256 Polyphony

Thanks Amen Ptah Ra for your reply. I think as you say, 'remain optimistic' for future developments in Pianoteq. After all Modartt have worked on all areas with this software and even improving aspects that I for one didn't even give thought to. Funny just as I'm writing I've noticed 6.5.2 released!!! Looks like a good day! again thanks Amen.....you and others have put some work into this forum and hopefully I will now contribute.

Thanks Key Fumbler...I agree that 256 is in itself sufficient(ish) but since 6.5.0 anything above 128 is difficult on my new HP OMEN i5 PC  (this is the best I could buy so won't be upgrading anytime soon). Version 6.4.1 however was very good at around 200 polyphony and it did feel 'light', giving an infinite polyphonic 'feel' (at least the elusion anyway). It could be me becoming used to my initial (new) PC increased performance but the polyphony readout and slightly lower performance index in 6.5.0 would say it is not all in my mind (placebo). So I suppose my question on 256+ polyphony is really just academic been nowhere near that at the moment. I think the CPU requirement is the bigger issue for me if I'm honest so it seems I've gone of topic on my own question? Thanks again for your reply

Key Fumbler...It's a desktop. i5 (8th gen) 8 DDR4 RAM, hex core, 128 SSD, 1TB HHD. I also found (off subject a bit but you might find interesting) ASIO4ALL using the line out from PC to a Yamaha DSP-A2070 (350 watt AV amp) gave a slightly better performance index and less jitter than my USB3 audio interface. Also using headphones (AKG702) through the Yamaha amp is great, as much volume as I want and the bass extention switch is like a high end sub woofer for headphones.

Key Fumbler...intel (R) Core (TM) i5-8400 CPU @ 2.80 GHZ, sockets 1, cores 6, logical processors 6.... My knowledge here is quite limited.

Thanks!

Great stuff and a brilliant vid. Thanks

Lanny, you funny guy, you have a definite sense of humor.  I like your word play, whether or not intentional!

In the trade off between polyphony and latency I doubt that there is any value in latency lower than that of a wooden piano.
I don't know what sort of timescale escapement movements and hammer flight times are on, but I suspect that it is significantly slower than the electronic delays that are discussed here.
Even after the hammer arrives at the string there is a finite time before the string develops resonance sufficient to excite the soundboard, etc.

Anyone have the physical timing of a Grand ?, say from key hitting key bed bottom at ffff to first audible complete wave ?
tens or hundreds of milliseconds ?

Not aware of any solid data on it aandrmusic, but a certain amount of latnecy is a definite plus to me, to make the digital keyboard feel more like a grand piano.

Anything over 5ms to even 15ms latency (and adjust settings to achieve it) from time to time.

Some pieces with certain repetitive physical rhythmic input seem better with more latency, than if I'm set up for lows around 1.3ms - some space in the air between muscles and brain/nerves or however best to describe it, is mostly OK and feels more like when at a real sizeable grand piano, to me.

I'm sure small/upright pianos with lightning action would better be simulated with lower latency but depends on the piano - I'm sure some large grands are faster than others too.

Mostly, with a large grand there can feel like that + amount of latency is not so much just keys, but the fact the strings are some feet away and we hear sound coming back at us from 2.7meters away (what I think of as the player bubble). It's analog, psychological and some physics which gets real for me over 5ms.

I suppose around 8ms or 10ms I'd be happy with in terms of realism for most things.

@MeDorian yes, definitely seems sympathetic resonance adds to polyphony. Using optimistic settings, it can really jump esp. with like you say a strong accent or some ballistic chords.

You must be an advanced player managing that level of polyphony, if that was a piece of music, rather than just going at it to get the most polyphony

As for the performance index that is dynamic so I guess it will depend what else your computer is doing at the time with the CPU. In other words no cause for concern when it fluctuates by modest percentages.

I don't know how the polyphony works with regard to what some companies call infinite, but it won't actually be infinite because piano strings do not resonate forever.

