Casio XW-P1 Acoustic Piano Sound

[eponn]

I really liked the sound of the acoustic piano XW-P1. That sound is derived from some of the Privia line?

[Technics]

I saw some answers at Harmony Central where there is an extensive review - dozens of pages long with pics and sound samples.

Look for post #71 from Mike Martin, he elaborates on the waveforms a bit - I think it was on the 4th page.

The whole review is worth reading. Tons of questions and tons of great answers till the very end.

http://acapella.harmony-central.com/showthread.php?2900825-Casio-XW-P1-Performance-Synthesizer/page1

[AlenK]

Given its title this thread really should have been started in the P1-specific section. But since the G1 has piano tones too (although they are not the same tones, according to Mike Martin) I suppose this location will do. Since I have more questions about the piano tones that were NOT answered in Craig Anderton's excellent Pro review over at Harmony Central, I'll revive this thread to add them.

 

Mike has previously stated that the stereo piano in the XW-P1 (the grand, at any rate; presumably most of the others are just tweaked versions) is triple strike, meaning sampled at three different velocities. (The equivalent in the G1 is only double strike.) First question: Does the P1 simply velocity switch between these three samples or does it crossfade between them? I believe that the Privia series of digital pianos do some sort of cross-fading or interpolation (Casio has their own tradename for whatever they are doing, of course) and my assumption is that the P1 doesn't do it that way. Please correct me MIke if I am mistaken. 

 

Second, are the piano samples in the P1 sourced from any other Casio product? I'm thinking they might be the same as those in the WK-7500/CTK-7000 keyboards. I'm also thinking that they (and those in the WK-7500/CTK-7000) are related or sourced from the samples used in Casio's Privia line, at least the older models. (I believe the current models use four velocity samples.) Note that Mike Martin didn't comment on where they came from in the Harmony Central Pro review. Surely there can't be any reason to keep it a secret; Casio can be proud of their piano samples, IMO.

 

Third, does the P1 implement any kind of simulation of the sound of the open strings when the dampers are raised by the pedal (sustain footswitch), as the Privia models apparently do? From what I hear with my ears I am pretty sure the answer is "no." If I'm right is there any way to sort-of simulate this with the P1's existng sonic resources? Don't get me wrong, the piano sound is good but perhaps it could be tweaked to be even better.  

[Mike Martin]

The piano sample in the XW-P1 is most similar to what is in the WK-7500/CTK-7000 although I can not say for certain if they're identical.  It is a three layer stereo piano sample.   No, the XW-P1 does not do any kind of damper or string resonance simulation.

[Gnomo Alegre]

I want to remind that the CASIO XW-P1 is a synthesizer with keys. I don't understand why gives more importance to the sound of the piano.

Based on my experience with the CASIO XW-P1 is used more electronic. The buttons and sliders are very effective in using them. For people who like electronic music will draw party.

If you want a piano, why buy a synthesizer? It will never sound real, never. They will sound similar.

Piano or synthesizer, XW-P1 is bestial.

[AlenK]

With all due respect, David, why do you think we are giving the piano sound more importance than it deserves? The XW provides it so why can't we ask questions about it?

 

It should be obvious that Casio themselves thought the piano sound was pretty important for the XW or they wouldn't have devoted so much sample ROM to it (less so in the G1 than the P1). The fact is, many - perhaps most - potential customers won't look twice at a keyboard, especially one designed for performance, if it doesn't have a decent piano sound. Casio knows this.

 

They also know that the market for the XW synths isn't dedicated synthesists, who nowadays will be looking for lots of knobs and a really clean virtual analog (or real analog!) synthesis engine. (The XW's solo synth is not exactly the "cleanest" sounding around.) Purchasers will expect lots of other sounds beside synth tones, including piano. I bought an XW-P1 _because_ it could make a wide variety of sounds, both electronic and acoustic. It isn't the best choice for either of those types of sounds, frankly, but having them in one board along with some intriguing sequencing makes for one fun product.

 

Personally, I'm happy that the P1 has a decent piano sound. It's yet another sonic resource in a long list of many that this keyboard provides. Granted, it's old technology (triple-strike velocity switching, 32 note polyphony for stereo and no string resonance, like the Privia PX-110/310 from ~2006) but I'd wager that it's more than adequate for almost everything an XW user will need it for.

