Yamaha, and others use their own hocus pocus, AWM, SCM, SN, AiFLM and so on. Do you, as a musician really care? I just want to known what it sounds like. If they want to keep their technology secret let them, and let another competitor produce an 'open source' type product. Oh wait, aren't there some software pianos already like that? As an engineer it's interesting but that's not why I learn piano and music. JMHO.

In summary UPHI seems to be better than PHI but no way we can quantitatively know how much. We are left only for qualitatively (hearing them) judging the differences between the two.

Maybe (just speculating) the companies do not want to release the exact information of how big are the samples they are using because that would show a lot of distance from sampled software pianos, which are, quantitatively (judging only the numbers), far superior. That would be specially true in the low end DP models.

But qualitatively the distance is not so big. In fact internal sounds are very well adjusted to the action of the DP, and latency is minimal (by the way, somebody knows the latency of a DP... it has to be very small, but how many ms?), and that makes them very playable and enjoyable compared to software pianos of an equivaleng size. DPs also tend to use any trick available to get the most from a small sample set (blending layers, stretching, simulating string resonance witch effects, etc).

So, hiding the numbers would benefit DP manufacturers. Also, a small increment (going from x layers to x+n layers) is not as apealling as going from PHI to Ultra PHI.

Or maybe is just a marketing trend...

Reverse engineering is the magic word. It has been done a thousand times, but it doesn't seem to be worth it for an enthusiast in the DP market. For those little iPod whatever gadgets there is alternative firmware, hacked apps etc. etc., but we are alone in the rain... *sniff* Guess what would be possible if someone would write an alternative firmware for the CA63... Longer sustain, more space for patches, even different algorithms for sound generation...

And I think if a company really wants to know what's inside another brand DP they can just rip it off and do their own reengineering.

I think thats the keyword.
They did do some magic and experiments with UHPI but dont understand themselves exactly why it sounds better. So they cannot explain ;-)

You must be right. It's magic - and you just have to believe it. We're still living in the dark ages it seems.

You're right - at least when it comes to DP firmware! Thank god we have our Galileo -> Dewster!

As luck would have it, the three Kawai DPs that we have DPBSD MP3s for also represent the three different levels of Kawai's Harmonic Imaging technology. Here is a quick comparison of the obvious technical features:

MP5 (HI - Harmonic Imaging)
- Stretched, 22 groups.
- Blended velocity layers
- Attack: 3.0,1.8,1.6,1.5,1.6,1.0,?,? => avg 1.6
- Loop: 1.6,1.2,1.1,1.3,1.0,0.8,?,? => avg 1.1

CN33 (PHI - Progressive Harmonic Imaging)
- Unstretched.
- Blended velocity layers
- Attack: 1.8,2.1,2.1,1.8,1.7,1.6,1.1,? => avg 1.7
- Loop: 0.8,0.8,0.7,0.8,0.7,0.7,0.6,? => avg 0.7

CA63 (UPHI - Ultra Progressive Harmonic Imaging)
- Unstretched.
- Blended velocity layers
- Attack: 3.0,3.0,2.1,1.6,1.4,1.5,0.97,? => avg 1.8
- Loop: 1.4,1.3,0.68,0.83,0.64,0.74,0.63,? => avg 0.8

So the big step from HI to PHI is the elimination of stretching (every note sampled).

It's not clear to me what the step from PHI to UPHI is, though it could easily be additional layers, more note damp samples, etc.

What's really fascinating is the fact that very little has changed in terms of average attack and loop sampling times. In the MP5 and the CA63 the attack sample times are weighted more towards the bass region, which is the right thing to do if you employ looping IMO.

And the lowly MP5 actually has longer loop samples than the two later models, a function I believe of "improved" loop processing techniques. But my stance on looping is this: if you must loop, the loop sample should be made long enough to realistically wobble without easily hearing the wobble period (i.e. sound like multiple slightly detuned strings) - which requires on the order of 5 to 10 seconds of sample time depending on the note. And ideally the looping would happen only after a lengthy attack sample, at the point where the note volume has dropped significantly.

Good point... additional sample memory can be used for string resonances and other things as well, besides the three I listed. As others have mentioned... note-off samples, longer loopable regions... probably other things too.

That's a good observation. I remember when I listened to an MP5 sample online, the thing I was critical of was how simple the decay sound was, as if it were a piano with only one string per key, without the richness of multiple strings which can never be so perfectly in tune. I think that is a lot of what makes DPs generally sound unnatural in the decay.

