So does anyone know why Setbfree doesn’t seem to register what velocity my midi keyboard plays? The velocity is registered in the midi log and other instruments does register the velocity. Is this just a limitation of the plugin itself or am I missing something?
setBfree emulates the B3 organ faithfully, including the lack of velocity. Most organs do not react to velocity. setBfree also emulates the B3’s percussion sound well, applying the hard attack only on the first note pressed.
I’ve previously gone on a rant about how midi controllers suck for playing B3, specifically because they all require you to push the key all the way down in order to sound the note, whereas the real B3 (and any good emulator hardware) gives you the full velocity note after a pretty small amount of travel. This difference makes it impossible to play a lot of famous B3 parts, if you’re in a cover band and don’t have a device that works properly.
I own a Roland VR-09 which is actually a fine B3 emu in its own right, with keys that respond properly in organ mode. I am under the belief atm that this is the cheapest way to get real organ action on a consumer device; no USB controller that I have looked into includes an “organ mode”.
Yep! I can’t understand why, but no MIDI controller in the market can do this “simple” thing. Perhaps it’s not so simple? Perhaps there is no demand because most players don’t care about this? I think there would be enough demand if there is some supply. At less i would buy one at a reasonable price
Regards
Yes it would be really cool to have a keyboard controller that operated it’s contacts near the top of its travel. Even better, one that allowed you to configure the action. I doubt we will see this though.
Aren’t there some keyboards that use the time from the top contacts connecting, to the time the bottom contacts connect to derive the velocity value?
If so, and if such a keyboard had open source firmware , it seems like it would be possible to use the top contacts in a kind of organ mode by modifying the firmware.
I’m not saying it would be easy, just possible.
Not necessarily “affordable”, but I have a Crumar mojo double manual that has true organ action. Interestingly, the lower manual can be configured to act as an organ or a touch sensitive keyboard. I believe the manuals are a customized fatar tp-80.
Inside the Crumar is a mini pc running windows CE and using VB3 CE 2 as the B3 emulator, which sounds better than setbfree.
It is an awesome controller and I like it better than the Hammond clonewheels, I haven’t played any others though.
In the computer keyboard world, there are magnetic keyboards that uses Hall effect sensors. They allow the user to configure the travel distance to fire up the event, some starting as high as 0.1 mm. And they are not too expensive. I presume there should not be that difficult to use this kind of switches in a MIDI keyboard…
Check this as an example:
That is how many (most) keyboards work but the contacts are mounted too low in the key throw / travel, often at the end. The organ action has a single switch contact that triggers right at the start of key travel allowing “tickling” the keys. This is how greats like Jon Lord were able to play such rapid and complex music.
That could be a really good solution but likely to be rather expensive. A Hall effect sensor plus magnet will often cost more than a couple of switches but can have longer life.
When I was (very) young I bought a keyboard from an electronic surplus supplier which probably came out of a VT terminal. It had magnets and coils on each switch. I had no idea how to use it at the time so it probably ended up in landfill.
Thanks for this discussion, I always wondered why when trying to play organ (Hammond) sounds it never felt right, I put it down to my piano fingers. Now it makes sense having the sensors higher would make a difference. I think I may have something of interest to further the discussion. This mkcv64smf MIDI Encoder for Fatar keyboards plug and play looks like a good way to go. Particularly with the mention of “NEW: scanning only upper contact with constant velocity (units produced since July 2023)”
It was cheaper to use hall effect sensors on my 13 note pedal board than it was to use switches of some kind…
My theory: conventional wisdom is that electronic keyboards use two switches/sensors to measure velocity - one at the top, and one at the bottom. I believe that my Roland stage piano does that, hence the ability to emulate an organ correctly.
However, I think the vast majority of, you know, affordable midi controllers just have a single, pressure-sensitive button at the bottom of the key’s travel.
