FLKeys Lachnit MK23

vagfilm Hey vagfilm,

thank you for your very detailed response! 🙂

vagfilm

1. Bottomline, the structure of the keyboard is still a Fatar action, and in Fatar actions many people complain exactly about its structure: non perfectly machined metal rods in which the keys pivot, keybed warping, etc...
2. AFAIK, the keys are not made of wood, but covered on the sides with a thin wood veneer (like the high-end Roland actions). The key itself (and all the pivots) are made of plastic.

1. Yes, I know, whereby I would expect (or hope) that these imperfections are corrected, but maybe they are not because they are integral the design of the Fatar action.
2. That's right, but that's not a problem. IMO the whole "keystick is fully wooden" in digital pianos is also marketing bs. Rolands higher end actions work very well without fully wooden keys, making them significantly lighter than Kawais counterparts.

vagfilm

3. MIDI resolution of 4096 steps. This a marketing mumbo-jumbo. All manufacturers at some point in their calculations end up with velocities with that kind of resolution (because the resolution comes primarily from the microcontroller timer, and not from the touch or optical sensor - velocity of the key is a difference between two sensor TIMES, not two sensor values). It would really difficult (and stupid) to downgrade the microcontroller to force him to calculate times with the 128 range of MIDI values... So, in all keyboards the timings have, at least, 4096 levels of resolution (I assume much more than that), that then are rouded to 128 MIDI steps. Lachnit rounds it first to 4096, and then to 128. Do you think this is relevant? I don't... The main problem with MIDI output is the way sequences of notes are grouped for transmission, because notes are not timestamped at the origin and are sent in sequence: two simultaneous notes on the keyboard (if that is possible...) are always transformed into 2 notes sent one after the other. In a perfect midi controller, those notes should me sent as fast as possible, with minimal gaps of fixed intervals. Many times, those gaps are unequal, and there is some jitter. This is much more relevant than the 4096 steps of resolution.

I agree fully with the second section. I had the feeling that the VPC1 had more of his "jitter" than my FP-30. The timing often just felt a little off.

Not sure about the first one. Now there are 128 steps for me in Pianoteq. If I play mf, I have, let's say, 10 velocitiy steps in that range I can use. Of course I can't exactly intentionally hit for example 80 a few times in a row. It's not about that. But it could happen that I hit a note twice with the same velocity and that would mean I hear twice the exact same sound. With more steps there is much, much more timbral variety (like in an acoustic piano, where it is more or less impossible that two notes sound the exactly same). I know that these differences would be very subtly at best or not even conscious. But it would be a step in the analog direction of an acoustic piano.

vagfilm

4. MIDI range. Again, this is irrelevant and may even be bad engineering. If you force a system to always output in the 0-127 range, you are cutting any events that go below or above the set values. It makes much more sense, from an engineering point of view (and that should be the point of view of a controller), to translate velocity in a linear fashion rather than a realistic fashion. I mean, there is no reason why the controller should output ppp as 0-10, and pp as 11-20, and so forth. Zero velocity of the key is known, and maximum velocity of the key can also be measured (velocity will plateau at some point). That range should be split into 128 equal steps. Translating that into realistic velocities IS THE JOB OF THE VST (OR OF THE PIANO ENGINE), and it is called velocity mapping. Is this how it is done by manufacturers? I don't know for sure, but I assume IT IS NOT because otherwise people complain about supposedly missing MIDI ranges, and so they build a velocity curve right into the velocity calculator (like you are complaining of ppp being 30 in MP11SE; but consider that if you remove the fake let off, then ppp can become 20; and if the key gets a bit sluggish, ppp can become 10. If Kawai had set predefined value of ppp as 5 in a new keyboard, after removing the fake letoff the value would still be 5, and in a sluggish key it would continue to be 5, because it cannot go below that. Is this good universal enginneering? Again, I don't think it is...