Absolutely. A second or two of looping will either give you a quick fake wobble, or if it is highly processed (like the recent Yamaha CP sounds) an equally fake non-wobble. A bunch of notes decaying & audibly looping together is something I really just can't abide anymore.

voxpops,

I'm afraid it is not possible to control the drawbars of the MP6's tonewheel organ simulator via MIDI.

Kind regards,
James
x

batak,

The MP6 does not transmit dump information via MIDI.

Instead, data is loaded from and saved to USB memory devices.

Note that the MP6 can receive the MP5's AllDump information (via MIDI), however there is no guarantee that the SETUPs information will map perfectly to the newer board due to its new/additional sounds.

Kind regards,
James
x

That's a shame - I wonder if it could be implemented at a later date...

Thanks for asking, James. Right now I'm talking to my local Kawai dealer about the MP6. I'm feeling it still ticks enough boxes to be a major contender.

dewster,



Not quite, there is actually a fourth standard - 'Harmonic Imaging, 88-key sampling', as utilised by instruments such as the ES6, CN22, CL35, and CA18.

The MP5 was developed prior to the advent of Kawai's 88-key piano sampling technology, which is why the stretch groups are rather noticeable. If you were to examine a DPBSD of one of the instruments listed above, you would find that the sound is unstretched - just like PHI and UPHI.

For an 'at a glance' comparison of Kawai's digital piano instruments, please refer to this PDF - note that there is a brief explanation of the various DP features on the inside back page.

Kind regards,
James
x

Harmonic Imaging is the technology used to recreate the rich sound of a grand piano, and ensures that all KAWAI digital instruments off er a smooth tonal transition from pianissimo to fortissimo. A hand-built EX Concert Grand is at the heart of every KAWAI digital piano, with 88-key piano sampling capturing the unique harmonic character of each individual note, and enhanced forms of Progressive and now Ultra Progressive Harmonic Imaging delivering the most expressive grand piano sound ever realised.

Thanks! So "HI" in all of its incarnations just means blended velocity layers, and doesn't have anything to do with stretching.

But after reading the back of the product catalog I can't say I have any more insight into the differences between HI, PHI and UPHI. Not trying to pin you to the mat or anything, but may I be so bold to suggest that it could be good marketing to let prospective customers know what they will be gaining by selecting and paying extra for a product with more advanced sound technology features?

To their credit, Roland has taken the bull by the horns by directly addressing layer switching, stretching, and looping with their SN piano media blitz - so the cat is already pretty much out of the bag, no? Manufacturers don't have to tell us exactly what they are doing under the hood, but at this point they should sufficiently convince us that they are aware of and working the problems IMO.

Do I understand from the DPBSD that the longest attack sample for PHI is only 2.1 sec long and the longest looping sample only 0,8 (not even a second) . That really worries me. When I play a note on an acoustic, the sound lasts (and keeps on changing!) forever in comparison. I wonder how the tone after the attack phase could sound like a piano at all with this little sample time, but perhaps I miss the point somewhere ?

FP7/ RD-NX still in the picture then...

RE: the Roland SN, is it necessarily much better, in terms of the quality of the note evolution as it decays? I know SN doesn't have any looping, but is it capable of producing a rich, slow, phasey sound as it decays? Or is it still rather static, aside from the fact that it does become mellower in a natural way?

Greg.

No, I don't think you've missed the point. This is exactly what brought Roland Supernatural into being...this and audible velocity layers. However, it is certainly true that whilst looping and layers are intrinsically undesirable, the actual implementation of them varies considerably...some are much better than others and if well done can be acceptable.

If looping really is an issue (I certainly don't like it) then Roland SN is the only current answer. I suspect however that the Kawai MP10 has a superior action...well, if not technically superior, perhaps a more natural feel...and I also suspect the Kawai is of higher perceived quality, ie, less mass-produced looking/feeling.

The state of the art in hardware DPs is undoubtedly Roland SN though...it should only be ruled out by people that don't relate to the Roland sound signature (slightly richer, warmer, darker than Yamaha and probably Kawai too). The MP10 would be on my list though...I wouldn't buy a top end stage piano without trying it anyway.

And if I were to re-visit my piano buying experience personally I would be very tempted by the FP-7F...On paper specs/look I would give it the nod over the RD...just can't stand the placement of the RD's pitch-bend/mod controller...this is subject to the key action being strictly comparable between the two.

