Sampled Pianos NOT up to par for playability AND sound in one package

[Jake Johnson]

heh, and it's $49? that's a killer deal! TSO is probably built from TBO samples right? That's pretty cool.

For me, I fine tune my velocity with a simple custom Jesussonic script I wrote in Reaper. It's just a nonlinear scalar so it's the usual exponential curve, just that I can dial it in with extreme precision.

kensuguro--

I thought jesusonic was just for creating scripts for effects. I didn't realize that you could do work with velocity and create a graphic interface. Can you point me to information about working with jesusonic in the way that you've used it here? I'd like to see what else can be done.

[K-Sounds]

The picture quoted in Jake's post reminds me of the interface for K-Sounds' new Signature Piano. Sound and playability were top priorities in developing this product, so I provided a graphic interface where users can draw their own velocity curve. When the instrument is saved, the custom curve is stored as well.

That part of Signature Piano's interface looks like this (without the caption):

[Stefan Kortner]

Hello maestro2be,
I was in the same situation two years ago, I didn´t find a piano that was close to the real thing. That´s why I recorded my own ones... :-)
I am new in this forum, and I don´t know if it is right to talk here about my pianos. But two years ago I was really frustrated because I bought some well-known grands and playing was not fun.
They all sounded nice, but their behavior was not smooth. I located the root of this problem in the fact, that all samples were played in by hand. That means, somebody sits in front of the piano and hits key after key, with a velocity that is more random than precise.
I thought, it must be possible to hit all the keys of a velocity layer with a precise velocity. That is definitely not possible by hand - I tried it. I looked at the MIDI output of my Yamaha-P80 and tried to hit velcocity 70, for example. The results (after 10seconds length of the previous touch) were somewhere between vel55 and vel85. Completely inacceptable.
So I started dropping objects onto the keys and measured the preciseness. They reflected from the keys and touched the key a second time. Inacceptable. Then I constructed a hollow round thing with an inner water-damped mass. What a mess... but much more precise.
Then my father gave me a heavy, old electromagnet and said, try this. You can give it a voltage up to 40 volts, depending on what force you want it to make. That was the solution! The magnet works as the finger that presses the keys. The precision to hit a certain velocity is within +-2. Then I spent much time to make this thing noiseless.
But the result was great. First I recorded a Boesendorfer with five velocity layers. The preciseness made it possible to program the Kontakt instrument very easily. No specific treatment of individual keys necessable. Each velocity layer has its filter settings, and that is enough. Playing the loudest note of one layer sounds exactly the same as playing the softest note of the next layer.
Then, my observation was, that a velocity curve is not sufficient, it has to match a corresponding brightness curve (that´s what I called it).
What am I talking, play it yourself.
When I started to make these recordings, I planned to do that for myself, to have an instrument that satisfies me. I am not a musician but an engineer, but I used to take a lot of classical lessons during my school time. I played all the great piano brands in all kinds of concert halls, whenever I had the opportunity. And now, when I play my Schubert Impromptu on my recorded grands, I am happy again.
Meanwhile, I also offer them for sale. Believe me, you can´t get rich doing that. But I am extremely satisfied when I see that someone in Japan plays my piano, for example. And I am very happy that I made it - despite all the comments of the so-called experts (whow, YOU want to record a grand? You need microphones worth thousands of euros, and the programming is very difficult... forget it!)
Again, I don´t want to abuse this forum. But I see now that there are others who would appreciate a playable piano. I recorded four of them. A Boesendorfer, a Steinway-B, a Kawai, and recently a Bechstein with my friend Roman Beilharz. For all of them, I made demo versions for Kontakt. Not to hear how they sound, that is possible also with mp3 demo songs. But to show how they feel when they are played. Playing all notes is possible, with sustain, with releases, with hammers, with pedal sounds. But to limit the size and that playing them doesn´t make too much fun, I faded most of the samples out after one second. That sounds brutal, but it indeed gives you a very good impression of how the pianos can be played. You can download them from my page www.pianowave.de and play them using Kontakt-2.
Enjoy, and tell me what you think!
Stefan

[Jake Johnson]

Oops. Egg on face. I should have seen that the graphic was from the new K-Sounds piano.

