My Concern with PianoTeq

The reality is that I'm at a disadvantage compared to most of the pianists on this forum. I started playing the piano in my forties and I haven't developed any real refinement in my playing. I lack technique and I don't play anything that difficult.

With that caveat out of the way, I've noticed while playing the Ivory 3 and Studiologic Stage D, which are both hybrids (sampled/modeled), that they play better than any pure sampled based instrument I have. This brings up a consideration for me. What's more important, how a piano plays and allows me to express my undeveloped technique, or how it sounds in comparison to a real acoustic piano?

Sound is definitely important because I won't be motivated to play a virtual piano if it doesn't sound good. Pianoteq was the first VST I purchased a year or so after I started playing. I shortly thereafter started a thread on PW titled "Pianoteq purchase gone wrong."

https://forum.pianoworld.com/ubbthreads.php/topics/2615309/Pianoteq_purchase_gone_wrong.html

I was lamenting how much I hated the way it sounded. 😀 This was before I knew the difference between sampled and modeled virtual pianos. I have the latest version of Pianoteq and I still can't play it. So sound trumps playability when it comes to Pianoteq (for me). Then I went down the rabbit hole of VSL and many other virtual pianos focusing on sound and not playability since I was only a beginner and couldn't really play much of anything anyways.

Now I'm 51 and I notice I'm willing to sacrifice some sound quality/realism for better playability. Not to the degree of Pianoteq 🤮, but the Ivory 3 and Stage D are nice compromises. I find myself more motivated to play and develop better technique when I have a virtual piano that makes me feel more connected to the expression of sound that starts in my brain and travels though my body into the instrument. Modeling does that better than sampling for me.

I'm not sure what's going to develop faster, i.e., pure modeling, pure sampling or hybrids, but I'm grateful there are some good options out there and we're not limited to only pure acoustic instruments with the additional considerations that they bring.

God Bless,
David

David B Indeed, when done well, hybrid libraries can be better both playing and sounding. AcousticSamples has their V series, and it's a good example.

However, so far I have yet to see a hybrid piano library (sampled + modeled) that's truly gapless. On Ivory III, I sometimes can hear the modeling bit, if you know what I mean.

David B With that caveat out of the way, I've noticed while playing the Ivory 3 and Studiologic Stage D, which are both hybrids (sampled/modeled), that they play better than any pure sampled based instrument I have. This brings up a consideration for me. What's more important, how a piano plays and allows me to express my undeveloped technique, or how it sounds in comparison to a real acoustic piano?

Both are important and I think the dividing line will fall in a different place for each person. I agree that Ivory 3 has superb playability and it really makes a difference - it feels incredibly cohesive and smooth. However for me, I think the tone leans a bit more towards the artificial side than I am personally comfortable with, and after a while I found myself gravitating back to VSL because on balance I found it a more satisfying overall experience, despite the very real gains in playability that the Ivory 3 offers.

However, I really hope that they continue to advance the technology, as I feel like it is so close and just needs a little extra touch of tonal fidelity to put it over the edge. I would also love to see how the engine responds to different pianos and if the timbre of that particular Hamburg D has something to do with my impression of it.

Irmin I'm a PT fan. I like it for what it is

Same, I don't think it perfectly replicates the sound of an acoustic piano, but it's a fun synth to complement my sample libraries. Pteq became more enjoyable (at least to me) when I started to think of it as its own kind of instrument where I have real-time control over hundreds of params and experimenting with those. Continuous pedals mapped to celeste, pinch harmonics, buff stop, etc., poly aftertouch for string bends like a guitar, etc. There is a characteristic "modeled sound" sure, but it's a useful tool for practicing piano and it runs just fine on my phone.

That said, if I want a Steinway simulator, I'll play Modern D. 🙂

I’ve been kicked out of Modern D land, but it turns out this here thread is the “real” modern world!

Not to be eager, but you all know Pianoteq 9 is coming soon? and yes, it will be the mother load, a paradigm shift, and only the second coming would make a bigger splash!

Yes, Modartt got into AI early (bet you didn’t know that), and “the machine” has been crunching numbers at an exponential rate, so this “Newteq” will push Pianoteq into the stratosphere and beyond!

