Not bad...sounds a little chorusy in some places, and the decay is rather short, especially noticeable in the first piece.

Good bottom end, and very nice and sparkly on top....a little thin in the middle, but still nice.

I'd say it is a Korg piano...not stock...probably modified on a PA series.

Ian

I can't tell what keyboard it is, but I can hear the chorus and decay problem.
It's a good sound though, just needs tweaking.


Have a listen to this one. http://www.irishacts.com/mp3/grandpiano-2.mp3

I used the same Purgatory Creek midi file as you did. The Piano is my own work, I sampled it myself, and it's playing back on an OASYS in this demo.

That Purgatory Creek midi file burns my ears though no matter what I hear played back from it.

For example, here's the same sound played live by me. http://www.irishacts.com/mp3/grandpiano-1.mp3

And here's the same sound in a full sequence. http://www.irishacts.com/mp3/grandpiano-3.mp3

The last two demo's in my opinion give you a better idea how it sounds than the Purgatory Creek midi file.

lol... I'll have to start my own Piano Shoot out website.

Regards.
James.

Didn't care for it. Very brittle IMO. I'm referring to the very first post.

[This message has been edited by zuki (edited 03-12-2009).]

THis is k-sounds piano remapped on a pa2x using the parameters of the grand piano rx preset. Probably can be tweaked better....

Which K-Sounds piano, mrdave?

If it is Vol.2, maybe it needs more work? Here are a few demos of the Kurzweil version of Vol. 2.
http://www.ksounds.com/Demo/KSV2_Kurz/KSV2_Piano_1_studio.mp3 http://www.ksounds.com/Demo/KSV2_Kurz/KSV2_Piano_2_bright.mp3 http://www.ksounds.com/Demo/KSV2_Kurz/KSV2_Piano_1_full.mp3

This is one of the few sampler pianos that fit in a sub-128MB sampler that I think seriously gives my G70 a run for it's money. And Kawai has it's own 'flava'... sort of a hybrid between Steinway and Yamaha.

James.... what about the P Creek file burns you? On the better sampled pianos, it sounds amazing... On lesser ones, perhaps not so much.

I have a feeling that quite a few piano sample makers cross over the bright samples too low down in the velocity range, in an effort to get that mix cutting sound. Trouble with that is, a real piano cuts in the mix because of careful EQ and compression, but the raw piano sound is still full and warm, and only gets that bright when you really spank it.

The P Creek file really shows this problem up, as it truly DOES use the full dynamic range of MIDI.

But please tell me what YOU think about this. I am fascinated to talk to someone that 'walks the walk' and tries what is surely the most challenging sampling project out there...

The midi files has no real emotion to it. It's just a barrage of notes in an effort to get through the song and to the end. The type of classical tune it is doesn't help matters either.



Maybe for on-board factory keyboard sounds as memory is limited, but not professional sample libraries. A pro library will usually have 3 or more layers.

My Piano above has 5 with each stereo sample only stretched 1 semi tone in either direction from the root note.



I've spent god only knows how many hours listening to this tune and I don't agree. I find it a heavy handed performance, and quite honestly I hate the tune at this point


Maybe that's my problem, I've heard it too many times

Regards.
James

Could be. It never fails to demonstrate the flaws a sample set has, though... you listen to it on the better sets (Ivory, TruePianos Diamond come to mind) and it sounds convincing. Others (in fact I am appalled by how many!), not so good...

And, I'm sorry, but for all their vaunted advantages with having unlooped, no pitch-stretched, high count velocity switched, multiple mic position sets, it amazing to me how few stack up to to or three of my favorite ROM pianos... (can't wait to hear the V-Piano play this file, too!)

Listen to the Nord Stage Yamaha, Kurzweil PC3X New Triple Strike, and of course, my favorite, the FantomX Ultimate Grand Piano, and you realize how careful sampling and obsessive attention to getting the dynamics just right can make quite small pianos (RAM-wise, that is ) still very effective.

What did you think of those K-Sounds Kurzweil demos, BTW? Did you see how it could be warm AND bright, depending on the program (EQ and dynamics, etc.)?

Anyway, I still don't think you can call the file heavy handed. It hits 127 only ONCE in the entire piece, seldom exceeds 116, and goes to a low of 28. Just eyeballing it, it looks like the average sits about 90-100.

I simply think that it shows how low down some of the pianos put their higher velocity (real force) samples. It doesn't matter if it is three, or ten layers, it is whether there is sufficient at the top end of the range to NOT make you sound like you are simply pounding the piano when you go much over 100. There should still be a fair amount of nuance in that upper range, IMO. What velocity points do your samples cross over at?

One of these days, someone is going to use a weight to play the samples, dropped from different heights, so there is force consistency across the action, then use weights on a target MIDI piano so that they match up... Then finally we might have some really accurate dynamics

Anyway, keep up the good work... it's a pleasure hearing about this...

My problem with the midifile is that the person who recorded it in the first place would have been greatly influenced by the sound of the Piano they recorded it on. It's hard to disconnect yourself and play that well in a sterile way that will suite one song to be a universal demo.

