My wife teaches private piano at home and has a Yamaha P-120, mainly to encourage the little ones to play the same thing over and over in different voices, and to practice her church service stuff which is often organ based (on an incredibly lame church keyboard unfortunately, a toy really). She plays her Young Chang 6.1' Grand otherwise, which has a nice bright sound and a light action.

I (an EE with a very strong interest in electronic music - but I play guitar, not keys) usually scour the web for info regarding recommended keyboards for her students. I put together a one-page deal over the last couple of days based on our trying things at Guitar Center and looking at PDF manuals on-line. The only keyboard in the list (complete list: Casio CDP-100, Yamaha P-85, and Yamaha P-155) that we couldn't see & try was the Yamaha P-155.

To me, what makes a piano truly magical is the sound of single, sustained notes with the pedal down. Looping (totally stone-age compression technology) kills this, as does the lack of sympathetic resonance modeling. Although I've never played with one, I think the GEM piano modules garnered rave reviews simply due to their sympathetic resonance feature.

Anyway, my questions are regarding the P-155 general sampling, sympathetic resonance sampling, and key-off sampling: Is there obvious looping (like nearly all digital pianos have - a total peeve of mine)? Is the sympathetic resonance obvious? Obnoxious? Real-sounding? Desirable? How about the key-off sampling? Could someone who owns one please comment on this? Also, could someone who owns one please post a sample at the Digital Piano Shootout at Purgatory Creek? Thanks!

The effects, resonance, key-off and so on are adjustable. You can turn them up or down. Look in the user manual where it talks about the "functions". I think the Yamaha has a clean sound not muddied up with over done resonance.

Thanks for the response ChrisA!

Is the resonance effect realistic in your opinion? Does it add significantly to the piano experience? Also, if you adjust it, does the adjustment "stick" through a power cycle? The P-120 forgets everything with a power cycle...

I'm really interested in looping artifacts too - do you hear it on single sustained notes? I can clearly here it on our P-120, and every other digital piano I've messed with (lots).

OK, I looked more closely at both the P-120 and P-155 manuals. It looks like you can set what things are remembered through a power cycle per category in both keyboards. The P-120 forgets it after a week or so, the P-155 appears to remember it until you change it or do a factory reset.

I really don't like some of the voices being relegated to that hokey "strings/others" key. On the P-120 each voice has it's own key. This is a clear step backward IMO. In fact, it's almost a deal killer in terms of how my wife and students use a keyboard.

I'm still rabidly interested in hearing what owners of the P-155 think of the sympathetic resonance effect, and if looping is audible. From the press releases this keyboard has 4 samples per note, but does that mean they sampled every key, or is there the possibility they are still stretching samples over multiple keys? From the P-155 brochure:

"Using Yamaha's Dynamic Stereo Sampling (DSS) technology,
the incredibly natural 4-level piano sample is a breakthrough
achievement for digital pianos. It means that, depending on
how hard you strike a note, you will hear one of four piano
recordings on each and every key for an unprecedented level
of dynamic range and tonal expression"

Blah, blah, blah (translation: we're only 10 years behind PC-based samplers.) The key phrase "on each and every key" leads me to believe they are sampling every key, but who knows? I wish manufacturers would come out and say what their product do in unambiguous terms. Should I even care what "pure CF sampling" means?

A P-155 sample uploaded to Purgatory Creek would be vastly appreciated! I'm on my hands and knees here people...

I just unpacked a P155. It will be January before I'm set up to record it.

BTW mine, unpacked just now, is a low serial number indicating that there are well under 2,000 thousand "in the wild" They are still new. They are made in Japan. I was surprised

I can't hear a loop on long sustains. But that does not mean it's not there. Maybe I just don't know what to listen for. But to my ears, no loop.

From your comments it sounds to me like what you really want is a Roland FP7 or RD700. These are hybrid sampler/modelers and have access to a dozen or so piano voices (plus hundreds of others) and have better key action. The cost is about double or triple. But you do get a nicer LCD screen and direct access to many different sounds via Roland's "scroll wheel" you get a steel case, no plastic or particle board. But did I say the price is double or triple?