You won’t even need to “Power-ON, boot, etc…” but rather simply say “Pianoteq Activate,” and as if by osmosis, a version better than yesterday’s and already building on tomorrow will be ready for you to play.

Powered by AI means it’s a “self-learning” machine, so the notion of waiting for version 10,11,12, etc.. will be moot, because every single day you will have a new & improved “model” at your service. I mean, how great is that?!!!

Unfortunately, the improvements in Pianoteq have slowed to a crawl, and the next version will be the first to actually have regressed from the pinnacle which was the previous version. But get that one while you can, because the following version will be much worse!

😁

QuasiUnaFantasia Unfortunately, the improvements in Pianoteq have slowed to a crawl, and the next version will be the first to actually have regressed from the pinnacle which was the previous version. But get that one while you can, because the following version will be much worse!

Where did you get this information from? surely, it is not the case (and I know that from “the most reliable source”).

Pete14 It is a forecast made by a statistical model based on the historical developments of Pianoteq.

Well, I’m sorry to break it to you, but that “forecast” you talk about was made by a static sample, and not a statistical model, and if there’s one thing I know about static-data that would be that it is, well, static (lacking in movement, action, or change)….

Pete14 I'm willing to update the model upon arrival of fresh data (i.e. version 9). And - to be earnest (for just a moment anyway) - I would love to see actual major improvement in Pianoteq. It's just that my expectations are now so low that I find it difficult to believe in any improvements to the tone; they will probably add some features, but the tone will stay essentially the same.

Pete14 Yeah, dream on, Pete! I wonder what kind of a "reliable source" is this that you got this stuff from? Surely, it's your own imagination, because you are an imagineer! Therefore, that's what they are, whild imaginings that can only be fantasy.

Based on historic developments of PianoTeq, however, we know that it has been rehashing old stuff over the past 6 versions at the most, and its improvements are small, if not at all. More piano models, yes, but the sound is still uninspiring.

As a pianist, that's why I'm still staying away from PianoTeq.

QuasiUnaFantasia they will probably add some features, but the tone will stay essentially the same.

Yes, and not a piano tone, unfortunately. Heck, even the Roland and Kawai models on their on-board sounds are better tone wise than PianoTeq.

I have a feeling (not merely a "gut" one, but based on various technical insights) that the real problem for Pianoteq is that Modartt are adamant in keeping the CPU requirements very low, and for this reason they have to avoid including a lot of high frequency material, and this entails distortion relative to the "correct" tone. Some people apparently don't perceive any unpleasantness to the tone, and so they happily use Pianoteq; maybe this user base is simply enough for Modartt.

IIRC, YAM and KAWA Digital Pianos are all samples, AWM or PCM tech based. Roland is in the league of PTQ, modelling. And IIRC, Sir Pete is using YAM digital pianos. Yam digital piano and sampled vis are all samples. Why sampled vis are to the blame? Was there any unhappy purchase experience of those samples?

Another thing. Pete mentioned version 9 is coming. Let's think about the release cycle of PianoTeq, if there is one.

If a Google search got the info right, PianoTeq 6 was released September 6, 2017, version 7 in December 11, 2020, and most recent version 8 in November 16, 2022.

So, why don't we just wait? We still have a couple of months till 2024 is out, will version 9 come this year? It's about time they release it. And then we can judge if Pete is dreaming or there's truth in his posts. I'm willing to be on the side lines watching it all happen.

However, from 6 to 8, the tone hasn't improved much, which is still quite sad.

QuasiUnaFantasia Good point. Also, the CPU power on phones and other smart devices doesn't allow very much live processing, so you can have PianoTeq on these platforms, too.

David Lai However, from 6 to 8, the tone hasn't improved much, which is still quite sad.

I upgraded from version 6 to version 7, but I did not feel the sound improved at all. New features were introduced, but by then I had discovered samples ...

David Lai

Pianoteq Activate! 😁

Pete14 Dear Pete, We're not joking around or dreaming here. Face the facts, man!

pipAT

no, the whole thing is because he's a child, unhappy, always whining about how the things truly are.
he just wants to be different then the majority
if we were in the modeling team, he would be talking about how samples are the future, how great and high quality they are compared to modeled pianos.