Cheers,

Steve

JFP, that's theory/just an indication and BTW true for almost all the DPs brands. You need to check the DP for yourself whether you are satisfied with the sound, and whether the looping is audible for you.

Again: For me the keyboard action is the most important part of a DP.

Greg, to my ears the SN decay is absolutely believable...there is nothing static or artificial about it to my ears...it is a towering achievement in my opinion, the biggest asset of the SN engine in fact.

Cheers,

Steve

From a highly pessimistic technical standpoint the SN sets might just be another case of clever use of filter/blended layers that is tailored to fool a spectrum analyzer. Listerners might never notice the difference anyhow. To brighten the tone it's 2010, and I don't believe they need to do this looping/stretch charade anymore. As dewster points out the cat is out of the bag and companies are fighting it

But as always. If it sounds good it sounds good. For example I kind of like the way the CP5 add an artificial flavour to the note decay. I didn't hear this so accurately before seeing it happen in dewsters test images. But now that I know what to listen for I can't un-hear it!

Ok, ok, I agree - the SN is indeed impressive. I had never listened to the DPBSD closely until now. The decays are very good.

Greg.

A big telltale for me is watching the VU meters in Audition while listening to the decay portion of the notes. A real piano will show L & R bobbling around rather independently, whereas a looped piano will display the same bobble pattern cycling over and over. Roland SN looks and sounds very real to me, which is quite refreshing after listening to so many DPs with criminally short samples.


Very well put. By constantly singing the virtues of SN I feel I'm in danger of giving people the impression that I'm generally pro-Roland or anti-other DP manufacturers, but it is really a game changer in terms of sound technology. It even gets all the small things right, like the silent replay test, which most DPs and PC samplers can't boast. There's literally nothing else out there like it, and the other manufacturers by comparison have been caught with their pants down.

Some tidbits of news for anyone still interessted in the main topic. A local dealer responded with this a few minutes ago as to my question how soon I would be able to pick it up at the store after the official release date.

"As a "main dealer" of Kawai i Norway, we can confirm that the new MP-10 will be in stock in Norway, around 1 of December..."

That should count for the rest of scandinavia as well. He went on to say it could be shipped shortly after that.

It should worry you.


No, you've got the point exactly right.


We will probably buy one of those for our studio to replace the P120 we recently sold. Just waiting for more feedback from users regarding the key action, and a physical trial in a store, to decide which one. I like the extra sounds, build quality, and UI in the NX. I like the music rest, built-in speakers, and compact length of the 7F.

Too bad they don't make an RD7F. Though, since the perversity of the universe tends towards a maximum, they will probably unveil something like that the second we buy the NX or 7F.

and by people who are not willing to carry around a piano that weighs over 50 pounds (not even counting the case). I wish they would come out with a piano with that sound quality in a box the size/weight of at least an FP-4, if not a Yamaha P-95, Casio PX-3, etc. Or a rack module (like the old MKS-20 and P-330) or sound module (like the SonicCell), and then we could get the sound on whatever size keyboard we were willing to haul around.

Even if I would love to have SN build into my Kawai I would rather buy the DP with the best keyboard action, full stop. That's by far the most important decision to make. If, in case of Roland, the sound signature of SN and your personal preference of best keyboard action meets in one of it's DPs, I can only congratulate you for an almost perfect combination!

When you decide to upgrade to an acoustic piano will you buy one with the "best action" no matter how it sounds?

Perhaps you could have a Steinway & Sons. klavier installed into an old Yamaha grey market frame to save money?

It seems to me that any reasonable person buys the total package. What is the point of playing on the theoretically best keyboard if you aren't happy with the sound of the instrument? It sounds a bit perverse to me.

Yes. (If it has a proper MIDI output.)



I don't think it would fulfill the MIDI requirement



Well, if you have MIDI, you can completely replace the sound later, with anything you want. Pianoteq and Ivory comes to mind, but a SN sound module would be nice, too...

The cost (and hassle) of replacing the sound is really small, compared to the cost of the original DP, so this seems to be a sensible option to me.

Of course one buys the whole packages, but (for me, at least) the whole package contains the expanded possibilities, too.

Agree with Csillag on this. All depends on priorities. If you're primarily using it as a MIDI controller, then the built-in sound is fairly inconsequential. I use a DP as a MIDI controller first. Occasionally I use the built-in sound when I don't want to wait for the computer to fire up and playing the acoustic would wake everyone up.