Forum

Congratulations on getting your Alpha, enjoy it!

Are you sure you haven't accidentally selected multiple splits at the same time? Press the main mode key and ensure that only one split is selected, otherwise you may end up with one keygroup 'on top' that you see the lights for, and another underneath that plays notes?

What Oberon describes does not sound like a planned feature.

Well so far (3 days) I haven't found this so much of an issue, yet. But I started this thread the same day my Alpha arrived. There's something wrong with either me or my Alpha.

Sorry, I think that was a false alarm (at least in 2.0)

More thought...

Hi Stefan

The problem Jim is talking about seems to be connected with the keygroup light blink. This can leave you to believe that you're in selection mode when in fact you haven't pressed the mode key hard enough to enter it. We may get to doing a fix for this in 1.4, but since it seems to be very easily avoided by pressing the mode key firmly as a matter of course (and only firmly enough that it seems you may be the only person who has ever pressed it lightly enough to notice this effect - you must have a very light touch on your instrument - we're all obviously thugs by comparison) it isn't on the top of the list of 'things to do' right now as we're really trying to get the next release of 2.0 out and there's a very pressing list of unavoidable showstopper bugs that need attention before all the ones that can be easily overcome in playing.

If this is a more serious problem (ie, if you press the mode key firmly and it is still occurring) then please let us know - we couldn't reproduce it here without playing the mode key very lightly indeed but it's quite possible that it's not this effect at all, it could just be a coincidence and we could have formed the wrong conclusion about it. If so please let us know and we'll look at it again.

Quite a bit of this stuff is being changed (work ongoing this week in fact) in 2.0 and it's more than likely that this effect will naturally disappear in that series - we combining the key number signals and the activation signals into one for the next release as it's become an obvious source of setup wiring errors when they are kept separate.

John

HI Stephan,

It's probably easiest to check this with bcat (since you were at the DevCon, we've introduced you to it already):

/usr/pi/release-2.0.36-experimental/bin/bcat -idDS '|alpha_manager1/1|#17'

(replace | with left and right angle bracket)

Should show you the current key data that is output from the Alpha. You can replace #17 by #2, #3 and #4 to look at pressure, roll and yaw. Doing it this way, you actually looking that the data combing straight from the keyboard, instead of looking at data coming into another agent, which goes straight to the source.

Let us know what happens.

Take care,

Geert

Hi Stefan

This has got us seriously stumped here. We can't reproduce this behaviour at all, however hard we try and we've had a really good think about how the instrument firmware could even let this happen. We really have no idea beyond thinking that starting a note very very slowly could make the hysteresis and thermal drift filters behave oddly. We can't get this to happen though, so it seems unlikely. The only other thing that we can think of is that you may have some strange or corrupted data in your calibration that is interacting in some bizarre way with the firmware. This would be very odd though, we've never seen anything like it in years of development , testing and shipping Eigenharps. Not to say there isn't a first time though.

The next step is probably going to have to be us bringing your instrument back here to look at it. Before we do that could you possibly try something? We'd really like to see this effect shown on video, in case we're missing something that you're doing in our attempts to reproduce it here. If you have any way of making a visual record of it, it could be a big help and there's always the chance it'll avoid depriving you of your instrument for a while.


John

Thanks Stefan, that's very strange to watch. The keys obviously activate as the light comes on (which is done in the firmware), which tells us something immediately. Quick question if you have a moment - does this also happen with 1.4.12 using a Factory Setup and if it does could you show us? This is quite an important question as there are a lot of code changes from 1.4.12 to 2.0.36 and it would be very helpful to know if they have had any effect, it will really help us try to isolate the problem.

John

It doesn't seem to be a problem of EigenD: http://www.youtube.com/watch?v=67Yl-SGh_EE
I switched to AU 1, because you probably couldn't hear when I played the piano softly anyway.

Maybe, could it be that this is the very Alpha we saw at the DevCon? And something went wrong when you demonstrated the calibration process? Maybe someone touched the Alpha during the calibration or something.

Btw, you may want to update the year of the copyright notice in the about dialog ;-)

Hi Stefan

We finally got to the bottom of your problem today, and it was about as weird as it comes with the Eigenharp. It's quite hard to reproduce on your instrument although after several hours of trying we were finally able to do it around one key press in ten or twenty, if we were very careful. We also managed to reproduce it on other Alpha's, though it was much much harder, to the point that it took many hundreds of even more careful key presses to produce the same effect.

Jim finally tracked the problem down to the key debounce algorithm. This disables a key for a short period (20 milliseconds) after it deactivates to prevent very short key activations caused by vibration and rapid movements near the keys. Without this we found a lot of spurious notes got triggered as a consequence of the key sensitivity. The way we were able to reproduce your problem was by brushing the key very very lightly on the way to pressing it - the brushing seems to set up a tiny sympathetic resonance in the key/spring system that activates the key debounce, which then ignores the subsequent key press as it considers it to be noise. We think that the reason it is easier to see in your instrument (and by easier I really mean 'possible' as it's incredibly difficult to get this effect to happen on other instruments) is that it has hard maple keys - these have a slightly different mass to ebony or plastic keys and we think that the changed natural resonant frequency of the key system is interacting with the key debounce window in a different way. We have made very few Alpha's with those keys (which are very nice but also very expensive and difficult to make), which is probably why we haven't seen it before.

Fortunately, the debounce window is an adjustable parameter in the firmware. Jim has just committed a code change in 2.0 that brings this up to the user level as a control in the Alpha manager. When we adjusted it to 25mS instead of the normal 20 we could no longer reproduce the effect on your Alpha. You should do that in your setups to avoid the issue in the future. You won't notice the change from a playing point of view unless you were planning on being able to play notes at better than 40Hz on the same key, which we think is probably unlikely.

We've booked your instrument in to be shipped back to you tomorrow, hopefully you'll have it for the weekend. Please accept our apologies for the hassle caused. Every now and then we get a reminder of how sensitive the Eigenkeys really are, and this one was a doozy.


John