[- T -]

David

 

In support of Alen's comments, we have only to look at Roland's Jupiter 50/80 synths.  While the "purists" might argue whether these units deserve the name "Jupiter" or exactly fit the classification of "synth", Roland defines them as such and obviously was targeting them at the synth "purist" market.  Even so, they still went to great lengths to provide a fairly decent set of acoustic instrument voices (pianos, strings, brass, guitars, bass, etc.), recognizing the performing artist's need/desire/expectation for inclusion of such sounds in the same unit as the "pure" synth sounds, and as Alen pointed out, the buyer has every right to expect these sounds to be of, at least, reasonably usable quality - particularly the piano voices.

 

As a matter of fact, Roland also uses this same architecture in their entry level and MOTL synths - the Juno Di and Gi, and I have no doubt that those buyers have equivalent expectations as to sound selections and quality.

 

I full well understand what you were getting at, but I think it is now more a question of market demand.  Both our technology and our expectation have advanced well beyond something that required 30 to 40 minutes of tedious patching and knob twiddling, only to produce something that sounded more akin to someone stepping on a cat's tail  -  kind of like the early days of driving or flying - where you took pride just in GETTING there.  For all the excitement and adventure that provided, the average auto or small plane buyer of today expects something a little more than that.  So, sit back, relax, and enjoy the "new" ride.

[AlenK]

[begin novel]

deleted

[end novel]

[- T -]

Alen

 

I am sorry you spent the time and energy writing your "novel" because of my (poor ?) choice of words.  To be quite honest, I personally feel that the Super Natural acoustic tones of the Jupiter 50/80 are nothing short of absolutely stupendous.  That is really why I bought the JP-50 to begin with.  I chose the milder "fairly decent" wording in order to avoid a myriad of distracting "Whose sounds are better?" wars with all those "argumentists" who seem to scan threads like these for no other reason than to jump all over such statements.  I was also trying to avoid insulting our hosts here by dumping lavish praise all over a competitive product.  I chose the JP-50 simply as an example to David that even the synth manufacturers recognize that performing musicians do not live by synth tones alone.  BUT . . . I enjoyed your "novel" anyway.  It was really nice reading !

[AlenK]

I actually do hope that what I said here, especially my praise for the competition, will start some debate. After all, you have to stir the pot regularly if you don't want the soup to burn. And Internet-searching argumentists would actually be a good thing for this forum. We need more participants. I'm gettin' tired of seeing my name in the recent topics list (and this last one won't help!)

[Brad Saucier]

I'm gettin' tired of seeing my name in the recent topics list (and this last one won't help!)

I thought I was the only one that felt this way.  

[CairnsFella]

On 6/22/2013 at 4:39 AM, AlenK said:

The machine really needs to allow modulation of parameters (such as whether a layer or oscillator is heard or not) or triggering of events (such as guitar strums) according to things like how many keys are being held down (one = non-legato, two = legato, three or more = chord). It's hardly a new idea. The ill-fated Technics SX-WSA1 allowed that nearly two decades ago.

 

I arrived at this old thread for an entirely different reason (virtual controllers to be specific)

 

As a WSA1 owner I did type out a whole 'contrary stance' to the notion that modulation could be controlled from the tone triggers. Having thought about it for a while though (and I am not in front of my gear at the moment) I guess what is stated is true. I am unsure it is quite as flexible as the testament implies, but I then I suppose that is being picky and down to interpretation. The reason I was going to contest this is because I seem to recall that the most directly related modulation applied is really layer driven.  The reason I recanted is because the layers can be trigger (or velocity, or key) driven.

 

What I will state though - despite the age of the thread and the fact that people have moved well on - is as much as the Technics did whatever it did 20 years before, it was - at release at least, several times more expensive. I am not going to get into inflation, or the opposing effect of decreasing tech prices, but still, it was a much higher end instrument. Yes, many mid and high end boards currently surpass it in most areas, and I am sure plenty of low end synths surpass it in some areas, but I am fairly confident that you still wont find a synth with the WSA1's build quality, multi-timbrality, polyphony, depth of programming, effects, aftertouch etc all in the same board for  the price of the XW. My point being that stating a current board doesn't do / should do something a a twenty year old one did is a little arbitrary.

 

I mean, my current three year old car doesn't have a vacuum operated windscreen wiper vanity panel. What gives. Even my 47 year old Corvette had that

 

 

[AlenK]

The point was simply that the "technology" (idea, really) is quite old. There was no intent to compare the XW-P1 to the SX-WSA1. If that's what you got out of my reply in that thread then I'm afraid you misread it. 