One of the best controllers I ever had was an M-Audio Axiom 49 - this line is notorious for having a small timer IC in the power supply fail, resulting in the appropriately-named blue screen of death, and that’s the current status of mine. I attempted an RC fix but it didn’t fix it, but that’s probably my ineptitude at picking the right R and C. :>
(edit 2: the keyboard is on a shelf, against the day that I can acquire an oscilloscope with two probes and measure the damn timer myself…)
Anyways, while I had it open I looked closely at the key mechanisms, and it looks to me like it’s just the one strip of rubber-covered buttons at the bottom of the key travel, which also provided a nice aftertouch functionality.
It’s probably down to the fact that any serious organ player is probably gonna have a real Hammond device anyways, and the engineering time and extra parts and support expense are not worth it, or would increase the cost beyond the tolerance of your average consumer of $300-500 midi controllers. Even I am constantly debating whether I’m getting an XK5 or an XK Pro or a straight up real B3 and Leslie, once I have about $6k saved up. I’m not even a serious player… yet.
edit: Or maybe one of these “Crumar Mojo” devices that Long & McQuade seem to have never heard of. They are truly an aggravating store to deal with, and also the only game in town these days here in Manitoba…
I don’t know, but I feel like the hall effect sensors would also be less subject to bouncing? Or do you still need to debounce? Debouncing still gives me a headache, even if the library just works my brain has to attempt to do the math/visualization and I literally get a hurting head when I’m coding for it…
No de-bounce required.
I tried several mechanisms and the hall effect sensors have proved to be excellent. You can either take the digital out and use that as the trigger which obviously isn’t velocity sensitive. This suits the pedalboard. Or use the analog to do your own dynamics from the analog output, but do be careful which modules you get. Modern copies might provide only the digital output…
If I was doing it for a 61 note keyboard or similar I would build it all from discreet components and design sub boards which I would get manufactured by a PCB supplier. They generally have a minimum order so this approach plays into that.
Using a 4051 CMOS chip to multiplex the analog would seem the way to go and feed the individual multiplexers into an arduino, with a commoned digital address bus to address the individual sub boards.
Some keybeds have two pairs of contacts on a PCB that the key pushes a carbon contactor onto. Because of the angle / arc of the key throw, one of these contacts is made before the other, hence the ability to measure the velocity.
It would be challenging to multiplex and scan Hall effect sensors fast enough to derive the data required to calculate velocity and after touch. It would probably work better with dedicated microcontrollers for each bank of keys. Microcontrollers can be very cheap so that child work.
Agreed, and introduces a rather pointless analog infrastructure…
12C linked or something else?
Would seem to be sensible to generate a USB based MIDI output from somewhere, but could one USB carry the 6 or so microcontrollers? I always feel with these projects I spend far more time perfecting power supplies rather than code and processors, once of course I’ve got the signals going to the right places.
I would use a microcontroller to aggregate the key sensors and interface with USB / DIN-5.
It seems at least that the fatar keyboard has some documentation https://doepfer.de/DIY/Matrix_88.gif
Here is an opensource project on Bitbucket that seems to do the job with fatar and could also be easily adopted to having an “organ mode” which would only use the top sensor. Of course with MIDI 2.0 taking off perhaps there is eventually scope for upgrading my very old Doepfer LMK2+ and not having to throw it away.
Scanning key matrices and calculating velocity (or ignoring it) is fairly trivial. It’s the position of the sensors (which is a hardware issue) that is relevant here. Most keybeds have the first sensor to low which means it is impossible to sense key press until the key is pressed too far. The only solution is to have a keybed that detects key press near the top of its travel which organs do but synths and pianos do not. I have tried lots of keybeds and not found any with the desired hardware sensor design.
Using a Hall effect with analogue reading may work but it may be challenging to calibrate. Maybe on power-up the sensors are read to calculate the rest position values which may be sufficient, assuming the value changes fairly predictably through the travel.
I’ve just measured the position of note off, my synth fatar tp9 is a not acceptable 5mm, interesting my fatar tp10 measures at 3mm. Of course after just reading through the manual https://hammondorganco.com/wp-content/uploads/2011/03/B3mk2.pdf there is no way can get anywhere near the real thing with it’s curious side-effects see page 62.
I’m wondering if anyone has tried Fatar 61TP/8O which is specifically build for organ projects, example of price a Doepfer Kit which includes all components MKE Info for only Euro 300.00 (okay you’ll still need to build a case)