That's only one part of the problem. The second part is depending on what Piano you sample, it's dynamic response curve will be very different. This must be taken into consideration when performing as how the piano responds will effect the level of emotion you put into your playing.

In short, you will play the same song completely different. This is why I don't believe a single midi file can do any piano sound justice.

Yes, it can give you a very good idea of what it sounds like, but to really hear the sound a demo needs to be recorded playing the sound. No generic midi files.

Pianoteq of all things is a great tool for proving this. You can load up a Yamaha C3 Grand and play a song with all your heart while it records the midi file. Now switch between two Pianos in Pianoteq and for one the song sounds fantastic, and the other just sounds wrong. The motion is lost because the dynamic curve response is wrong for the sound and you don't get the most from the sound board, hammers and everything else that makes up the sound. Yet re-record the song played on the second sound, and you will actually play it differently.



Yep... it does happen. Sampling a piano is also nothing short of your worst nightmare and it took me a few attempts to get anywhere at all.

That said, on the flip side you would be surprised how listening to the factory Piano sounds also colours peoples opinion of what a Piano actually sounds like. I know quite a few people who believe the KORG M1 Piano is the best they ever heard. ..lol



Yep, that's the joys of unlimited time at the Piano in a Studio. A Piano can make or break the sale of a keyboard, so a lot of effort goes into creating the sample data for the factory sound.



I know Ksounds, he's a member on KORG Forums and we first spoke when he wanted to develop those sounds for the Triton Series. They are very good. The EQ / Dynamics would be the Engine, the filter modulation and MFX/IFX. His sample data is very likely to be 100% natural and untouched by EQ.



Yeah heavy handed is really the wrong word to describe this. It would have been better to say that the performance is more focused around a single layer of samples.

For example if the average is 90 to 100, and the majority is played on a single layer. Your 0 to 127 needs to be divided by the number of layers to work out the range each layer covers, and at 90 to 100 you can be sure a lot of the sound is coming from a single layer.

That at what I've said above about dynamic velocity curves are in my opinion what make the midi file nowhere near as effective as a proper performance played live by a human on the sounds being demoed.



The cross over points between layers vary depending on the keybed of the workstation being used to play back the sounds, but for something like my 5 layer piano.... your looking at something like this at the sampling stage.

Layer 1 – 0 to 25
Layer 2 – 26 to 49
Layer 3 – 50 to 75
Layer 4 – 76 to 101
Layer 5 – 102 to 127



They have been for years, it's called a Keyboard Thumper.



Thank you, I'm enjoying the conversation.
Most people don't question me over on KORG Forums and I end up just being a Moderator answering peoples tech questions. It's nice to be asked real questions and talk about something interesting for a change.

Kind Regards.
James

Thank you... I enjoy reading your well articulated posts, too.

I love the piano, and after having tried since the original Ensoniq Mirage ( ) to get something onstage that sounded as good as the studio, I have come to realize that anyone that can come even CLOSE is a very talented programmer and recordist...

I have to admit, perhaps I'm not using an 88 wooden controller enough (though in the studio, I use a K2500X for the piano stuff) but having TWO (and a half) samples completely under 64 seems a bit of a waste, as little as you are down there. Filters and dynamic EQ can do wonders in that pianissimo to piano ranger with just the one sample. I honestly would prefer two samples in the 100-127 range.

I think that P.Creek ought to allow a simple velocity shift to allow some of the wrong curves to shine. Perhaps one file for untouched, and another allowed that the user can adjust global dynamics to fit the samples.... But some kind of force standard really needs to be adopted to let pianists NOT have to adjust every single time. Yes, your playing makes a huge difference, but overall, I am happier that there IS a P.Creek standard, rather than the way it used to be done, with no standardization at all...

At least it gives you a good apples to apples comparison of timber and phase coherency (the bugaboo of sampled pianos, IMO) and image stability (I can't believe how many sets 'swim around' in the stereo field).

To be honest, I wish there were something similar for drums and horns, etc... Standardization of touch and expression isn't necessarily a BAD thing! There's be a LOT less work translating styles, for one thing! I am getting very annoyed at how much GM/GS/XG is diverging from those standards. Interchangeability of sounds in styles and live playing is one of the things that made the eighties and nineties arrangers so effective. No surprises, it all just worked (within the limits of single layer sampled sounds).

Today, maybe better sounds, but less compatibility. It might not be a loss, per se, but it's a lot more work for the poor player...

mrdave hasn't responded yet about the K-Sounds piano he posted. I'm a little worried about the chorus-y nature sometimes in it. Perhaps he needs to double-check that the Korg is playing the stereo samples in phase, or maybe there are a few overlaps in the velocity cross-switching?

But if it IS the Vol. 2 piano, at least the Kurzweil programming shows that it CAN be done. It isn't the samples themselves. I used to convert a lot of Akai sets (still quite like some of the EastWest stuff) for my Kurzweil, and am only too aware how difficult it is getting the envelopes just right. It's rare you can use the envelopes for a different piano on another one. They are all just too different...