Nord User Forum

I can confirm that option B is the case. If you play a midi file to the NS3, switching KBD TOUCH clearly changes the sound.

Also, you're right about the extern Keyb Velocity parameter. Switching this to Hard gives more dynamic control. To actually do this, you have to set Local OFF and MIDI chan Prog to 1 (Extern menu) and Panel A to chan 1 / Global Off (MIDI settings)

However, the NS3 doesn't respond to other input then (switching on effects or the like or program changes, turning knobs, nothings works anymore).

Streef wrote:Also, you're right about the extern Keyb Velocity parameter. Switching this to Hard gives more dynamic control. To actually do this, you have to set Local OFF and MIDI chan Prog to 1 (Extern menu) and Panel A to chan 1 / Global Off (MIDI settings)

If you turn Global Off and instead enable MIDI transmission via an EXT function, then yes, of course, the EXT functions will be used. However if you leave Global enabled, and do not enable any EXT transmission (or not on any channel the Nord is set to respond to), none of the EXT settings should be relevant.

anotherscott wrote: It is not true that "everything that is sent out MIDI goes through this."

Agree. To my modest defence, I had a later "disregarding the global output options" caveat

I will readily admit that I haven't played around with these functions too much since I tested it quite extensively at or before 1.04 -- which had a lot of bugs. So I simply wasn't sure about the present state of things.

baekgaard wrote:disregarding whether you're in global MIDI mode or only use the Extern -- there was at least at some point a difference/bug so that you could get both "raw" output with no curve applied and then a curve applied one too

It would be a bug if it did NOT behave that way.

As far as I remember, the bug was that you could not switch off the global output even if you tried to (or some such). But it works now as is intended, sort of -- but not how I think you expected...

In theory, you should be able to transmit three different curves from the keys simultaneously over MIDI:

You can, but not entirely the way you described it... and this could be a small part of the reason why it's somewhat confusing to others too?

There are 3 velocity scaling options in the menu that can be applied to the extern A and B output. The global output has no selectable curves.

The default scaling for the extern setting is the "med" value.

The "curve" that is always in effect for the global section corresponds to the "hard" option.

Note that "hard" transmits the lowest value: If you play a key that is transmitted with velocity 40 at the global setting, then hard gives you the same velocity 40, mid sends out 59 and soft 73 -- so selecting soft means you have to play really soft to get the same sound as hard gives you with harder playing.

This means that for those testing loopback, there will be a difference (even with default settings) depending on whether you do a global loopback (as most probably do) or use the extern section to send out midi data back to the keyboard.

Obviously, what the keyboard plays should be the value it will play when the same velocity notes are received via MIDI. I haven't (yet) tested whether there are some settings where this happens -- maybe the default (global) curve should have been something else than hard?

As also noted above, it means the setup is as follows (now that I've had time to test it):

Keyboard -> Global MIDI Out (with a scaling that corresponds to KEYB VELOCITY "hard")

Keyboard -> KBD TOUCH (0, 1, 2 or 3) -> Piano Section

Keyboard -> KEYB VELOCITY PANEL A/B ("soft", "med" or "hard") -> Extern MIDI Out

Midi In -> KBD TOUCH (0, 1, 2 or 3) -> Piano Section

... so it is essential (at least) for the MIDI playing via sequencer that you have the same KBD TOUCH settings in the Piano section.

Note that the KBD TOUCH settings DO NOT change what is sent via MIDI out (I tested this quite rigorously also).

Corrections are of course very welcome!

baekgaard wrote:The "curve" that is always in effect for the global section corresponds to the "hard" option.!

Ah, I had assumed the global default would correspond to the middle/medium setting, thanks for the correction. It's unusual that all the "adjustment" from default only goes in one direction! I'll correct that in my earlier post.

baekgaard wrote:Note that the KBD TOUCH settings DO NOT change what is sent via MIDI out (I tested this quite rigorously also).

That is surprising, but is at least kind of philosophically consistent with Streef's finding that "If you play a midi file to the NS3, switching KBD TOUCH clearly changes the sound." You probably wouldn't want the KBD TOUCH function to apply to both MIDI Out and MIDI In, since the Touch settings would compound each other if you recorded into a DAW and then played back (and it would definitely create a problem for our loopback scenario). Though if I had to pick one, I'd have picked putting it on the Out, so that it naturally, automatically records the performance just as you played it.