[CairnsFella]

22 minutes ago, AlenK said:

There was no intent to compare the XW-P1 to the SX-WSA1.

 

In which case I must apologise. You are correct that I "did" read it as saying the the tech should be in the XW. (Which I wouldn't have contested anyway 'as such', it was more the point of where does one draw the line, as despite being old tech, there will still be a limit as to what can be incorporated into a relatively budget board).

 

22 minutes ago, AlenK said:

I'm afraid you misread it.

 

Guilty as charged I guess. 

[CYBERYOGI =CO=Windler]

Can Casio XW-P1 and -G1 split keyboard midi input from the sound generation?

If yes, you may connect a PC (or even programmable mobile device?) and use an external midi mapper program to switch sounds through velocity, pedals etc. through sysex commands or whatever the Casio synth supports to modify its sound parameters by external controls. Or is such a loopback mode impossible by latency or lack of parameter input?

With piano sounds I guess that velocity controlled envelope or brightness (VCF?) control is more important than having many switching velocity sample layers. E.g. the 'piano' sample of "Miracle Piano Teaching System" and Yamaha PSR-230 (both 1990th) sound terribly boring because they only change volume and nothing else. A piano should grow duller with less attack when played softer (easy to implement e.g. by truncating the start of the sampled attack phase and compacting the envelope shape). Even a cheap FM sound IC can do this. I e.g. would love to hear an editable velocity sensitive version of the ancient Casio SA-series piano timbre (made from Phase Distortion or Triangular Wave Modulation?) in modern keyboards. The count of sample velocity layers is least important for piano-like feel and expression; it only starts to matter once you want to emulate the exact timbre behaviour of an acoustic piano. (But rom capacity is dirt cheap now anyway.) String resonance may be approximated by a kind of reverb, which is anyway part of most modern synth engines.

By the way, even the stairwave based Casio CT-6000 from 1984 had already some complex velocity controlled features (e.g. accompaniment changes), so this is nothing new. Casio should add a modulation source matrix to their synth engines to make more things user definable.



 

[AlenK]

@CYBERYOGI =CO=Windler

 

You can disable local control for each part separately, which has no affect on MIDI output for each part coming either from the keyboard (the four zones) or from the step sequencer. So it should be possible to do what you're describing. I suspect, however, that latency could be a big problem if you route through a PC. I would give it a try anyway. 

 

Regarding the timbre variations of piano sounds, this is a big topic. Just changing filter cutoff doesn't sound particularly authentic: that's how early samplers did it because they had so little memory (e.g., Korg DSS1 had only 256K 12-bit words = 384 KBytes!!). Samples at multiple velocity levels are required for decent emulation. Casio, Roland and Korg all have algorithms to more-or-less cross fade between those samples to avoid obvious switching. The XW-P1 of course simply switches between three samples, which while not optimal is good enough for non-critical applications. 

[CYBERYOGI =CO=Windler]

I don't own a Casio XW-P1, so it is nothing I tried out (but USB is likely faster than legacy midi). May be I think a bit much the 1980th way.

What velocity dependant filtering and envelope change can do is approximating the musical feel/purpose of a piano (hard played notes get much more emphasized/pronounces than soft notes), not perfectly simulating an acoustic one. E.g. important part of the piano "feel" is also that the bass range (wire wound strings) has a rougher grunting timbre while the rest sounds duller. Such things could be easily approximated by using e.g. unequal or multipulse squarewave in the bass range and a duller waveform for the rest. I always wondered why average FM home keyboards lack this feature. So either the bass range is way too dull (hollow or sinewave-like) or high notes too bright (harpsichord-like). Before sample based digital pianos arised, since Rhodes apparently nobody really cared about simulating this timbre difference.
 
Designing something like a Privia by sampling every small bit of mechanical noises of an acoustic piano and putting things back together in a musically plausible way is certainly an art of its own. I remember when I programmed VisualPinball games and had to attach sound samples to playfield events, and it was very hard to get a credible feeling of a electro-mechanical pinball machine despite I added relay and reel counter samples to all internal events of the emulated mechanism. Making an exact digital "clone" of an existing piano is certainly hard. E.g. there is the "Hauptwerk" software instrument for emulating existing church organs by such an approach. While this is also meant as a tool to document historical organs (and their sound changes after restoration) it may be not the perfect method to make the musically most playable organ.

I enjoy e.g. old tablehooters and toy keyboards for their simplicity and technical elegance rather than perfect simulation of existing instruments.