I can follow you, but somehow that was not my experience with the MP11SE. Using Pianoteq and playing the ppp, pp or p the resulting midi velocity made more or less no difference - around 30. There was no good differentiation of my touch in the first place. So when I mapped it to the lower velocities to get a better pianissimo experience this low resolution of my touch was just relocated and spread, and I could get softer tones, but the MIDI response was still not as reliable in that range as with PHA-50 for example. I couldn't get it right… I am not able to explain it well. The Problem is not that the 30 - 110 range is to narrow per se. But that ppp, pp, p all output at around 30 from the Kawai. If it was like that 30 is ppp, 40 pp, 50 p, 60 mp, and so one it would have worked better. I don't know, maybe I just have to try it again with the MP11SE.

porco_rosso The Problem is not that the 30 - 110 range is to narrow per se. But that ppp, pp, p all output at around 30 from the Kawai and after that it works and differentiate fine with mp, mf and so on.

Mapping does not need to be linear. Setting a velocity curve correctly is a back and forth process between hitting the keyboard at ppp-pp-p-mp, etc and checking what values are output, and its range; then checking which midi values (output from a midi test file using the entire 0-127 range) result in the piano timbres that you would like to have as ppp-pp-p-mp, etc... Then match both scales... Maybe you will be surprised at the end result, which may be a bit wavy. Just avoid having very steep sections because that will result in uncontrollable playing. But there is no point in having 0-127 range, if 0-40 in your VST sounds always the same, and then 50-80 is very dynamic and goes from p to ff, and above 110 does not sound like a piano at all...
Sorry if all this is well known to you and it is already your approach...

vagfilm I know - my FP-30 Pianoteq curve also looks wavy and not at all linear - because it's too easy to play ppp and pp with the Roland I have a slight rounded "hill" in the lower velocities; then it's back to linear because that works fine with the p - f range, and because it's to easy to reach fff the top end is "rounded".

porco_rosso Does anybody of you has ever seen or played or heard anything additional of the Lachnit keyboards?

The short answer is no, exactly because they are so expensive.

Deleted You realize that 5 grand get you a complete acoustic piano from Yamaha WITH continuous optical sensors, USB MIDI and CFX + Binaural + Bösendorfer samples (VRM) already built-in.

Well, the digital part of that acoustic may not be particularly, good, but that's an excellent point, for why the FLKeys are so seldom seen (well, beside the small manufacturer, but if there were demand, I'm pretty sure they could ramp up production).

porco_rosso I know - my FP-30 Pianoteq curve also looks wavy and not at all linear - because it's too easy to play ppp and pp with the Roland I have a slight rounded "hill" in the lower velocities; then it's back to linear because that works fine with the p - f range, and because it's to easy to reach fff the top end is "rounded".

Frankly, with this comment and many similar others you have posted, I think you can be happy only with a custom made instrument. I am not sure whether or not that custom made instrument for you is the FLKeys, but it might.

One other thing to ask yourself is why are you buying a digital piano instead of an acoustic. Ask yourself this question with an open mind. Then, look at this instrument: https://keybird-instruments.com/ -- would it solve the problems that made you think you could not get an acoustic? It's lightweight, it's inexpensive, it has a good action and good sound (well, at least it is said to be so, I haven't tried it). Could it be what you seek?

Alternatively, I've already told you so, but I'll repeat myself: if I were you (wait! as I told you, I think I indeed am like you 🤣) I would build one to your specs and tune it (pun intended) to your perfect liking. A-la Cybrid, as I am in fact doing for myself. The budget you have is very generous for a DIY project, and as such, you can take some shortcuts, which I am not taking. For example, you can start with the smallest/cheapest upright which has a satisfactory action, improve the latter with regulation, and install the spring-based repetition improvement kit which has been discussed on this forum before (sorry, don't remember its name now). Removing the plate will be complicated/expensive, but perhaps you can leave it in place, as @vagfilm plans to do, or if you buy in a store (easy with your budget) you can ask their technicians to do it for you at an additional fee. Alternatively, you can ask Keybird if they'd sell you only the action, and/or action+cabinet (no strings), which I think they would. Boy, I kick myself for not having asked them, I thought I had not left any stone unturned when I was trying to search for somebody providing a complete action. Even better, we could discuss with them if they'd be interested in building a hybrid piano using the piano-conversion kit. First just as a custom for you and few others (again, @vagfilm ?) perhaps later could be a business for them, so they might be interested in such custom work (which usually companies turn down).

