Somewhat dissapointing first impression (sounds very syntesizer-like) (Page 1 of 2)

Must add:
I also tried to use 3 presets for Privia 150 I found on your site (which is very close to Privia 160, so I believe the keys are the same on both), but no one of them produced sound I could like, so I resorted to using default "Moderately slow keyboard" preset again.

Alex, you need to learn all important details, adjustments, before state such remark.

Velocity curves it's very adjustable on pianoteq, and it's very important to get it right before judge the sound.

Real acoustic pianos have all these strange noises and overtones.

However, it's possible there's something misconfigured at the hardware/driver level.  Can you make a recording of the sound your system is making?  Or download MIDI files and compare them to the what's on the Pianoteq website?

Thanks, billd. Now it worked smoothly. Sounds became even better than with this default preset I've been using. Though during the calibration it said that my keyboard can't achieve velocity value 127, which, as it said, most keyboards should be able to achieve. Have you also seen this during configuration? Or keyboard on PX-780 differs significantly from one on PX-160?

I also would really love to learn what actually did you mean by "smoothing" it after that? I'm still observing some strange discrods when I'm trying to smootly (legato-style) play adjacent keys; for example when playing D -> C repeatedly  in the 5th octave while trying to achieve sort of "seamless" transition, I keep hearing for a very brief moment how sounds overlap, creating some unpleasant discord sound. I can't see it on other keys, even on the same keys of different octaves. Is it how real piano behaves, or mb my keyboard needs configuration, or defected? But in general sound became much more clear and natural.

I'll try to do it tomorrow.

Those are nice, I would like to achieve the same quality of sound in my installation now.

I've tested it a bit more, and now I see that with the same velocity curve this effect cannot be seen on some pianos/piano presets (or may be on them it's not that noticeable).

Alex:

What I meant by "smoothing" the velocity curve generated by the calibration function is to reduce the highs and increase the lows, thus taking the curve closer to a straight line between the calibrated low and high point.  By the way the low point is not zero and the high point is not 127 on my PX-780. 
Also I should note that I do not always use this curve.  Some presets seem to me to benefit from a completely different curve.  For instance one of my favorite presets right now is "Model_B_Closed_Mic_modarrt_Video" which was supplied by Modarrt.  For that preset I use the velocity curve supplied which is a straight line from 0,23 to 127,127.  Any other curve I tried did not improve the sound.

Bill

Though Cassio sounds nice, it falls short of what can be heard in any good recording of acoustic piano play you can find on youtube or other resources (I mean those that are performed on really good - and expensive - rig). It also can't be compared to demos that can be found on Pianoteq's site.  So I'm looking for a way to improve quality of the sound, while not investing too much money in it (really good pianos are insanely expensive). I.e. to find some compromising variant.

It seems Pianoteq has a great potential, but to uncover it additional configurations are required. I'll be trying to find some preset I can be fond of now. Thanks to all for your suggestions and ideas.

Maybe you don't like it and that's fine. You should try the Model B though before you dismiss it altogether (also the R2 is quite nice if you're into Rhodes style electrics).

Wrt the velocity curve calibration (which you now have figured), I would consider that a bug. Obviously the software can in principle tell that there is something wrong when it fails to proceed with the calibration. It should at least try high-res MIDI and switch to it automatically when called for. High-res MIDI will become the norm soon enough and Pianoteq needs to accommodate it more seamlessly without every Casio PX (etc) owner having to come to the forums to figure out why their Calibration Assistant is not working. And while changes are being made to the software, the high-res MIDI issue should be clearly documented in the manual.

I'm 100% sure that I'm not 100% sure! Just that my standard MIDI keyboards feel a bit lifeless compared to the PX-160. When I am playing a piece or a section of a piece that only moves between say ppp and pp, I'm limited to just 10 or so levels --- might as well have a sampled instrument at that point. But since I cannot control my touch to within more than a couple of levels out of 127 at best, the jitter idea sounds very good. There is no reason I can think of why it should not work. I'll give it a try. Thanks for the tip!

But there is another shortcoming of 7-bit (0,1,...,127) MIDI. If I want to radically alter the velocity curve in software, there will be big gaps in places, so I am limited in how far I can push things. It's like photo editing/processing with 8-bit colour vs 16-bit colour. The 16-bit colour gives you "headroom" (extra precision to allow for crazy things you might do to the data). Same again with 16-bit vs 24- or 32-bit audio: 16-bit is fine for the final mix but you want more when you want to tinker.

I would not be so harsh on the OP. If he/she does not like Pianoteq that's their call. That's why Pianoteq offers demo versions so people can try it out and see if they like it or not. Maybe the OP would prefer Ivory or Ravenscroft or True Keys etc. I know a lot of people do. Or maybe even a Roland or an older RealPiano hardware module (they were the s**t back in the 1990s). Choosing an instrument is a very subjective thing.

