Reply by miiindbullets

  • Avatar
  • miiindbullets
  • Member
  • Offline
  • From: Idaho, USA
  • Registered: 31-03-2022
  • Posts: 251

Re: Pianoteq 9 When and What

Beto-Music wrote:

Please vote:

Are you in favor of a quite more CPU demandaing version variation of pianoteq, to get more sound details/texture, or do you prefer CPU demanding to keep around average CPU personal computer.

1-Yes, I prefer a more CPU intensive.
2-No, I want CPU demands to stay around average personal computer.

How about a slider for performance vs. fidelity? I'm not sure if this sort of thing would be technically feasible, but I'd be interested in something like that if it is. It would be nice to find a sweet spot for real-time playing on whatever hardware we each happen to be running it.

YouTube

miiindbullets's Website

Reply by stuart_tetley

  • Avatar
  • stuart_tetley
  • Member
  • Offline
  • Registered: 21-09-2024
  • Posts: 1

Re: Pianoteq 9 When and What

Beto-Music wrote:

Please vote:

Are you in favor of a quite more CPU demandaing version variation of pianoteq, to get more sound details/texture, or do you prefer CPU demanding to keep around average CPU personal computer.

1-Yes, I prefer a more CPU intensive.
2-No, I want CPU demands to stay around average personal computer.

I'm fairly new here.  First post in fact.  I've been testing Pianoteq for a while on various hardware, ranging from at the lower end Raspberry Pi 4, 2012 Mac mini, iPad Air 2 to higher end A14Z & M1 iPad Pro,  M1 Max MacBook Pro.  I took the plunge during the recent sale, and am very happy with how that turned out.

A few weeks ago, I would have voted Yes, but now it's definitely No for me.  Pianoteq already appears to scale very well across hardware capabilities.  Sure, some glitches in the lower end devices, but I see this as something Modartt can (and probably are) improving alongside the piano models themselves.  And maybe there already is a CPU intensive switch in the form of the two polyphony settings: Auto (Optimistic) and Auto (Pessimistic).

For new piano models, Fazioli F308 first for me, and maybe Bôsendorfer 290 Imperial alongside the 280VC. 

Also, for iPad version, which is amazing, proper portrait mode support would be really nice.  Sure, you can kind of 'fake' portrait mode via Stage Manager, but proper support would really be the icing on the cake.  It would also remove the need for many of the scrolling widgets.  An option to 'lock landscape' maybe needed for those who prefer to keep it that way.

All v best

Stuart

Reply by Beto-Music

  • Avatar
  • Beto-Music
  • Member
  • Offline
  • Registered: 02-05-2007
  • Posts: 2,776

Re: Pianoteq 9 When and What

It could always have different option in terms of cpu use.

The first problem is human brain always get used much easier with better quality than the oposity. So, after try an "turbo version" that demand 3x more cpu, for example, with more quality, many people could get problems to accept the standart quality fit for a laptop.

The second problem is about create cpu demanding version, computing more elements to render more textures and get more natural tones. I don't know how much work and money this would require to reengineering of the algoritims.

ACKeys wrote:

For the sake of being a more portable solution, I would argue for #2.  Not very many other solutions can be loaded onto a modest laptop.

Reply by Beto-Music

  • Avatar
  • Beto-Music
  • Member
  • Offline
  • Registered: 02-05-2007
  • Posts: 2,776

Re: Pianoteq 9 When and What

Hi
No we are not talking about CPU use by poliphony.
We are talking about the details of piano physics computed by pianoteq algorithm.
Think on pianoteq like a CGI computer model of a dinossaur. The more details th3 computer model have for skin texture, eyes, teeth, tail, nails, the more natural the final CGI animation will look.

stuart_tetley wrote:

And maybe there already is a CPU intensive switch in the form of the two polyphony settings: Auto (Optimistic) and Auto (Pessimistic).

Stuart

Last edited by Beto-Music (22-09-2024 00:28)

Reply by Irmin

  • Avatar
  • Irmin
  • Member
  • Offline
  • From: Second star to the right
  • Registered: 20-04-2012
  • Posts: 97

Re: Pianoteq 9 When and What

stuart_tetley wrote:

A few weeks ago, I would have voted Yes, but now it's definitely No for me.  Pianoteq already appears to scale very well across hardware capabilities.  Sure, some glitches in the lower end devices, but I see this as something Modartt can (and probably are) improving alongside the piano models themselves.  And maybe there already is a CPU intensive switch in the form of the two polyphony settings: Auto (Optimistic) and Auto (Pessimistic).

I don't think Pianoteq scales at all. The engine runs on the lowest common denominator. (ignoring polyphony) You essentially get the same quality on a dog slow cpu e.g. Pi 4 as an intel 14900k.

Today, even an entry level to mid range laptop CPU/iphone CPU is probably 100's maybe 1000's of times faster than the PI 4 cpu. Why not tweak that model for that level of performance. Perhaps a new model for current performance and a legacy pianoteq for people who want to run on PI's etc.

Reply by Key Fumbler

  • Avatar
  • Key Fumbler
  • Member
  • Offline
  • Registered: 20-10-2018
  • Posts: 1,514

Re: Pianoteq 9 When and What

Irmin wrote:
stuart_tetley wrote:

A few weeks ago, I would have voted Yes, but now it's definitely No for me.  Pianoteq already appears to scale very well across hardware capabilities.  Sure, some glitches in the lower end devices, but I see this as something Modartt can (and probably are) improving alongside the piano models themselves.  And maybe there already is a CPU intensive switch in the form of the two polyphony settings: Auto (Optimistic) and Auto (Pessimistic).

I don't think Pianoteq scales at all. The engine runs on the lowest common denominator. (ignoring polyphony) You essentially get the same quality on a dog slow cpu e.g. Pi 4 as an intel 14900k.

Today, even an entry level to mid range laptop CPU/iphone CPU is probably 100's maybe 1000's of times faster than the PI 4 cpu. Why not tweak that model for that level of performance. Perhaps a new model for current performance and a legacy pianoteq for people who want to run on PI's etc.

If Pianoteq 9 (or later) also had a greater complexity non real-time offline mode that could also work for the lower end hardware too. It would just take longer to render on slower hardware.
Pianoteq keeps getting better each generation anyway  - the sound quality difference is going to be pretty subtle despite substantial computation anyway and is probably going to be mostly lost on real time listening as we also play and get caught up with performance.
All the better potentially to leave the extra complexity model to offline rendering?

It might be interesting to see what GPU processing could bring to the table too - and that could be real-time. If feasible that's probably requiring a massive change. Who knows. I doubt either concept is even being developed for Pianoteq really, and we will be pleased as punch with the upgrades we eventually hear. It's great as it is.