I got the P155 because I'm mostly going to be playing the "piano 1" sound in headphones. It is the best DP in it's price point for that.

dewster, I don't believe Yamaha sample all 88 keys separately.

However, this is not necessarily a negative point, as it could potentially allow more memory to be allocated to the keys that are sampled, thus longer samples (and less noticeable looping).

Cheers,
James
x

ChrisA: Turn off the reverb, and perhaps the resonance too, then play and hold an individual note for 10 seconds or so. You might have to try different notes until you find one where the looping is really obvious. The sound will go from a rich, realistic, three strings slowly beating thing (the sample) to something more static and fake sounding (the loop). It may be more obvious in headphones. I hate looping!

Interesting about the low serial number.

Regarding the FP7: We briefly played one at Guitar Center but didn't like the feel of the keys, they bottomed out hard. Also didn't like the look or feel of the buttons. From the FP7 manual Roland says "Every note of an 88-key grand piano has been sampled..." and at Purgatory Creek it lists the two piano sounds as "Each Note Sampled: Yes" but there seem to be no resonance samples, which is what I'm really interested in. I also listened to the samples at Purgatory Creek and Grand Piano 1 struck me as overly metallic; GP2 was better, but somewhat phasey in my AKG K271 headphones, the notes seemed to be coming from multiple random directions.

KAWAI James: It appears you work in the industry (I'm jealous!) so my question to you is this: I can buy a 4 GB flash drive for ~$10, so why are keyboard manufacturers still treating sample storage memory like gold? Why does every keyboard still use stone-age compression i.e. looping? I really just do not understand why keyboards, and digital pianos in particular, are so far behind the technology curve. Are we all just sitting around waiting for the last old-school designer to retire / die? I'm completely baffled...

UPDATE:

Just got the following vague reply from Yamaha tech support regarding samples on the P-155:


The notes are individually sampled.
Thank you,
R.R.
Pro Audio & Combo Support
Yamaha Corp. of America
(714)522-9950

Application/Technical Support
Is every note sampled on the P-155 digital piano? Or are samples stretched over multiple keys? Why isn't this basic and important information included in the sales brochure or user's manual?

I wanna know the same thing: if you can get a 500gig hard drive for $55 and 1gig of ram for $23 there should be no problems getting a 800MB piano sample (or 40) on there. heck! you could fill up the entire midi standard with 1gig samples and still have plenty of space for your 300gb of rhythm data

I doubt that they can use cheap flash RAM in a DP. The controller can never predict which note will be played so it would have to keep the entire sample set in fast dynamic RAM. Flash is to slow. The processor needs sub-millisecond random access to the sample set. (think about 128 note polyphony with latency under 1ms. If the samples are all 44.1K 16-bit they have a fairly healthy memory bandwidth requirement. This means they can't be using iPod style flash RAM.

I spend an entire afternoon in the DP section of a guitar center and could not help to notice the other customers. Every one of them, and there were lots seems concerned about a $20 difference in price between two models. None seemed to care much about the sounds. I was the only customer out of 100 that even thought to bring in his own headphones. Most would play the keys only for a few seconds then move on. I did meet only a couple other customers who were listening carefully.

So my opinion, at least as it applies to under $1K keyboards is that people care a lot about price and little about sound. So Yamaha, casio and the like give the customer what he wants.

But this does not explain the over $1K DPs.

At this time of year, they are probably buying DPs as presents for their children, not for themselves, so it makes sense they would be more concerned about price. And to non-players, the DPs from various brands probably don't sound all that different anyway.

I agree with Martin - and would add that this applies to any musical instrument being bought by relatively clueless parents. It's rather tragic, actually.

I think there are many on this forum and elsewhere that can appreciate better instruments, and are striving to find them through the open exchange of information (unlike the manufacturers).

Or perhaps the "clueless parents" are right in this case: Compared to a bad acoustic, the typical Casio or Yamaha entry-level digital sounds pretty great, and the differences in touch are not that huge. So they mainly buy based on price, what the sales guy says, and whether the DP color fits with the other living room furniture. Hopefully the second instrument will then be bought with their children having a say...

I've got 10 years of embedded logic design (FPGAs) under my belt, so I think I can comment on this :-).

