Pianos you want Modartt to model? (Page 3 of 10)

Can't really agree with that as I consider there is no ideal piano.  Even in terms of mechanical condition, I think the condition slider demonstrates that Modartt themselves recognize that real pianos have random deviations from the perfect.  I tend to think of the 100% version as a "factory perfect" model and sometimes a little random deviation is needed to "flavor" the model a little (there's just no English way to describe the quality of sounds well).  I think I've seen other posters say similar things - for them there is no ideal and the sound they're seeking is actually not a perfect one, but one with the imperfections of a real individual piano they've played.  So everyone has a different idea of ideal.

Assuming the K2 is an abstract model and not based on anything else then I think I'd say it represents a type of piano sound that the other models don't quite have.  My own idea of an ideal piano has changed a lot since I started using the historic piano models more.  They're totally changed my classical piano experience - e.g. Beethoven's Moonlight sonata using the Conrad Graf or the Frenzel (a much under-rated model, IMHO) just sound so very different and seem to need a different way of playing compared to a modern instrument.  Period music with different instruments now seems very much feels like an exploration to me.

So no idealization of a piano is possible for me, any more.

That said, the more I play with the demo Steingraeber I start to think of that as what I personally want from a modern piano sound - sounds like I'm planning my next purchase. :-)

I'd like Pianoteq to model a Steinway D.

I understand the interest for additional libraries, but I'm way more interested in Pianoteq getting their core discipline highly refined before pumping out more of them.  I'm sorry, but the Steinway D is not good enough...not for classical music anyway, and I'm surprised Steinway finally buckled and gave them the licensing.  I own the Pro version of Pianoteq and just purchased the VSL Steinway D because it sounded fantastic but - more to the point - actually like a Steinway.  PT6 was definitively better, but I still can tell very easily that PT is synthetic in a way that makes me feel uncomfortable.  I believe in modelling as a method and I don't think it needs to be identifiably synthetic, and though I considered selling my PT after getting the VSL I am almost keeping it as my "investment" in the future of the modelling method, because I think they can get there and I'd like to be there when they do. 

But then I see them start to release organ models when a standard classical piano is still not fully developed, which imo is a diversion.  The mindshares and labor can only go to one place at a time, and now apparently they are being funneled into OrganTeq, after already being funneled into new model releases.  I'm for a stronger trunk, not more branches on the tree.  I'd like to see PT model a Steinway that professional classical pianists cannot tell is synthetic.  THEN maybe get into organs.

Sampling technology is getting better, too.  They also have the leading benefit of the sound of the actual instrument.  So I wouldn't slack or get diverted.  For myself, after purchasing the VSL Steinway the PT Steinway is not usable or desirable.  Why would I use a Steinway that doesn't sound like a Steinway when I can use one that actually does?  I understand that others may have a preference for a type of sound that PT has for their specific genre, but essentially what I'm saying is that as someone interested in classical music there is no need to have a Steinway that sounds synthetic when you have one that does not.

I know the VSL is sampled, but I'm a believer - I think modelling can pull it off eventually.  But it's not there yet.

Lol, not necessarily!  I can only speak about what I know, perhaps it's not good enough for pop pianists, I have no idea.

Eh eh..

The VSL is a rather heavyweight package and uses a hideous hardware key (31+ years in IT and I personally would ban hardware keys as a disaster waiting to happen).

My impression of the VSL D-274 demos is that with tweaking PTQ's D can get the same effect.  Most of it sounds like reverb and equalization to me, maybe microphone placement as well.  Try using an impulse reverb (like "Berliner Hall" from the set I use) and adding an equ3 - presence setting to get somewhere near those demos on Vienna's website.  I don't particularly like that VSL demo sound (sounds like it's recording from a mile away) and prefer PTQ's less intrusive reverb, but YMMV.

But with something like VSL or another sample lib I'm in resource hungry territory.  With PTQ I've a small, resource friendly, easily installed package that provides me with lots of instruments and endless variations and customization options.  The historic pianos are a joy in themselves and I don't have to carry a laptop costing more than any car I ever owned to access everything - all the piano models and customization.