vagfilm MIDI resolution of 4096 steps. This a marketing mumbo-jumbo. All manufacturers at some point in their calculations end up with velocities with that kind of resolution

I agree with the spirit of this comment, but I have not tested any of such keyboard, so in a scientific mind I abstain judgement, and just say all other things being equal I'd take the 4096 instead of the 127 velocity values, but all other things are never the same.

porco_rosso But it could happen that I hit a note twice with the same velocity and that would mean I hear twice the exact same sound. With more steps there is much, much more timbral variety (like in an acoustic piano, where it is more or less impossible that two notes sound the exactly same). I know that these differences would be very subtly at best or not even conscious. But it would be a step in the analog direction of an acoustic piano.

That's a good point, but given how bad (to my ears) are the timbre of all the virtual instruments I have heard, I think it's a moot point today.

Read the comments of the owner in the video below. I think he has more comments in other videos also.

There was a US firm that sold a customized VPC1, but that website seems to have closed.

Unless you use PianoTeq, I don't know if there are any other VIs that utilize more than 128 MIDI velocities.

RIP One other thing to ask yourself is why are you buying a digital piano instead of an acoustic. Ask yourself this question with an open mind.

Okay. I love acoustic pianos. In Germany I could buy a young (80s, 90s) german upright with the size of 116 - 124 cm of a brand as Schimmel, Grotrian-Steinweg, Ibach or a slightly older (60s, 70s) Bechstein upright for 6000 € - even at good piano stores. Here in Berlin someone at ebay kleinanzeigen sells privately a Steingraeber & Söhne 124 (!) for 6999 € from the 80s in perfectly fine condition. There are plenty of nice pianos here. Availability and price are not the problem. I would be perfectly content with the sound and the action of such an upright. I also have the space (even for grand up to 2m).

But I live in an apartment. A total of 9 residential units live in my house entrance, including a friendly policewoman who often works night shifts and understandably wants to sleep in the mornings. A nice young couple lives above me with a baby who is only a few months old. Below me is a very nice single mother with a behavioral problematic, sometimes aggressive son. Everyone is considerate of each other, that basically works well. And now I'm supposed to (and I've already tried it with a rented Hailun upright...) rob the policewoman of sleep after a night shift in the morning? Or wake the baby in the afternoon? Or upset the boy with repetitive practice of individual passages (he's sick and can't help it) that it is even more difficult for the mother? All of this is much more important than my piano playing. And even if you agree upon on a golden time window from 3:30 p.m. to 4:30 p.m., you don't have exactly that time every day, and an hour is not particularly long if you buy an instrument for thousands of Euros. I would have to move away if in doubt. But that really is not an option. Acoustic pianos are just too loud. I could not play them.

RIP Then, look at this instrument: https://keybird-instruments.com/ -- would it solve the problems that made you think you could not get an acoustic? It's lightweight, it's inexpensive, it has a good action and good sound (well, at least it is said to be so, I haven't tried it). Could it be what you seek?

Thanks for the suggestion. I heard of them and really like the concept. It would solve some problems - I think it wouldn't be as loud as a regular one. Maybe tolerable for the neighbors. It's also inexpensive and modular. Very good. But unfortunately there are only 69 keys for now - it really has to be 88 for me (or at least 85…). And I am not 100% sure about the tonal qualities. It sounds good for itself. But I am not sure if classic literature will sound well. But it could. There is a very sophisticated (but not that leightweight) alternative for a piano like this with Clavins Una Corda with 88 keys - for 25000 €. Oh boy…

n-player Read the comments of the owner in the video below. I think he has more comments in other videos also.

