VI Labs: Modern U Review

CyberGene Is a “multi” file containing all the settings one makes

Hello,

I'm not completely sure whether it also contains "global" settings like the MIDI inputs/sound device outputs.

On the other hand, since my settings had crystallized as I like them, I always launch from a shortcut to the "Multi" file (extension: .uviws) as posted earlier today answering @vagfilm's question. Even automatically from PC cold boot. And then it all just works (given that all peripherals were switched on first, so before booting the computer).

I'd be happy to mail/post my .uviws containing the settings as posted earlier in this thread.

On a pedantic note: I learned a lot from making these settings (the whole VSTi domain was utterly new to me at the time). So using a prefab'ed set may deprive one of some valuable learning… 😌.

Cheers and happy playing, all new and existing Modern U'ers,

HZ

Well, I had to play this, so I played a little in the last 30 minutes. Mainly with the default preset. The first impression is of a very warm and intimate sound, maybe a bit bright in the higher notes, but mostly due to the touch response, not because of the actual frequencies. I put the N1X to a hard touch response to tame that down but it affected the entire keyboard. I will have to play more with the touch response. As a side note, when working on the Cybrid, I noticed that if I use the same velocity measurement across the entire keyboard, this would make the higher notes louder than they should be. I believe the reason for this is, it's easier to throw a light hammer with a higher velocity, however on a real piano that won't result in louder notes because the higher velocity is offset by a lighter hammer (hence the momentum is the same which is what translates into note energy) and also the higher strings are shorter and under higher tension which means they are more rigid and would require higher energy to generate louder volume (might be wrong but the general consideration is the same). Ultimately I had to gradually compensate the velocity curve on the Cybrid and reduce the produced MIDI velocity from the same measured physical hammer velocity the higher on the keyboard I am. Well, that's with Garritan CFX. I'm not sure how other manufacturers program their touch response to sound volume but I've had this problem with Pianoteq and now with the Modern U.

All that being said, it's a subtle effect, I think I can live with it. Another observation I have is the pedaling and re-pedaling are slightly abrupt and unforgiving compared to the N1X and Garritan CFX (with my "fix"). I couldn't always blend between harmonies and create a nice legato melody. But I may have to see the half-pedaling setting.

I enabled the sordino samples to be triggered by the left pedal and it was a fairytale experience! Perfect for lyrical and dreamy pieces.

For now, it seems like a nice piano I can use from time to time. I also purchased it because I will need it for some modern type of music where an upright sound works better and where I don't care too much about very precise half-pedaling, etc., so it's definitely a keeper 🙂

For the moment is still work in progress and you need to prepare spreadsheet-like config files, but my velocity mapper tool can apply velocity curves per key (or decrease/increase volume per key and/or depending on velocity). And it may work in mac, if willing to try it 😉

HZPiano I'd be happy to mail/post my .uviws containing the settings as posted earlier in this thread.

Given their size sounds like something like wetransfer.com is best

CyberGene believe the reason for this is, it's easier to throw a light hammer with a higher velocity,

I believe you are correct with your assessment (even the subsequent one which I am not quoting)

CyberGene Well, that's with Garritan CFX. I'm not sure how other manufacturers program their touch response to sound volume but I've had this problem with Pianoteq and now with the Modern U.

How did you solve it with Garritan and/or Pianoteq? If you already posted it and are too busy working, perhaps a clue on location (PW?) and possible title of the conversation would help.

CyberGene Another observation I have is the pedaling and re-pedaling are slightly abrupt and unforgiving compared to the N1X and Garritan CFX (with my "fix")

If I understand https://www.garritan.com/products/cfx-concert-grand-virtual-piano-lite/ correctly, they fixed this issue eventually? They say

Re-Pedaling and Partial Pedaling
Added realism has been added to pedal behavior. Re-Pedaling “catches” a note with the sustain pedal after it is played and Partial Pedaling allows you to use a continuous sustain pedal to vary the amount of dampening applied.

Did they say that back when you got it?

vagfilm or decrease/increase volume per key

Mmmm, isn't that tool "just" a MIDI thing? How can you change the volume per key? Sure, you can change velocity, but that (at least for regular virtual pianos) will change both the volume and the brightness.

Thanks!

RIP vagfilm or decrease/increase volume per key

In fact, there is a midi command for note volume, from 0 to 127, and by default it is set at the fixed value of 100. But it can be changed on the fly... So, for example, in the VSL D274 the higher velocities are a bit aggressive in metallic timbre; so, I decreased the mapping so it does not enter that top range of samples, and I compensate for that by increasing the volume in those velocities. You need to go back and forth to adjust it properly, but this is an alternative way of increasing dynamic range: mapping key velocities to different sample AND volume curve. I have been playing with linear mapping of 0-127 velocities onto 90-110 in volume. The effect is nice. And you can do that by key, so that bass notes have volume range 90-120, while shrill treble notes have range 90-105 in order to keep glasses intact.

vagfilm there is a midi command for note volume, from 0 to 127

It's been a while since I did look at those detail, are you saying that's a per-note volume? I seem to remember only global volume, not per-note volume. A quick search did not provide me with the answer ☹️

RIP
Yes, it is a CC#7 message, not a parameter in the note-on message, so it is universal. What I do in my tool is that each note-on event results in two messages: a CC#7 setting the volume (depending on the note velocity), and a second message with the actual parameters of the note-on event (note, channel, velocity, time=0).

vagfilm Yes, it is a CC#7 message, not a parameter in the note-on message, so it is universal. What I do in my tool is that each note-on event results in two messages: a CC#7 setting the volume (depending on the note velocity), and a second message with the actual parameters of the note-on event (note, channel, velocity, time=0).

Ah, I see, so it's a trick of yours to hijack the global volume as only for a specific note by changing it right before and right after. Quite astute!

