Pianoteq 5 impressions (Page 4 of 7)

I have the same reaction to most of the early pianos, I find most of them almost unlistenable.  I also thought, 'how were some of these delicate and beautiful early pieces written on such coarse sounding (to my modern ears) instruments'?

It's a sharp reminder of how much things have changed, and how much more refinement was achieved by piano makers over time.  And how you work with what you have, and how a great artist can use anything to express themselves. Quite incredible really.  That's also why some of us give PT such a hard time - piano is such a complex instrument.

I'm sure if I spent much more time with the older pianos, I could gradually warm to them.  But right now I find them grating and empty-sounding.  Spoilt by progress : ).

ReBased, did you listen to the Kremsegg demos? For example:
J. Brahms Sonata No. 3 played on the Streicher, page https://www.pianoteq.com/kremsegg1
F. Mendelssohn Rondo played on the Bechstein, page https://www.pianoteq.com/kremsegg2
Don't you find them exquisite? I know of course all this is a matter of taste...

Note that not everybody agrees on the progress. I heard people saying that with the modern pianos we have lost a great variety of sounds that existed in the XVIIIth Century, because of the many piano manufacturers who were building pianos, each one with his particular receipt.

Listening to the tracks you suggested:

- Streicher: J. Brahms - Sonata No. 3 in F Minor Op. 5.  Yep, beautiful.  Much better than the top demo.

- Bechstein: F. Mendelssohn - Rondo Op. 14.  Also very nice.  Interesting how dark it gets at the lower velocities, yet it still sparkles at higher velocities.  What causes that in the piano design Philippe?

The Dohnal is indeed the earliest piano among the 4 pianos from Kremsegg 1, so its sound is the most "distant" from the contemporary standard. It takes a certain time to get accustomed to this kind of sound, but once you are, it can be very pleasant and refreshing. Of course, it cannot be used in such a broad variety of music as contemporary pianos do. This might be the principal contribution of modern pianos: you can play every music on it, from Bach to Debussy (just to take two extremes in the sound palette). This is obviously not true for the Dohnal.

I guess it might be possible, but not easy. The difficulty is not specifically in mixing the characteristics from a physical/mathematical point of view. As long as equations are involved, it's OK, we all know how to compute (a+b)/2  and more complicate things. The difficulty starts when aesthetic chimes in. What would be the interest of an hybridization if it does not sound nice? The problem is that I have nothing to measure the beauty of a sound!

I'm very much in favour of the sort of documentation as provided in this case by Joshua's waveform examples. This evidence can be highly instructive, if interpreted with due consideration to all pertinent factors. As things stand we are not privy to all possible pertinent variables.

So, with this hard documentation as a great start, the following questions have occurred to me, as I attempt to understand these comparative illustrations:

[first, I assume from the Joshua's comments that 'his real thing', the actual acoustic piano decay recorded, is a 100-year old Chickering baby grand]

1) I assume that the environment in which the piano is located will have imprinted itself upon the resultant waveform i.e., we are not looking at the 'raw' output of the piano in isolation (which would require an anechoic chamber – with a good stout floor membrane...)

2) I assume that the Pianoteq models are 'running' consistent virtual environments as part of a meaningful comparison i.e., if a room mode or reverberation enhancement is applied to the piano engine output, this is at least loosely the same in character for the illustrated waveform examples. This would make comparison meaningful (and fair), at least for the Pianoteq 3/5 simulations.

3) In that case, we have the following consideration: To what extent does (can) the applied virtual acoustic be fairly compared to that of the Chickering's native environment? And, if the Pianoteq examples are indeed 'raw' piano engine-only, what are we to make of the comparison with the real thing, again reasonably assuming the unavoidable presence of live ambience/reverberation.

4) The convolution effect of a real acoustic upon the real piano would be generally similar to that of an electronic equivalent upon the modeled piano, in that it will colour the original significantly, though not beyond the point of recognition. That is to say, the general characteristics of the piano waveform will 'survive' the reverberation. So we can still learn 'something' even in the case of a possible slightly apples/oranges testing comparison.

These questions can of course be answered to a great degree by Joshua himself, and I would be very glad to learn more details of the recording process.