Thanks, thats interesting. I will quote him and highlight the interesting parts:

I could write a book... this synthogy american grand is good enough for me, also the pianoteq plug-in is ok. After a lot of research I found out, that it is very important, to have a controller keyboard with fast electronic, the USB/MIDI connetion. We have to factors: At first, if you play a chord, some DP/controllers have big difficulties to transport those notes in time (close together) to the software plugin. Recording them in a DAW you see the notes spread, dispersed. This makes it very unnatural. I feel a big difference to the real piano! Second thing is, even if you play one single note, when will the MIDI- information appear in the DAW? Some are really bad, some not so bad. The MK-22 is definately one of the best concerning this aspects, but it is in my opinion too expensive and the keys are too light, the fingers have not enough counterweight, or better say oscillation weight. It's a FATAR- Problem. I have the VPC-1... there's no Fatar inside, but it is too slow in both aspects, sorry.

My hot favoriteactually: Roland a-88. See how much fun some of the best players in the world have on demo videos for the new spectrasonics keyscape plugin. I'm absolutley sure, they do not only enjoy the plugin, they feel good because the a-88 has very little delays... compared to most other kb. It is quite fast in both aspects. I measured many keyboards by recording them in logic. The a-88 keys are not as nice as the VPC-1 ones, but it is not expensive, not too heavy and it's timing closer to the natural pianos than most other controllers.

Our big problem is, that a real piano needs only 0,5ms, the dispers- aspect does simply not exist! And the big companies seem to have no idea about the problem. If they build a to some degree good controller keyboard, they seem like doing it inadvertently. The more players are aware of this inaccuracy and demand for better keyboards, the better!

Thats so interesting, not even a few hors ago vagflm mentioned this timing-issue and I remarked that the VPC1 felt somehow off to me. And Holger Engel even liked the Roland A-88 (with the same PHA-4 standard action as my current FP-30) best at that time. Btw. that where I felt the NU1X shined - the combination with Pianoteq was just so perfectly seamless in every sense, like an acoustic instrument…

Then I listen to such a recording of this brilliant pianist who uses his VPC1 for years and I ask myself why the VPC1 was not "good enough" for me? Or better: Not significantly less a compromise than the FP-30?

The slow transport of MIDI signals at 31.25 Kbaud serial relates to the legacy 5-pin connector. The potential problem is big chords at fast playing getting bottlenecked; I've never seen any serious research around players actually perceiving this problem. I won't claim it is not a serious problem, and I personally am irritated by increasing VI buffer size (latency).

Whilst MIDI over USB COULD run faster in theory, the MIDI Association wrote that controllers typically throttle to 31.25 probably to guarantee compatibility. (MIDI Association wrote in several posts that they never found a MIDI instrument outputting at a higher speed). So whilst there may be controllers running faster MIDI over USB, I have never seen any evidence.

https://www.midi.org/forum/4704-reduced-jitter-more-latency-with-midi-2-0

It is possible that the internal sounds on a Yamaha N1x, may not be tied to 128 velocities or 31.25 Kbaud serial data schemes. But nobody knows.

EDIT - Internal keyboard sensor scanning & processing will vary by controller, but that should be orders of magnitude faster than the MIDI transport speeds . . .

People don't like to discuss this stuff so I will stop writing. lol

The "classic" MIDI transport takes 1 msec to send a three-byte key-on (or off) message.
Latency in the virtual instrument software and in the audio chain are often (usually?) a good deal more than that.
So I don't worry about MIDI latency.

I wonder if you can simulate the "feel" of the action by removing the rubber dome strips from a typical DP action, to eliminate the resistance from that interface? If you did that with a section of the keyboard, you could potentially feel the difference between a rubber-contact sensor system and what a "non-contact optical sensor" would feel like on a given action.