BTW, forgot to mention something I was pleasantly surprised about. I was pretty sure the piano would not be Apple Silicon compatible and would run through Rosetta 2 translation. Well, it's M1-native and runs ultra smooth 🙂 I was able to configure the lowest possible buffer size of 16 which is just crazy and didn't have any audio glitches.

vagfilm What I do in my tool

Hello,

This is quite interesting, and might be of use for @johanibraaten too.

So, the way you defined this, the CC#007 only affects the next note that is played, and not anything that is already sounding at that moment?

In fact, what you propose on a per-key basis is the same as what I attempted to accomplish with Modern U's 'Bend' knobs: velocity curve somewhat more concave, gain curve somewhat more convex (in fact, less concave). That of course applies to all 88 keys, and although a good usable result, yet not fully towards what I was looking for.

Therefore, I was also thinking about intercepting the MIDI messages with per-key corrections, however, wasn't aware of CC#007, so couldn't see how to affect timbre without affecting gain (volume). Which resulted in discarding the idea -- having too much to do as it is already.

So do keep us posted!

Cheers and happy fine-tuning,

HZ

CyberGene Well, it's M1-native and runs ultra smooth 🙂 I was able to configure the lowest possible buffer size of 16 which is just crazy and didn't have any audio glitches.

Cool. My new Mac Pro should be here next week. Can't wait to have the Modern U available at church.

God Bless,
David

HZPiano CC#007, so couldn't see how to affect timbre without affecting gain (volume)

Hi David: it's the opposite - with CC#7 you control volume (gain) without significantly affecting the note character (timbre). I only go as far as ~ 10 % in overall change in gain in a range of velocities, to keep the effect under control without too much "jumpiness".

I have not tested thoroughly, but I assume that CC#7 affects everything that is sounding at the moment in a given MIDI instrument (it is a controller for any volume knob). You can set CC#7 events per midi channel, so it will not affect other instruments in backing tracks, I hope. Moreover, since the (small) change in volume is applied simultaneous with a new sounding note, psychophysically I do not perceive a change in volume on the previous decaying notes, although it is probably happening and it may become counterproductive in some circumstances... Experience will tell. And AFAIK, CC#1 would to the same but since in some applications people may already use CC#1 while CC#7 is almost forgotten, I prefer CC#7...

I think that @johanibraaten mentioned having bought VelPro, which a very good and comprehensive program to manipulate velocity curves, and it is able to do exactly this. The difference between my tool and VelPro is 40 USD, customer support, and good quality vs "whatever-works-well-enough" 😀

vagfilm Hi David

Hello,

Me not David 😉.

Thanks for your further explanation!

Being the purist that I am, I probably wouldn't want the overall volume of already sounding notes/resonances etc. fluctuating with each newly played note.

I really wish there was a solution, though, for having a per-key independent velocity-to-timbre and velocity-to-gain curve/adjustment/influence.

I do salute VI Labs for having separate curves for timbre and gain, that already works very usefully. If that was expanded to per-key adjustments, that would be perfect.

Cheers and happy playing,

HZ

HZPiano do salute VI Labs for having separate curves for timbre and gain, that already works very usefully. If that was expanded to per-key adjustments, that would be perfect.

Yep. I've gotten spoiled with VSL's user interface. Individual note gain/volume adjustment is very helpful with some of their libraries (280VC). I could get way more out of the Modern U if it had those adjustments.

God Bless,
David

David B Individual note gain/volume adjustment is very helpful with some of their libraries (280VC).

Isn't that common to all synchron pianos? The 280VC has more controls?
Anyway, what the VSL (and others) don't have is the possibility of applying changes according to combinations: ie, use velocity AND note to apply a gain (volume) adjustment.
I will try to find some time during the weekend to run some tests and will post.

vagfilm Isn't that common to all synchron pianos? The 280VC has more controls?

It's a common feature in their software. I just find it more necessary with the 280VC. Not so much with the Bösendorfer Imperial.

God Bless,
David

Well, as we say in Bulgaria, "you can't carry an awl in a cloth bag", meaning I can't wait until my vacation and decided to record something with the Modern U. I made my own patch by mixing two of the microphones, applying some EQ to tame down the highs, using custom half-pedal values to avoid the abrupt pedaling (it's pretty good now!), etc and then recorded a Scriabin piece I'm currently learning and still playing from the score and towards the end it's too sloppy and I'm dragging it too much while seeking for the notes... And because I've been so busy at work, I have a total of 1 hour at the piano for the last 2 weeks 🙁

Anyway, here's the recording:

CyberGene I'm dragging it too much while seeking for the notes...

You're being too harsh on you. I am often sloppy, dragging it too much and seeking for the notes myself -- if in that condition I were playing 10% of the quality you are playing, I'd consider myself and accomplished pianist.

On the sound, I can't comment because the kids have taken all the headphones and left me only with computer speakers for now (I could use the NU1 speakers, but I'd need to boot another computer and don't have time now). If you ever record this same piece again with another virtual piano, please share so we can compare apples-to-apples.

In any case, thanks for sharing!

CyberGene decided to record something with the Modern U.

Very beautiful song and nicely played.

I'm curious, since you grew up playing an upright, does the Modern U provide the illusion of actually sitting behind an upright when you're playing it?

In my limited experience, it seems to me that there wouldn't be much difference playing the Modern U compared to an actual Yamaha U3. Of course, I could be way off, that's why I'm interested in your experienced opinion.

God Bless,
David

David B it’s a very familiar experience indeed. My upright is a rather short one and doesn’t sound as rich and “in tune”. If a real U3 sounds as good as the Modern U, I’d love to have one.

That being said it’s an upright sound and I prefer Garritan CFX for those commercial-grade polished grand piano recordings 😉