[ElDuderino]

Thanks. Demos on www.scarbee.com are played quite softly with pedal down.

Transformation (dry)
Transformation (wet)
Flying (dry)
Flying (wet)
Baroque_inspired_melody (dry)
Baroque_inspired_melody (wet)
Tomheten (dry)
Tomheten (wet)

Songs by Nils Liberg

Hello Scarbee,

the piano sounds great. Which virtual piano was used for the above pieces (which are lovely, btw.) ? The Alicia Keys" piano ? I couldn't find any reference to it on your website.

[JohnGrant]

Playable and totally real sounding? No, doesn't exist... Yet....

First: loudspeaker technology prevents even the best sounding piano sample (or piano recording for that matter) from sounding like a real, live piano. That's obvious.

Second: since the only issue therefore is whether or not a credible recorded piano sound exists, the answer is 1) the software is there, but 2) the hardware isn't: we have yet to conquer the dreaded delay between hitting the key and hearing the sound.

So I've settled for playing in material on a reasonable facsimile of a piano keyboard (to get the right feel, at least), and then listening to the midi file via various sampled pianos.

Only using this kind of two-stage process can you get something close to what I'd call a "real piano recording sound."

A "real piano" sound and feel (playability)?

NO.

Doesn't exist.

JG

[ohernie]

I'm guessing on a grand piano the hammers are about a foot in from the keys. That's the point of sound generation. If you sit a little back there's probably another two feet. My guess is the distance between the point of impact and your ear will vary from 3 to 4 feet. Sorry, I don't have a grand around to confirm this.

Because of the speed of sound that's roughly 3 to 4 milliseconds of latency. Assuming of course, we ignore any actual hammer travel time.

Based on what you have said, wouldn't that ruin the playability?

And does it follow that playability would be maximized if one performed with their head under the lid?

(Sorry, two gigs today, time for some fun)



Ernie

[JohnGrant]

Red herring. Natural latency (whether a result of the instrument itself or the environment in which it is played) is one thing. Unintended hardware-related latency is quite another.

I'm open, however. Somebody tell me what set up will get the biggest piano samples to play and record simultaneously with 5 millisecond latency .....

Is that possible now?

JG

[ohernie]

No, it's not a red herring. Latency is the delay between action and sound. The latency of trying to play through an amplifier 20' (20ms) away is no different from having having a controller with 1 ms of midi latency, 16 ms of computer latency and a speaker that is 3' away from you.

I don't do any real orchestral work but on the (very) humble 2700+ Athlon XP system I use for gigging (solo instrument) I can get below 3 ms.


Ernie

[P.T.]

IMHO, when people decide to start seeing all these virtual instruments and sample libraries as tools in their own right and dealing with them as such, and stop hoping for/expecting them to be identical to real instruments, they will make better music.

I agree with this.
I keep saying it on various forums, but people are locked into their idea.

A sample is a sample.
It's not a guitar or a sax or a piano etc.
It is a sample of those instruments and that sample has it's own characteristics. Playing the instrument at hand (in this case, a sample) will yield better results than trying to make it behave the same as the real instrument, or do the same things as the real instrument is a prescription for trouble.

Though, I suppose this is off topic and should be a different thread.

I also am bothered by the idea that you can't have more than 4 hits when using a sampled drum kit. Or that you must use guitar voicings when using a guitar sample. I use a guitar sample when I want a guitar sound and I don't care about the voicing limitations of a particular instrument, unless I want those voicings, in which case I will use them.
I'm using a sampled kit on a computer. I have no interest in trying to fool people into thinking that it is anything else.

I am working with the tools at hand and have no interest in being tied down to the past by rigid attitudes of how things are supposed to be.