Yes, sure, it is an argument but again, in reality how radically are you REALLY going to push the velocity curve? ;)

Yes I will, very radically! I had a longer reply but my dog ate it. (seriously, I pressed some funny key combo and it vanished!)

@AlexS

I'm not sure I'm hearing what you're hearing.
There are a lot of instruments there, try them out and see if you like any of them. I don't believe any of us like all of them.

I recommend the Model B (probably the most refined of Modartt's creations), the 1899 Bechstein from the Kremseg#2 collection, maybe the U4 and also the K2 (the latter is a 100% virtual instrument, not a modelled physical instrument), and maybe even the YC5 if the other grands don't float your boat. But there are lots more (including freebies).

If you were to find you don't like any of the instruments then I'd be very surprised. But if you find you dislike most of them then maybe Pianoteq is not your thing.

I'm sorry if I won't be able to answer everybody atm, it's pretty late there

First of all, good news: I've found at least one preset I liked This is YC5 recommended by somebody before. Especially I liked "Vintage" preset of it. The funny thing is it sounds VERY close to my Casio, but somewhat richer. And it doesn't have this annoying discord issue between C and D notes of 5th octave.

Thanks for info on Bluthner, that was interesting. But still it's suspicious: I know about sympathetic resonance, and I hear it a lot in 6-7 octave when strings ring long after they were played. But this particualr issue happens ONLY in this specific place in 5th octave (it does not occur during smooth transition between D and E notes of the same octave), and occurs on different instruments (and does not occurs on some at all, like YC5 I've mentioned above, and when listening to my Cassio's internal sound). That's really strange.

I'm using Rolland RH-5 for piano play. It's cheap, but relatively decent pair of "studio" headphones I bought especially for this.

PS. Re something odd in a specific region of the keyboard (only few notes affected like you're finding with the Bluthner) have a read of this recent discussion: http://www.forum-pianoteq.com/viewtopic.php?id=4221
--- I think it concerns the same notes you're having issues with (edit: no it does not, it's the octave above). Subsequently I was able to create an FXP file using Pianoteq Pro that seamlessly changes the settings just around the few problematic notes (problematic for that person's ears, not so much for mine) without affecting the overall timbre of the instrument. Such FXP files can then be used as new presets by anyone with Pianoteq Pro or Std (but not Stage, unfortunately).

Human brain tends to recognize things as normal based on references.

For example, people who is used to play in out tune acoustic pianos, in music schools of low resources, tends to thinks that's how a piano should sound and, after play in a very well tunned piano, will think that something is wrong.

In a similar way, people used to play with samplers will strange a bit a real piano harmonics.

My suggestions is to take some time playing in real and well tunned pianos, in a good music school, and you may find pianoteq characteristcs are more natural than you think.

1) You can press a key on piano VERY lightly and it won't hit the string, but there will be mechanical noises sounding out. Also, if you hit any other notes WHILE you have this silent key pressed, you get sympathetic resonances from the silent key (because its damper is not damping the strings).

Chopin87, well if you suddenly come to me on the street and touch my cheek very slowly, I will definitely react somehow But the problem is my Casio is dead silent in that case. And I can't even say I always press keys THAT slow, i.e. often when I just try to play something in very low volume of sound, I produce such "dead notes" involuntarily even with internal sound (and in Pianoteq too). That's why I thought mb I need to do something with the sensitivity of the keyboard, or speed curve, or both.

Another question (I hope billd or SteveLy will be able to comment on this as they seem to posses similar devices) is what should I do with emulation of different hammers' weights my piano does. I believe it just do this by setting different sensitivity for different parts of keyboard (like, on lower octave keys it's harder to trigger hammer action because hammers are much heavier there). Here is a quote from their site:

I saw disputes on whether it's a good or a bad thing over the Internet. I saw opinions that modern digital pianos shouldn't do this as modern acoustic pianos try to get rid of this differences through design enhancements. I believe I can disable this feature in my Casio, do you think this may enhance its compatibility with Pianoteq as a midi keyboard?

@AlexS

The link I referred you to was not so much to "solve your problem" but to give you some insight into what goes on under the hood and to show you that you're not the only one who wants little things "fixed" in the default setup of instruments. You don't need Pianoteq Pro to use an FXP file but you would need Pianoteq Std to take advantage of an FXP (customised preset) that includes note-by-note editing. And the forum is great if you want some help with creating a new preset that requires note-by-note fine tuning (and can't afford Ptq Pro) because other users who do have the Pro version will likely be able to whip one up for you. And in Std you can explore the parameters and decide how you want things to sound near specific notes. But we're getting way ahead of ourselves here. With the trial versions of Ptq Stage or Std alone you should be able to find good instruments and if need be tweak them to your liking.