I guess you could record the sustain of each and every note on the keyboard (per piano model, because it probably varies between them too) and work out just how many notes would be required to sound on full sustain simultaneously  to work out if it can be truly regarded as infinite, or unlimited (whatever marketing word they choose).
I wonder what the maximum polyphony is possible in a real piece of music, IOW do we need infinite polyphony ever in practice anyway?
Also given the masking effects of the loudest keys (most recent notes) would our hearing even register the dying echoes of the quietest keys even as a collective sound?

Another consideration is the question that potentially a digital piano may even have the advantage of sounding cleaner than the real thing because a model can potentially have greater control when a note starts and stops than a real imperfect instrument with all it's mechanical physical failings!
Perhaps when Pianoteq is indistinguishable from the real thing all the time on every model then they could even look into models that improve on real pianos by deliberately avoiding undesirable resonances.
So we could have the option of real piano sound or musically improved cleaner versions of the sound. Such heresy! 
Actually I imagine many people have long sold on their more modest real world grand pianos because it couldn't compete with Pianoteq which presents piano sound optimally.

Yeah, I really don't believe that infinite polyphony is useful for anything other than specsmanship.

Now my advice to you, MeDorian, as a newbie, to this forum is:

1. Raise your awareness.

2. Consider your source.

3. Break contracts.

4. Be specific.

5. Look out for cyber bullying.

6. Remain vigilant.

As you enter any form of music business, even a forum about harmless musical instruments and virtual instrument software, it is today as though you stepped seemingly into bloody shark infested waters.

Let me caution you against some who just like to subvert!

Once you’re targeted, your inquiry quickly goes from a bodacious, “Shouldn't we be seeing/hearing 500+ polyphonic instruments these days?” to a whimper “'Do we need infinite polyphony in practice anyway?”

Have you forgotten the terms Allegro and Bravado appearing in the scores of some of the greats?

I’ve included a video where a speaker is entirely enthusiastic about the possibilities presented by unlimited polyphony:
https://youtu.be/a6nB8nmkVhI

Undoubtedly, Specs, is a real word.  If you will, just have a look at some of these: https://www.roland.com/global/products/...ies/lx708/

See: Roland Corporation, referenced.

Well old timer, someway you are now in for a treat!

Accordingly, to Roland Corporation a distinction in fact does exist between 250 maximum polyphony and limitless polyphony.  It for musical applications occurs inside Roland LX700 Digital Pianos.  The series boasts two (2) processors, one for each of its modeled piano instruments, one European grand, and one American.

Its specs are available from the Roland website linked (above):

Within your PIANOTEQ interface, as you cursory click its Options button, you’ll see Polyphony selectable as a Maximum polyphony.  Notice: the word, Maximum capitalized?

Edit: I maybe will change, “Notice: the word, Maximum capitalized?” to “Do you notice: the word, Maximum capitalized?” if it is continually a cause for some misunderstanding.

My posts are self explanatory, as I do post those as I like!

Have you a specific question you like to raise?  Your usage of these is vague to me.

Perhaps you like to rephrase your question.

Amen Ptah Ra:

You might want to take a look at the discussion pertaining to polyphony (Roland and Pianoteq in particular) at http://forum.pianoworld.com/ubbthreads....19/1.html. It goes on for, at the time of this posting, 3 pages. From a technical perspective, "limitless" is not possible, nor even particularly meaningful. "Infinity" (implicit in the term "limitless") is a concept that may have meaning only in the realm of mathematics. As to whether Pianoteq is indeed limited to 256 voice polyphony, that may or may not be true. The Pianoteq folks may well have set the limit in the GUI to 256 based upon that being an amount likely to cover most piano music playing. It clearly would not be adequate to cover the many layers that can be assembled with a DAW. The models Pianoteq uses may not have such a polyphony limit. The fact that the polyphony may be set to a very low number in Pianoteq lends credence to the notion that the major concern with polyphony from their perspective pertained to such issues as latency, directly tied in to computer capabilities.