So I'd agree, "it is essential (at least) for the MIDI playing via sequencer that you have the same KBD TOUCH settings in the Piano section." It's too bad that there appears to be no MIDI command associated with the KBD TOUCH button (unless maybe it's sysex and so not documented in the CC chart in the manual). If there were, you could at least select the desired Touch setting at the start of your track, and know that it would be played back at the desired setting (the one you recorded with) as opposed to having to remember to make that setting manually on playback to assure that you get the same sound/dynamics that you performed with (if I understand what's happening correctly).

I just made a recording of a few single notes at different velocities on the NS3 88HA.

Audio and MIDI was initially recorded at the same time, using local mode On and with no MIDI routed back to the NS3 from my sequencer, so the first audio track was from me playing on the internal keyboard.

Then the MIDI track was played back to the keyboard, and I recorded the audio on the second track.

Comparing the two tracks, they were virtually identical (almost down to the sample level). Analyzing corresponding parts of the two tracks showed that they were within 0.1 dB for all measurements I could do (L/R, Peak Levels, RMS levels, different sections, etc).

So apparently on my board, there is NO difference between 1) playing locally and 2) routing MIDI out and back again to the board -- at least with the settings I used, which were KBD TOUCH 0 and a MIDI "channel global" mode (which would correspond to using the Hard settings if one would record MIDI using the External Section routing instead).

If one would record MIDI data sent out via the External Section using default settings, the results would be different, as a different curve is then applied to the MIDI data before it is sent out.

Very interesting, baekgaard! On one hand, it's good to know that MIDI OUT-->DAW, DAW-->MIDI IN yields results that sound like the original playing, that's just as it should be. So then the question is why people are seeing some of the other unexpected results with MIDI input. Since we've established that KBD TOUCH (unexpectedly) affects MIDI IN, maybe the setting of that button has something to do with what people are hearing.

anotherscott wrote:... Since we've established that KBD TOUCH (unexpectedly) affects MIDI IN, maybe the setting of that button has something to do with what people are hearing.

That (e.g. different settings of KBD TOUCH when 1) playing vs 2) playing back) -- or some may also get into trouble if they record/play MIDI using the Extern Section, where the default is not Hard? Or something else entirely

Also, for people using other boards to control the NS3, I think the combination of settings and the resulting scaling can cause a bit of confusion.

I kind of ran out of time here, so I didn't double-verify that KBD TOUCH works exactly the same way for keys played locally and via MIDI in by doing the same experiment again with all 4 settings of KBD TOUCH. But it would be really weird if there were two different "chains" through the Piano section that applied almost identical but exactly the same scaling; based on what we know currently, the KBD TOUCH affects everything coming to the Piano section -- whether from local keys or MIDI in.

But it's actually nice that the NS3 has options for scaling velocity both going out of the board and coming in to the Piano section. Lot's of options there!

I have to admit I get a little confused . Basically, what Baekgaard writes is that there should be no difference if you play piano from internal keyboard, or if you play it via the MIDI cable routed directly from MIDI out to MIDI in (with MIDI channel global, not using Extern). However, to me it feels that there is a difference between the two !? I would describe it that to me it feels that playing internal, there is a Med setting, while playing via midi loop cable, setting is Hard.

However, the test from Baekgaard seems to prove otherwise (i.e. there is zero difference).

@Streef: Yes, it appears that playing with the Hard setting / through Global MIDI is equal to playing internally -- at least when I tested it yesterday on my board (running 1.24) it appeared so. Maybe there are some parts that I have not controlled for or have forgotten?

I often find it hard to judge differences just by playing when you know what you do/expect (i.e. you're being biased), so I tried doing some "blind" tests as part of the process. Also, to verify that there would be no differences in the transmitted velocity between the KBD TOUCH, I played more than 100 notes and used statistical inference to conclude on the variation within similar condition vs between conditions, seeing whether any difference was just me being a "random" contributor to the experiment (think variance and standard deviation, SEM, t-test, Bonferroni corrections, ANOVA, ...)

As to whether there is difference between playing the board directly vs input via MIDI (local off, global channel loop back), I think that could be well evaluated by playing something (something simple, but with lots of dynamic variation, like repeating single notes at increasing volume, at different points in the keyboard) while simultaneously recording the audio and the MIDI into a DAW. Then play that MIDI recording into the Nord, and record the resulting audio into the DAW as well. Then compare the waveforms of the two audio tracks. (You also need to make sure you use the same Kbd Touch settings for both your "live" audio recording, and when playing back the MIDI track into the Nord.) For simplicity's sake, I'd use the mono button and record just a single mono feed, so you'd only have two waveforms to compare rather than four.