There are much larger and somewhat faster parts out there, but let's take a relatively slow commodity part like the ST Microelectronics M29DW323DT Parallel NOR Flash which is a 4MB part (4Mb x8 or 2Mb x16) with an access time of 70ns. For a 44.1kHz sample rate, we need a sample every 1/(44100Hz) = 22676ns. 22676ns / 70ns = 324 samples per sample period. We could double this by doubling the width of the flash, thus retrieving two samples per memory access, which gives 648 samples per sample period.

648 polyphony!

Of course, this is just to a first order and there will be a need for interpolation between samples, amplitude adjustments (multiplications), filtering (multiply & accumulate), etc. But even relatively lame flash has a port bandwidth that is entirely capable of massive sample playback, and buffering the entire sample set in faster RAM is not necessary.

Having said all that, where the hell is my fully sampled, non-looped, 16 layer (8 pedal up, 8 pedal down), $1000 digital piano? There is no excuse (other than laziness) for this product not to exist at this point. The first manufacturer to build it will blow everyone else out of the water!

Innovation comes from research and research costs money. Manufacturers are MOSTLY in it for the profits. I know I shouldn't really bring in economics, and such into this thread, but hey, who doesn't like their money. And maximizing their profits is what humans do best (or some at least).

Side note, you ever tried the V-piano? If you did, what do you think of it?

dewster, I'm afraid I do not have the technical knowledge to provide a satisfactory answer to your question.

However, while the sound of most consumer digital pianos is certainly no match for dedicated software-based sample pianos such as Ivory or Garritan Steinway, I do believe that the technology employed has progressed considerably in recent years. Furthermore, with innovative new products such as the V-Piano and Avant Grand raising the bar, it is surely only a matter of time before your prayers are answered.

ChrisA makes an excellent point about consumer expectations. I do not believe the majority of individuals who purchase a digital piano are terribly interested in how many layers have been sampled, the resolution and bit depth of the voices, or the maximum polyphony of model X versus model Y. Most consumers simply expect an instrument to feel and sound like an acoustic piano - provided it fulfils these expectations I believe they will be quite satisfied.

Kind regards,
James
x

You can have all of that ... but not for $1000. Could it be made and sold for that price? Maybe. But put yourselves in the manufacturer's position. Why do it? People are willing to by $6000 (?) for the V Piano, or $20,000 for the Avant Grand. So why give it all away for $1000?

FYI..

I do not own one, but I'm pretty sure the P-155 does not have sympathetic resonance, which is different than damper/sustain resonance.

But really, in the grand scheme of things, at the price point of ~ $1000 USD, it is a respectable board.

Just curious why you are only looking at the above listed boards? Is it a price thing? If not, and you are looking for the best sound in a DP, you should also consider the major players' flagship models.
Without getting into the console DP units, these are definately worth looking into @ ~ $2k
The Yamaha CP300 or S90es/xs
Roland RD700sx/gx
Kawai MP8 or MP8II

I've tried the Pianoteq a fair amount and like it very much. I also have the Vintaudio Yamaha C7 samples and like the close-mic sample a lot. There's no reason on Earth why you couldn't buy either of these in a rack-mount or keyboard for way, way less than the V-piano. Pianoteq even has a Linux port now. Get some of the younger coders on it if the oldsters can't handle it. That's what they do in the US anyway (our corporate masters would gladly throw them under the bus for a little more corporate jet fuel $).

I'd buy it, that's why, and you would too. What are we talking here? 1GB of flash, 1GHz processor, a bit of ram, almost nothing these days. You can put it together yourself at newegg. I'm so tired of the table scraps they've been peddling. They're so used to know-nothing parents looking for a deal on a first keyboard they don't know how to make anything real anymore. Sometimes I wonder if they are colluding in keeping our expectations low and their prices artificially high...

The P155 manual says it has "Sustain Samples". You are correct, this isn't the same as pedal down velocity layers (if that is what you meant). But that is better than nothing, particularly when compared to other digital pianos (which isn't saying a lot).


Yes, but only when compared to other digital pianos (which isn't saying a lot).


CP300, RD700sx/gx, MP8 all have looping (so says Purgatory Creek). No way in hell am I laying serious bucks down for inferior technology. PC samplers had it all over these guys like 5+ years ago, who are they kidding? Not me.