I think there are also realistic limits to how many "notes" wooden pianos can (continue to) respond to.
The sound board is, after all, a board of wood.

I speak only for myself, but when I count up the number of "notes" in a measure (of the literature that *I* play) the number is quite small.
I don't play many 1/64th notes, 1/32nd., or even 1/16th.notes for that matter.
Then if I look across a few bars ....there really aren't many that I want still sounding at the same time - i.e. I lift the pedal if I want it to sound at all "musical".

{Forgotten quote about the importance of the space BETWEEN the notes }

I take it as a given that an implicit goal of Pianoteq is to emulate wooden pianos and PART of that emulation includes some of the limitations (often described as "character") of those pianos.  Again , a Soundboard is a board.

If I wanted a mere MIDI mill I would be looking for one, it would be something very different to Pianoteq.

===================
Beware spec wars.
A lot of things digital follow Moore's Law and that leads marketeers to boast another power of two every couple of years.
Many times that "improvement" is meaningless.
It seems no small coincidence that polyphony is often spec'd as some power of two.

The polyphony is based on the piano model (main note and any sympathetic resonances that have been modelled within the limitations of an average home computer's CPU capabilities in real-time) not the external microphone models.  Hence it is unsurprising that the computer continues to model the piano if you have not told it not to do so.
At least that's my layman's understanding.

I truly am at a lost for words.

I don't see Pianoteq having a problem with polyphony but rather wanted to know if 256 was sufficient for an extreme scenario like a piano duet or super heavy bass tremolo stuff. Having given this more thought and with the forum input I now think the 128 polyphony 'area' is where polyphony seems to peak in most heavy pieces, but with a 256 setting for 'polyphonic headroom' (only term I can think of here). This though would leave me with a question, 'Is 128 polyphony (general peak playing area) with a 256 setting for headroom sufficient to give an 'infinite polyphonic feel'? The words 'infinite feel' might be better as 'unrestricted polyphonic feel' or something but I was trying to emphasise the point.

Although I am a Sci-Fi fan I believe the above to be more in the real world and actually this is how my PC and software are presently performing.

Heya sandalholme,

my machine specs seem similar - when I'm not requiring less than 10ms latency (focused playing/recording) I will set the buffer to 1024. That allows a great deal more room to continue mixing tracks, adding plugins etc.

I find that no matter the Pianoteq piano/preset, complex Liszt is a great benchmark MIDI playback test

Then if I want to re-focus on recording a certain new track or whatever else, I'll set the buffer back to 512 for the lower latency. When done, back to 1024 to avoid Liszt type complexity from causing any pops and clicks.

Some day, I'll update hardware at home but hopefully, we can still enjoy the C. Bechstein DG without that for now.

Sincerely hope that is the key to your dilemma. Makes me to think you rule that piano out - but would make for a big if that solves it.

Ah, I see now - just realising, with apologies, you're speaking of playing Liszt (beyond my abilities), where I was assuming playing back MIDI was at issue - projecting much, I suppose there

A buffer of 1024 would likely be waaay too much for our hardware specs for playing purposes - certainly fine for glitch free playback of Liszt MIDI though.

That said, definitely do still suggest trying out a click higher on buffer, perhaps from 512 up to 576 (bringing 12ms here)?

Would be interested if it solves that particular passage's difficulties. I can enjoy higher latency, between 10.7 to 15ms for some physical rhythmic playing but with Liszt?? Well, not so sure higher latency would be quite the golden bullet Hoping though, of course.

Yes indeed, the Steingraeber is such a extraordinary instrument. I like your description of a gold sound. I feel it's descriptive and understood in few words. Would love to play an acoustic one some day - and the Grotrian too, also much loved.

Wishing the best, happy Liszt-ing sandalholm!