5) This whole exercise leads me to ponder what might result from a comparison of a new, mint Steinway D (or similar) and the Pianoteq 5 D4, set up and recorded in similar real/virtual environments.

Another poster, njaremko, has offered this observation:

"The decay of the D4 behaves and lasts exactly the same as my Steinway sample libraries."

The conditions under which these sample libraries, themselves, have been created, is of course a natural and essential consideration, in turn.

So I remain eager to learn more on this whole issue. For my part, I have often (from v4) deliberately held down long notes, listening to the decay – not to critique it, but to marvel at its mysterious complexity (knowing it is a synthetic creation, albeit a very sophisticated one). Real pianos, no matter how diverse in all areas of tonal generation, of course sound 'real' simply because they are real. Too many of them would do better as toxic firewood than as musical instruments, sad to say (not something I am remotely recommending, in case I have aroused unintended offence by that wry remark)...

Cheers,

Stephen.

New feature request.

Option to replace damaged mics and broken strings, buying new ones from the pianoteq store.

Ah...Sorry. I still wonder how much amplitude and possibly the velocity scaling are the major differences. I do see that the initial decay stages, right after the attack, seem more abrupt in V5. On the other hand, note the final decay on the last chord in all three instances. The decay is as long as the decay in the other recordings. It is less loud, however.

@stephenphillips Good points. Reverb was "on" in both versions, and I didn't touch the settings.

[scratches head] Same MIDI controller. Same velocity curves. Asking if amplitude is the difference is basically paraphrasing the issue. lol Amplitude after attack is attenuated in PT5 relative to both PT3 and the real thing, particularly when repeating notes while holding down the damper pedal.

That's just a fraction of the decay. If I posted a pic of the full track it'd be several pages long. lol Plus most of it would look like a straight line (0 gain) even though all 3 were still sustaining. And "less loud" is exactly the point.

Let's keep in mind, the impetus for this was my ears. I was playing this piece I've been working on, and when I got to that very part (which is near the beginning) I immediately noticed that the bass note was virtually gone by the time I changed A-sus to A-9, so it didn't sound as rich. That's the type of anaemic response I expect from older digital pianos.

That's what I would have written too, not as well as Jake though. Another observation is that one can see from the three figures that v5 and the real piano have a better "pronunciation" than v4: each new chord is visible, whereas in v4 the 2 last chords are drowned in the previous one, which may be the reason why people qualified it muddy.

Nevertheless, I'd like to insist on this Joshua: you have the hand on this, first by reducing the Direct sound duration (make this a priority underlined by ReBased), then by playing with the Impedance.

That's EXACTLY where I was hoping to go with this! I didn't want to presume, though. Since you said it, here's how I look at it. Muddiness was indeed an issue with version 3. Attenuating sustain so that attack transients are more "pronounced" is NOT the answer. It's a workaround, and a poor one, since you sacrifice sustain. And you can see from the waveform for the real piano, that's not how it behaves.

I'll play around with the other parameters, though I would think the default settings are meant to be as close to reality as possible. I.e., if I wanted to artificially pronounce each chord, I could LOWER the impedance and sound duration.

Joshua, it is not a workaround at all and we did not sacrifice anything - that's your interpretation but it does not correspond to the reality - it is the consequence of having a model that now comes much closer to the sound of the acoustic pianos that are modelled: Steinway D, Bluethner Model One, Kremsegg pianos, etc. I can guaranty you that the wave forms of the current model are extremely close to the reference we used, as well as to the many wave forms I have observed since many years - hence the default settings are indeed meant to be as close to reality as possible. This has been validated by a very competent beta team who has worked one full year in refining the models in their finest details.

Perhaps thru being in phase they drive the soundboard harder, and so energy gets lost to the air sooner.

So, when NOT in phase so well, loss is slowed. You'd imagine, if this is right, a tuning fork would sound longer when NOT touched to a big surface.

I used the term 'narrow lens' above in an attempt to express the phenomenon of expectation vs satisfaction/inspiration (the two are related...)

I do not believe (nor did I intend to indicate) that I consider the modern stable of concert grands (all makes) 'better' timbrally than their predecessors. That would be too great a leap. And we live in an age of historically-informed sensitivity!

