they do. It appears they indicate sensitivity thresholds from what I can understand.

There is no way to manually calibrate the Erae Touch.

Been two days now since I last emailed support mentioning the possibility of damage during shipping. I hope they took it on board. Apparently mine was good before it was shipped and defective when I received it.

I’m sorry that things turned out this way but obviously none of it was my fault. The first one I had will be sent back very soon. The faulty replacement will also be sent back once I receive my 3rd one.

Fingers crossed that the 3rd time is the charm.

37

Yes, it generally appears that way, but my most under-sensitive area is clear on the detection map, so that’s not encouraging.

By “manual calibration,” I mean do the same calibration that happens at power on, but via Alt/-, not power cycling. The couple of times I’ve done it to clear an issue, it seems to not clear things up as well as power cycling.

38

Of course, power cycling and autocalibrating by pressing Alt and - keys isn’t going to fix something caused by hardware damage.

My guess is the boxes are being thrown around by careless postal workers. Admittedly I’ve looked at the packaging and while it seems to be adequate there isn’t any foam or bubble wrap, while there is a sturdy box and a two-piece insert that obvously offers decent protection but maybe not sufficient for long distance shipping where things are generally more likely to get thrown around by careless postal workers.

There is also the possibility that the hardware itself is quite fragile.

Now that this had happened twice in a row I am left wondering what went wrong.

The bottom line is if I don’t get a replacement I am left with something that I as a musician cannot realistically use.

Hopefully these cases are in the minority. If I do get another replacement (fingers crossed) and I end up with the same problem and there has been no change in packaging materials then I’m going to have to assume that 1/ the damage is occurring during transit or 2/ the hardware is inherently fragile. I don’t see any other possibility.

40

People are on holidays. I’m not going to post in here for a while. See you guys later.

41

Received 2nd replacement. It has to go back as well. I was literally crying last night. I tried again this morning. Problem still there. Nothing I can do except send it back and hope that I can get yet another replacement. After auto-calibrating a few times, there is one spot that has a very poor response. Might even be just one faulty sensor, I can’t be certain of where the exact problem is. I am pretty sure this was caused by some kind of damage during shipping (AGAIN!!!). It looks like there may have been sort kind of impact on the box, a small gouge in the brown shipping box from what I can see.

I’m not sure what to do now. I’m literally tearing my hair out. This unit is almost perfect, except for that one small area. I absolutely cannot believe this.

I think postal workers in my country must treat international parcels like rubbish or something. When (or maybe I should say IF) I get my 3rd replacement I would be hoping that the next unit be shipped with extra padding or some way to avoid getting any damage during transit.

I feel hopeless. This has me utterly heartbroken. I almost can’t believe this is even possible. 3 times now, and all units likely damaged during transit.

I won’t blame Embodme for shipping failures, but they are hopefully insured against damage in transit. I cannot, absolutely CANNOT use this unit either. So it has to go back. I’m just not sure what to do, aside from smashing my head into a wall. I am really this frustrated by the whole experience. I JUST WANT AN ERAE TOUCH THAT WORKS PROPERLY FOR GOD SAKE!