If you want perfect, buy a real piano, because ultimately nothing else will replace it.  For me Modartt have done a superb job and the Steinway D is the model I tend to return to (still eyeing that Steingraeber mind you).  I can't imagine the VSL lib giving me any better experience, but it's as personal thing.  It may work for some people, just not for me.

have you tried playing with the condition slider and using the randomization function of the various parameters?  i find this goes a long way toward producing a convincing (and far less synthetic) sound.  if you're running with "perfect" settings then, yeah, it's going to sound artificial.  also, maxing-out the polyphony setting (ie to 256) and using as high a bit-rate as possible will contribute much to making for a natural sound; but more than anything, it's the introduction of "imperfections" to the model that, when artfully applied, can give it a truly life-like nuance and character...
cheers,
dj

Yes, I have, quite a bit.  Not for some time now because my hard drive crashed, but yes, I have literally gone through all the settings a number of times to the point of exasperation (in both 5 and 6 and a bit in 6.1).  What I look forward to, once I get a new machine up and running, is essentially attempting to match the PT with the VSL to see precisely what it is that is generating my feeling of PT's synthetic sound to see if I can bridge the gap.  I want PT to be it, frankly, and am disappointed that it isn't there yet.  I wasn't thrilled to find myself buying another piano library.  I feel like it is very close, but that they're missing something basic but essential that, once they get it, may unlock the whole shenanigan.  But I think we're in really nuanced territory now and the thinking needs to be more nuanced as well.  I've read some messages elsewhere on the forum talking about "hearing the wood more".  I understand what they are saying, and am not terribly convinced by the responses they were given.

as you suggest, the difference between using digitally modeled instruments for recording purposes vs live performance is indeed vast, but perhaps the latter is not entirely impossible to bridge with the use of transducer-soundboard arrangements (though requiring customization of the modelling engine beyond what we currently have publicly available)... so i guess, as with most any audiophile pursuit, so much of the endeavor's "success" lies in the hardware being used (ie the amount of money thrown toward it).  beyond that though, i'd further offer that while a good real-world steinway can (perhaps) make a poor player sound pleasant enough (as the old adage goes), conversely a good player can (perhaps) make even a poor, synthetic instrument sound still better...

Jconl,

Thanks for your response and I see where you're coming from (I think :-) ).

First suggestion : start a different thread about this particular subject.  This discussion is a little different from the intent of the original poster of this thread, I think.

Í'm not sure you're going about the issue the right way.  You're talking about producing a more realistic Steinway D-274 for Pianoteq, but you're comparing it with a different virtual instrument, not a real piano.

This seems basically flawed as a way of achieving your goal.  I'm 100% with you on improving the piano sounds from Pianoteq, it's just that when you use another fake sound to compare it, then you're really not using a good target, IMO.  You're automatically open to arguments about processing differences, rather than the practicalities of the differences as you perceive them.

Your response (above) to Bruno's post was more useful, IMO.  You're making specific statements about what the difference is.  I'm not sure I agree with the comments about the trills and runs, but that's not the point.  The point is you're identifying what you want and you feel PTQ doesn't deliver, which is far more useful than comparing it to VSL, IMO.  It certainly does improve the odds of someone making a better FXP (don't look at me - I get lost in the "parameter soup" too quickly :-) ).  More detail is always better.  I'd prefer comparison to real piano recordings - it just makes more sense to me, but YMMV of course.

Alas you've cursed us all now to trying a dozen FXP's and tweaking them to try and reproduce those VSL samples.  :-)

True!

I'd love to, Bruno, but I recently had a hard drive crash and don't yet have a new machine.  Once I am back up and running perhaps I will revisit this thread, my intention would be to match the two as closely as possible.  I'd actually love to see the process unfold and prove myself wrong with being able to accurately match the PT to the VSL close-mic'd and dry.  It very well may be possible, but I find it unlikely.  There are a number of examples of people using the PT Steinway D and they all sound synthetic to me, especially when the instrument is exposed and not being drowned out in reverberation or obscured by music with thick piano textures or other instruments.

For me two or three voice close-mic'd and dry Bach is what a virtual piano lives and dies by, and so far they all have died.  That's where you find out how attentive the instrument modeler was, and that's where most of the faults and shortcomings in detail are exposed, especially if played with ornaments.  The VSL being a *possible* exception, but I don't know as of yet, unfortunately.