I wonder if you could buy a 10 yr old K200 silent piano to use with a VST for 5000€?

TomLC I love this idea. A few years ago, I was looking for used pianos with MIDI/silent functions exactly for this purpose, on the assumption that the electronics are quickly deemed "obsolete" and you could maybe find one for cheap. I didn't see anything that stood out as an absolute steal though.

Would love to know if anyone here uses an acoustic almost exclusively in silent mode with a VST.

porco_rosso Yes, I asked myself why they didn't use a GF action as a base… would have been even better for sure.

It's been done. Ravenworks VPC1.

EssBrace It's been done. Ravenworks VPC1

Hello,

😲 Ravenworks don't replace the VPC1 RM3 Grand II with GF actions do they?!

Cheers and happy playing,

HZ

Gombessa Would love to know if anyone here uses an acoustic almost exclusively in silent mode with a VST.

Hello,

Gamma1734 (who not yet made it over here as it seems), if I understand correctly, solely uses his Kawai Kx00 upright as his MIDI controller -- and loves it that way.

Cheers and happy acoustics,

HZ

HZPiano Ravenworks don't replace the VPC1 RM3 with GF actions don't they?!

Of course they do... But if the ravenworks vpc1 was 3 times the price of a vpc1, they only need to buy a ca79, strip the action, replace the sensors, and charge you 2 times the price of a nv10... For that price, I can do it too 😊😊😊

Ravenworks Digital VPC1 details

https://nebula.wsimg.com/7563cfd9d30c030a316233d822a5f31e?AccessKeyId=FF959A32CC82E300548C&disposition=0&alloworigin=1

n-player Whilst MIDI over USB COULD run faster in theory, the MIDI Association wrote that controllers typically throttle to 31.25 probably to guarantee compatibility. (MIDI Association wrote in several posts that they never found a MIDI instrument outputting at a higher speed). So whilst there may be controllers running faster MIDI over USB, I have never seen any evidence.

I would be interested by references. The only reference I have is the MIDI over USB specifications where I read :

The USB transfers MIDI data at rate hundreds of times faster than the original MIDI 1.0 hardware specification. In addition, this specification takes advantage of the USB's bandwidth and flexible data handling to enable the transfer of many "virtual" cables worth of MIDI data.

It seems weird but not impossible that a MIDI controllers throttle their USB port which are designed for communication to a PC (if not named « to PC ») which can handle high speed streams.

Let’s hope MIDI2.0 will make (soon) controllers compatible with higher speed.

vagfilm Then, how would it be to just buy a CA79 and give it a custom-designed vinyl wrap instead? 😁

And if you leave the legs, pedal bar, and rear panel off, it makes a nice back-breaking stage piano just as well 😬

Cheers and happy useless ideas,

HZ

Frédéric L

https://gearspace.com/board/showpost.php?p=13786652&postcount=37

n-player I am also very interested, but the references you provided are just "a random guy on the internet" aren't they?

Most importantly, bandwidth is not speed. Speed is (the inverse of) latency. You can have (and in fact often do have for internet) extremely large band connections, which are horribly "slow" from a latency point of view, to the point of not being usable for anything "real time" (such as video conferencing).

Speaking of bandwidth: suppose one plays semiquavers (sixteenth note) at a speed of quarter-note=180, i.e. 3 quarter notes per second and hence 12 sixteenth note per second. And suppose that one does that with 8 fingers at the same time, or 96 note-on per second. At 3-byte per note-on message, that's 288 bytes per second, counting also the note off it's 576. Put in pedal, double the tempo (if you can play that way) or whatever you want, but that's a far cry from the 31.25Kbaud limit! So as far as I can tell there is absolutely no problem in bandwidth -- however there can be in latency. Even though as @MacMacMac and others have said it's not the MIDI transport the main source of latency.