The mindset of the great composers and all others living in those former times were informed and contextualised by the experiences to which they had access. Who knows what they would have made of later 'developments'? I suspect they would have rather approved, but cannot be sure (nor does it matter).

The important thing is that these men and women were enabled (compelled, even) to find a voice for their spiritual and aesthetic expression in the instruments they knew (instruments we might now be able to better acquaint ourselves with, in the luxury of our own homes, even in the middle of the night...) and in this, they felt no constraint or sense of inadequacy. This lesson, for us, is at a fundamental level, a philosophical one.

Thank you, Pianoteq.

Josh,

Could you post the audio files files that correspond to these three images? I'm not simply avoiding responding what you are saying--I just want to hear the difference. I particularly want to hear your piano. It's entirely  possible that it does have a slower amplitude drop-off than the source Steinway D. It may be older, but since it's a good piano and well-maintained, with no loss of crown, it may just differ from a Steinway D, although V3 of Pianoteq may have more closely approximated its decay rates. In other words, V5 may more accurately emulate the sound of a real D4, but it does not emulate the sound that you want.

Maybe a piano like yours should be investigated for modelling, if it has a precise attack and a slower decay that does not encourage what was called muddiness. Please understand that I am being serious. I'm hoping that we will see a modelled Steingraeber, and a Mason and Hamlin, and an older Baldwin grand, too, although I fear that many people in Europe have not heard older Baldwins.

An aside--some people might argue against the assumption that an older piano will not sustain as well as a recent piano. Many guitar players prefer the sound of older instruments, so long as they are well made and structurally intact. There are various arguments about why older instruments may sound better. Better construction and materials, for one. The age of the spruce tree cut for the sound board. And the wood may dry a little, so that vibrations radiate across more of the sound board--not the same thing as radiating faster, which would kill the sustain. I've seen the same arguments about older pianos--that an older Steinway B has a musicality that newer ones do not, for example. And old violins are still widely praised, although again the exact reason for the tone is argued over.

Please do post the new recordings.

It's a bit more complicate. In short:
1) If the unison is in tune, there is the well-known "double decay", a fast initial decay (direct sound) followed by a slow decay (after sound).
2) If the unison is less in tune, then:
- the initial decay is not as fast than when in tune,
- the final decay is faster than when in tune.
3) When the unison is largely out of tune, the "double decay" disappears, the decay being almost constant and moderately fast during the whole life of the note.

In the meanwhile, if you want to listen to a beautiful new demo:
J. Brahms - Clarinet Sonata Op. 120 No. 1,  see https://www.pianoteq.com/kremsegg1.

Philippe: re: the unison issue, thanks for the explanation.

Greg.

That's the one, njaremko. I had the number memorized wrong.

That's a beautiful performance. Was it performed on an historical instrument? Is it not the case that the MIDIs used for the demos of the historical instruments were actually performed on modern pianos? That could explain why some of them have a rather dry tone. The shorter sustain of the historical pianos demands a different style from the performer. Pianists like Malcolm Bilson, Ronald Brautigam and others make strong arguments for these instruments.

Cheers,
Clive.

Good demo. In this (non-blind) demo, regarding tone, I think Ivory sounds a bit better and more authentic, but I mainly just hear a different overall tonal balance, and that difference is just a difference - without having a preference for one or the other.

Greg. (not an Ivory owner btw)

Hi, TJ! Thank you for this good comparison. For my ear Ivory has a little bit round and rich sound with more presence of low frequences and action noises. D4 is slightly thin and has a lack of "footing". But you are true - it's more vivid. Still some job for Modartt about sound and its timbre (and its richness).
Can you provide us with this midi-file to experiment with PT settings? Unfortunately, I don't have Ivory II for the time being, and your recording is a good material for comparison!

I forgot to mention that I was using the Bluethner on this PT Recording

Ah. by the way. I would have preferred the preset list in Pianoteq to be more based on popular mic positions and named after them. That's my first wish for the future. See what I mean. The collection is not exhaustive:
http://www.youtube.com/watch?v=3XlN3kc1Hx0