I'm so happy with the beautifulissimo Ant Petrof, and the K2..
Maybe I would better like if people at Modartt to add a few controls to the engine..
Musician noises, as on some other engines.. Maybe..
And more probably, I would like to hear the pedals hitting the wood, when we release them.. Not only this "pssshhht", but alos the "knock"..
The only example I can think of  is in Rickie Lee Jones album "The Magazine"..
https://www.youtube.com/watch?v=UBAAeBeMepg
Even though still faint..

And of course, I would like emulations of Dulcitone and Unacorda..

Did my own comparison with twl's post (thanks for posting it!) and the Pianoteq portion of bms earlier rendering:

https://www.forum-pianoteq.com/uploads....20only.mp3

At 1:21 PT sounds almost cartoonish to me.  The trills really bring the deficit out imo.

Lol!  Quite!

Thanks for the links to the American and Hamburg Steinways.  The American in the youtube video you sent sounded brighter, indeed.  I had heard otherwise, most likely it's determined by a variety of things.

Sorry, but I think the comparison Bach demo I posted says it all.  PT doesn't cut it.  I'm willing to be proven wrong, but I've not heard anything yet that convinces me.

It's their loss more than it is mine.  There are other options.

The purpose of all this being - in the first place - to express that maybe before PT releases more libraries they should get their engine right.

Alan stivell's harp has been physically analyzed and digitized.
Wouldn't it be interesting to model the sound for a new Pianoteq instrument?

https://www.20minutes.fr/arts-stars/cul...HPyTnZykFU

Hi Professor Leandro Duarte,

At the moment I don't have a machine, therefore no comparisons are happening here.  I spent a great deal of time before my machine went down getting PT 6/6.1 sounding acceptable and I never got there (I was also attempting additional case reverberation with convolution reverb's after adjusting the settings in PT - again I never got there).  I listen to all of PT's samples and the majority of them I find easily identifiable as synthetic within a few seconds of hearing them.  It seems really obvious to me, and here we are in PT 6 and it's still so.  I bought the VSL because it was recently on deal and a close-mic'd, trilled sample someone posted for me to listen to convinced me.  For me, so far, the VSL is the most satisfactory Steinway library I've heard (from what I've heard).

Not that VSL is flawless by any means, their solo string libraries also sound completely fake to my ears within very little listening time.  But every piano library I've heard or used before fails the dry-and-close Bach test.  VSL may as well, but it hasn't so far.  Everything I've heard here in terms of samples has only convinced me more.

...oh, sorry, if you meant the comparison I posted earlier it was the VSL twl posted compared with the PT audio file bm posted.  I chopped them up with audacity to get a side by side.

Jconl,

I agree with you.

I confess that I already liked Steinway D more, for me it was already an irreplaceable piano. However, currently, I prefer K2, because it sounds better, lighter, has more sustain and a clearer and more defined harmonic series.

However, the best Steinway preset on Pianoteq is the "player wide". It has all 5 microphones on and the entire effects palette tuned to provide the best possible timbre. So if you love a Steinway and this player wide does not please you, Pianoteq is not for you, stick with VSI.

Very good!

These are comparisons like this that make me want Modartt to create new pianos like K2, which are not submissive to existing models, but which are original, authentic in themselves, like the surprising K2.

Thanks, Bruno, for the immense work you had to give us with these comparisons!

My word, it seems, you’ve definitely gotten off topic!

Maybe I should re-phrase and say that I'd enjoy having a Pianoteq instrument which sounds like the instrument sampled for the Estate Grand. I'm not crazy about sampled pianos, either, I find them very disappointing.

Ditto.  Bring on the big Baldwin SD-10

Lanny

I believe it would have been a review of the original Pianoteq release:

https://www.soundonsound.com/reviews/modartt-pianoteq

" recordings of a Fazioli F212 (a seven-foot Italian grand retailing at a sickening $90,000) were used as part of the analysis for the modelling of Grand C1"

Well, I said in beginning of this thread 25.1 that my feelings go more to Fazioli. If the next Ptq would be a Fazioli, then the circle is complete, they are back where they started. If so, it’s fantastic. I listened to Fazioli 278 on youtube and, wow. And the Fazioli 308 - very powerful bass strings and clean sound. It is an Italian Ferrari of Grand pianos. That’s what I think.