Watto's upright piano sampled

[Watto]

Not sure if I explained my thoughts on that very clearly, so I'll try again. If you were sampling a piano using some sort of robotic technique, you could probably achieve much higher velocities than the average player is capable of (although, having tried it, you'd be surprised how a piano will only let you go so far). The thing is, I don't see the point in using 1 out of 3 velocity layers to produce a sound you couldn't physically achieve in real life, or would be very hard pressed to reproduce.

I hope I don't seem stubborn or something, but maybe that's what will make this piano a tad different for people out there (there are hundreds of piano sample libraries), plus the fact that it's miced from the back. As I said, you make some really good points, but it's not really the vibe I'm going for here, which means it's probably not up your alley, but that's cool

[Lotuzia]

Sounds nice and clear to me.

Allways a good thing to have a different piano imho.

Good luck

LtZ

[Watto]

Cheers, Lotuzia

Here's the same demo, with the velocities raised slightly to get a couple more hard strikes:

http://simonwatson.com.au/kvr/piano_demo_vel.mp3

Thoughts?

[Jake Jackson]

Sorry if I sounded dogmatic. You should of course do what you want--you know the sound you want. And as I think about what I wrote, I have to admit that I'm not sure what effect having the wall close will have with really hard strikes. You'll be getting a lot of reverberations.

Just heard the new demo. I like the upper velocity level better, now. The lower levels sound a little bright. Have you set filters and envelopes yet?

Please don't take this too seriously but: (Here I go again)

No, no. I wasn't urging you to create a really bright, hard sounding piano. Remember I'm only talking about the very hardest strikes. I'm not saying that you should only record hard strikes. You may never hit the keyboard hard enough to hear the loudest strike at its highest volume. Yamaha, etc sample pianos at very high velocities for several reasons, but basically the idea is to get in the most harmonics\partials. The harder the strike, the more partials you'll have. So when you apply a filter or velocity curve, you are only adjusting from the greatest possible number of harmonics. If you instead start with softer hard samples, you have fewer harmonics to start with, and those will be at a lower volume than if you record at the highest level. And then you reduce those still more when midi tracks amplitude to velocity, and still more if you use a filter.That's partly why the pianos on Yamaha's synths sound good (when they sound good) They use really hard strikes and filter them slightly to create low velocity sounds, but you can still hear the upper harmonics because they were so loud in the first place. Otherwise the filter would cut them entirely.(The exception here is for close micing near the strings, when the hammer thump can get too audible. Near micing and signal\noise ratio for soft strikes get to be big problems.)I understand the problem with the wide scale between soft & hard strikes in a 3 layer piano. Thats always the big problem and where both recording and programming take planning and experimentation. But...really hard hard strikes also give you the maximum loudness so you can scale it down with the volume knob and the amp envelope. If you record at a comparatively low level, you can only reduce the volume further. When you then adjust the volume scale of the lower velocity levels to match the upper samples, the result can be that the piano multisample will never sound as loud as your piano. It's likely to sound slightly distant, no matter how much you turn up the volume.

But I'm lecturing and I worry about that wall. Sorry. Record it so that you like to play it. Looking forward to the sound you get.

[multree]

it sounds so totally Watto....

two octaves is fine.... I wouldn't need more than that anyway

although I usually like the lower octaves... so if you wanna add an octave - take the lower one...

and yeah... do share this...

at least with me

hehe.. cheers,
mully

[Watto]

Hey Jake, did you have a listen to the second one? It's only subtly different, so probably still not your cup of tea

Mully:

[tee boy]

Sounds great! Very good.

I wouldnt mind doing this myself actually - sampling my piano and using it my projects. I was always a bit stuck on how to get the velocities enough though across the entire keyboard. Sadly, my dynamic control is not quite up to the rigors of such a challenge!

Do developers of sample libraries usually just wing it (with a decent pro pianist in the driving seat)?

Or is there a technique that helps to keep the velocity layers consistent?

I assume the highest has to be as loud as possible, the lowest as quiet as possible. My problem is judging the increments between levels and keeping them consistent across all the keyboard.

Appreciate any advice.

TB

[Jake Jackson]

I did like the 2nd recording for the hard strikes. Did you apply filters and envelopes yet, if you plan to?

[Bonteburg]

Sounds great! I assume the highest has to be as loud as possible, the lowest as quiet as possible. My problem is judging the increments between levels and keeping them consistent across all the keyboard.

Appreciate any advice.

TB

maybe a robot of sorts ??

[Watto]

I did like the 2nd recording for the hard strikes. Did you apply filters and envelopes yet, if you plan to?

Nup, that was exactly the same as the first recording, only with the velocities turned up a smidgen - bear in mind that I'm playing one of these:



Terrible for playing piano

[tee boy]

Sounds great! I assume the highest has to be as loud as possible, the lowest as quiet as possible. My problem is judging the increments between levels and keeping them consistent across all the keyboard.

Appreciate any advice.

TB

maybe a robot of sorts ??

Lol, like some virtuoso Terminator!

Does my playing sound mechanical?




Steady on, it wasnt that funny...






Seriously though, I am actually wondering if there is some mechanical method of standardizing these velocity layers. Im sure I remember when one company started talking about their 'special' method (obviously keeping details close to their chests). But it got me thinking - how exactly would you do that??? Particularly given that the action of the piano varies with pitch.

TB

[koolkeys]

Yes, there is. Many companies, like Sampletekk, have built devices that allow them to play at an extreme amount of velocities with a consistency that a human just can't play. That's why their libraries have up do a few dozen layers. They can line it up to play every note at the exact same velocity.

That being said, I like the sound that Watto has here and I agree that the highest velocities may not be needed. I consider myself a VERY strong pianist, and I like to play loud. However, whether on a keyboard or a real piano, during normal playing, I just can't hit those velocities.

I have played a few pieces over the years which were played rather fast, with some alternating octaves(which to me is when you're hitting the strongest) between the hands, playing as hard and loud as possible, and I have libraries that have layers that exist higher than I can reach in those times.

I know what Jake is trying to get at, and he seems to have good method behind the madness. But I really like where this piano is going and the way it sounds. Plus, my controller has 12 velocity curve variations, so I can adjust as needed anyways!

Brent

[Gregjazz]

Yeah, you get a disklavier (you can also get the hammers on the piano rigged to operate as a disklavier) and then you can feed the piano specific velocities on notes. That's how you get 16 evenly-sounding velocity layers. I don't know anyone who can do multiple velocities by hand with the precision it takes to make a solid piano library.

[Watto]

You would have laughed if you'd seen me doing it, because I was looking over my should at REAPER to see if the note I'd just played was a similar size waveform to the others

[Watto]

Oh, and to precision - I hate being precise about anything, which is why I dislike most sample libraries.