Re no sound for lightest touch: that's how a proper piano KB should behave because that's how a real piano behaves. If you don't give the hammer enough momentum, it won't fly far enough to hit the string. Some piano KBs don't work that way (like my SL Numa Nero) and that's basically a bug; but it can be fixed using Ptq's velocity curve. When calibrating I interpret "Press a key as softly as you can" [while still making the softest sound you can] --- but I don't think it matters a great deal if you make a sound or not. As long as you see some MIDI output (green vertical line) pop up on the v-curve graph. Even if there is no sound MIDI level 0 should be registered. You can always manually refine the velocity curve after the initial calibration (and typically you would).

As for the Casio hammer response timing thing, I think it's just a gimmick. There are more significant delays associated with the speed of sound in air. But maybe it's just a badly worded bit of marketing spiel and it's meant to say that the KB responds to the movement of the hammers, not just the keys themselves. To the best of my knowledge the three sensors on each PX-160 note consist of two sensors for the key and one sensor for the hammer. Ideally you'd want all the sensors on the hammer but that's trickier to implement. So maybe the marketing talk is meant to be a substitute for something more honest like:

We have sensors on the hammers for more realistic response, unlike some of our competitors who put all the sensors on the keys only and ignore the hammers [e.g., SL88 Grand]. But to keep costs down we could only manage one sensor per hammer and the other two are still only for the keys. But we fixed this up with some clever calculations that combine the input from key and hammer sensors to best mimic having all three sensors on the hammers.

But don't take any of this too seriously. I'm only guessing, trying to read between the lines of marketing talk smoke & mirrors. (It could also be that all three sensors are on the keys and the hammer movement is calculated from that; I've seen conflicting illustrations online. One of these days I'll open my PX-160 up and will look for myself. But I've already done an experiment that strongly suggests there is a sensor on each hammer. I'm a physicist btw. I like to know how things work.) Bottom line is, don't worry about it, just play the thing because it works very well out-of-the-box for such a cheap KB.

Same.

Noob remark here:
May be it is all much simpler than you think. May be you've got resonance happening between particular notes and the system that makes room imitation and hull imitation? . Did you try to turn off all effects to see if the problem is still there? Did you try to move virtual mics in some weird position, like behind the deck?

One year ago I used to play sampled pianos and I liked them more than the pianoteq demo. Then, since I never managed to tame Windows to run without mysterious system processes randomly consuming resources, I decided to "settle" for pianoteq in linux.
After a few months, I tried the sampled pianos again and surprisingly found out that now I prefer the pianoteq sound... I don't know exactly why. Maybe it's because I made a custom calibration (I have a VPC1 but was not satisfied by its factory pianoteq profile).
I'd love to try a real acoustic grand piano, though. :-)

That's true. I've played some very nice grand pianos, but the really good ones are few and far between, and beyond the economic reach of most people including myself.

Most of the pianos that one might find locally, in the homes of family and friends, in schools, churches, community buildings, nightclubs, etc., are usually somewhat out of tune, not well maintained, and have a significant number to many eccentricities both of action and of sound, making them a much less than truly satisfying experience to play. One exception to that general rule are the fairly nice pianos kept in good shape in college auditoriums for student, faculty and visiting artist concerts and ensembles.

There is no direct connection between the hammer and string on an acoustic grand.  The hammer is flipped up at the string.  There has to be a certain amount of velocity in the key press for it to be flipped up.  A big issue with electronic controllers is the electrical contacts under the keys and exactly how they respond to nuance of touch and speed or force of the touch.  This then has to be interpreted by the computer voice (Pianoteq piano simulation).

There seem to be two weak areas in electronic instruments.  The feel and response of the keyboard.  Is it musical? Can it be controlled with nuance, and does it feel predictable.  Secondly, the speakers themselves with must replicate the sound of an actual acoustic instrument as authentically as possible.  The reason Pianoteq is popular is that it sounds "real."  I like it because it is very musical, and it's quite possible to play it emotionally.

This is very interesting for me. My setup is a Casio PX-160 going into a PC and a Focusrite Scarlett 2i2 going out to KRK Rokit 6 G3's. I'm not 100% happy with the sound. My biggest dissatisfaction is with the mid-range, but not having experience with concert grands, I assumed the most important upgrade for me would be to a better keyboard.

I have a few questions.

Has anyone here played around with the KRKs and with some relatively expensive monitors and noticed a big difference?

I play in a relatively confined space like 12'x16'. With more expensive monitors, is the subwoofer that important? I would have to put the subwoofer in the corner. Isn't this supposed to be bad?

Another question I have is that with my monitors pointing at me diagonally, each of them is practically touching the drywall. I don't have much space to pull them away from the wall. How important is it for me to acoustically treat the wall to reduce vibration?

I think it would cost me about $500 to upgrade to the Dynaudios without the subwoofer, and I'm wondering if it will make such a big difference.

I've played a few home pianos by Yamaha and other companies that had the monitors places under the key bed, facing the floor. This arrangement has often sound good to me--the sounds are dispersed a little as they bounce off the floor and back up to the player. Not sure how to replicate this arrangement with large monitors, however. They take up so much space that, even if I built a stand designed to hold them in that position, they would protrude so far downwards that there would be no room for things like legs and knees....