You can check if your DP handles "sympathetic resonance"(some manufacturers call it "string resonance") or not.

Without pressing the sustain pedal, press C4(center C) with ppp and then press F3(F which is lower closest to C4) with staccato with strong accent. If you hear the sound of C5(one octave upper than C4) your DP handles sympathetic resonance.

This happens because the 3rd harmonic of F3 is close to the 2nd harmonic of C4.

It hasn't progressed nearly as fast as the PC sampler / emulator market. No disrespect, but what do the engineers DO all day at Kawai? Sit around designing the next micro-incremental advance in digital pianos by adding/removing a button here and there? Look at ancient PC samplers and say "we could do that"?

My prayers were answered years ago with PC samplers / emulators. I just want someone to put it in a rackmount / keyboard. Is that seriously too much to ask?

This board is full of people who are quite interested in these details, otherwise it probably would not exist. Also, leaders lead, they don't sit around following blind consumers and their idiot expectations, they are busy dreaming up and implementing the next big thing.

Please excuse my rudeness, my hair would be on fire 24/7 if I were able to work at Kawai (or Yamaha, Roland, Korg, etc.).

It is a riddle wrapped in a mystery inside an enigma...

I learned guitar on a very lame Harmony that had rather serious intonation problems - there was really no way to tune it correctly. It was a serious impediment to my musical education and enjoyment. The first instrument can make or break an interest in a whole area of interest of one's life.

[edit] If you're talking Casio CDP-100 as as low a one should go, I'd agree. But many of my wife's students play on crap toys, and it is often impossible to get their parents to upgrade once they have made the initial investment (however small). Everyone suffers as a result.

I've been thinking lately, I'm a computer engineering student so how hard would it be to design something like this on an FPGA? We could keep it simple at first, and just use a single piano sample, off some flash. it looks like the memory involved would be pretty cheap. I have quite a few ideas about simplifying the system, especially if I can give myself some ridiculous polyphony (704 polyphony, or 8*88, would make me happy).

I have concerns with hardware costs and how much processing power I'd need, i'm especially doubtful that a cheap FPGA could pull it off. nonetheless, I like it!

I think I have a senior design idea.....

I learned on VHDL but switched to Verilog for a big project a couple of years ago - I highly recommend Verilog (less verbose, etc.). Do your project on a Xilinx Spartan3 demo board (can be had for <$200, has flash, DRAM, LCD, AD/DA, RS232, VGA, etc.). I use the simulator in Altera's Quartus tool (very nice). Not sure if FPGAs are the way to go for a full-blown commercial synth, but the smaller ones are showing up in all kinds of things lately. Most now have hardware multipliers / DSP sections associated with 1/2 or more of the block RAMs.

I really love digital design, it is much more creative than I ever thought it would be (I was an analog hobbyist in my youth)! Look at www.opencores.org & anywhere else you can find ideas.

For this kind of thing you will probably end up needing to understand DRAM interfaces (complex) and the DSP blocks.

You can really differentiate yourself from the other EE coders by making FPGAs your thing, it is a niche job that requires a special skill set. It is true digital design. I enjoy working with both the board guys and the C coders, and I never really have to get my hands too dirty or worry about external deadlines too much.

If I were an EE in the synth / digital piano department of Korg, Yamaha, Roland, KAWAI, whatever (drool, drool) I don't know if I'd be able to get out of bed in the morning and hold my head up after having my ass repeatedly kicked by cheap PC software / hardware for the last decade or so. Don't those guys have any self-respect? How do they show up for work every day and manage turn on their PC, knowing that it pwned almost their entire product line years ago?

Sympathetic (or string) resonance is the simulation (be it sampled or modeled) of undamped strings being excited by other strings, due to harmonics as mezzo-poor points out.


I guess I am a bit confused. I (mistakenly?) thought the intent of your post was to gather opinions on the realism of DP's for recommendations for students. Perhaps it's more of a discussion on the state of the art and/or price and demand.
PC ROMplers are a far cry from a quality stand alone final product which offers musicians a viable alternative to an acoustic piano.
Inferior technology?? Hardly!!