I just found out the my mini form PC with a i3-4160T
canNOT support the K2 prelude
but can support the layered Steinway NY/HB (left/right)
I'm very surprised... I didn't know.
My first thought was "maybe the K2 prelude has more mics".. but no.
Interesting, on a fast sys, in the performance tab, the K2 seems lighter on resources than a preset with 2 different layered Steinway D pianos (NY and HB).
My first thought was "maybe the K2 prelude has more mics".. but no.
Interesting, on a fast sys, in the performance tab, the K2 seems lighter on resources than a preset with 2 different layered Steinway D pianos (NY and HB).
It's even worst with the C. Bechstein!
This makes no sense I think I could use the K2 and the Bechstein on a Celeron J3450 (probably with different settings.).
PS - I just ordered an used i7-6700T... there goes my PT-pro upgrade budget...
I used to run PT on a odroid XU4q (ARM based) and I remember the historical pianos used to run easier than the modern ones.
I used to run PT on a odroid XU4q (ARM based) and I remember the historical pianos used to run easier than the modern ones.
... there must be something else wrong with my system...
Edit: I found out what it is:
I use internal sample rate of 96kHz with a buffer size of 512 samples which gives a latency of 5.3ms
greater latency confuses me - of course, other parts of the system will add to the latency.
when I tried to reduce the latency with 48kHz I found the sound to be different (read: bad)
I just tried the C.Bechstein with 48kHz and it's usable. Why the layered Steinways are less demanding on the system I do not know.
192kHz was one of the reasons I wanted to try the Pro version of Pianoteq.
(thank you "found my laptop to be underpowered [...]" thread!)
You're welcome - Glad you found something to adjust
At 192kHz I don't hear much diff from 96 or even 48. I sense it's better at times, in certain ways - but also fail too often when trying to A/B test this. I mainly use 96 or higher for recording/processing but eventually downsample.. so still to today I'm usually at 48kHz.
I guess some audio units may have different quality applied to the DAC processing at specific different rates - maybe?? (haven't considered it before - and would be interested in finding if that's common.. like having 1 expensive patented process for the high spec, but the cheapest DAC IP for lower rates.)
Good luck with you new setup.
@Marcos - absolutely love playing the older pianos - they seem lighter on resources for sure - and still improve with the engine - def. worth checking out if a system is low spec.
At 192kHz I don't hear much diff from 96 or even 48. I sense it's better at times, in certain ways - but also fail too often when trying to A/B test this. I mainly use 96 or higher for recording/processing but eventually downsample.. so still to today I'm usually at 48kHz.
Neither do I. In fact I think most of us cannot hear above 16 kHz, so according to Nyquist, 32 kHz should be enough. I also bought the pro version looking at the 192 kHz but at the end the most important thing about it is the extra parameters available to tweak.
Qexl wrote:At 192kHz I don't hear much diff from 96 or even 48. I sense it's better at times, in certain ways - but also fail too often when trying to A/B test this. I mainly use 96 or higher for recording/processing but eventually downsample.. so still to today I'm usually at 48kHz.
Neither do I. In fact I think most of us cannot hear above 16 kHz, so according to Nyquist, 32 kHz should be enough. I also bought the pro version looking at the 192 kHz but at the end the most important thing about it is the extra parameters available to tweak.
I mean 192 kHz internally just to increase the number of samples without increasing the latency...
That's how I use 96 kHz - I don't even know if it works that way - but there are other features that I might also use.
Good luck finding the cause.
It is a bit odd to title a thread "Performance for lower systems" then to start extolling the dubious benefits of going beyond 48khz, something that for the most part needlessly gobbles up a lot of extra CPU power.
I would instead concentrate on finding out why at your system settings Pianoteq does not sound right at 48khz.
The Nyquist frequency of 48khz is 24khz. Comfortably way, way above any meaningful output of a piano that humans can genuinely hear, and above the maximum frequency that can be perceived by a newborn baby in tip top health - and the perception of those super high frequencies falls off pretty quickly anyway.
Perhaps we have miniscule amounts of frequencies that can possibly just about be felt rather than actually heard at the extreme ends of the spectrum. Really nothing to do with music or tone.
I don't think Pianoteq has audible aliasing when using the instruments normal frequency range.
I would test the Windows Audio and ASIO drivers with different buffer settings. "Exclusive mode" and low "latency mode"
Have a look to see if there are clashing settings in the audio settings in the operating system that might cause any audible distortion.
Good luck finding the cause.
It is a bit odd to title a thread "Performance for lower systems" then to start extolling the dubious benefits of going beyond 48khz,
[...]
I would test the Windows Audio and ASIO drivers with different buffer settings. "Exclusive mode" and low "latency mode"
[...]
You're right! - I just try desperately to find a difference between 96/512 and 48/256 and couldn't really tell. (I can't do a blind test)
Seems obviously that I'm wrong about this and I'm glad I posted and got help from the forum!
BTW changing enough times from 48/96, 512/256 Pianoteq v7.5.4 will break on my system, I don't know what part of the system breaks but restating PT fixes it...
My system changed from when I last "figured" all this out:
- faster machine (i3T vs Celeron)
- PT 6 vs 7
- faster MIDI to USB interface
- different audio drivers on Linux
Key Fumbler wrote:Good luck finding the cause.
It is a bit odd to title a thread "Performance for lower systems" then to start extolling the dubious benefits of going beyond 48khz,
[...]
I would test the Windows Audio and ASIO drivers with different buffer settings. "Exclusive mode" and low "latency mode"
[...]You're right! - I just try desperately to find a difference between 96/512 and 48/256 and couldn't really tell. (I can't do a blind test)
Seems obviously that I'm wrong about this and I'm glad I posted and got help from the forum!BTW changing enough times from 48/96, 512/256 Pianoteq v7.5.4 will break on my system, I don't know what part of the system breaks but restating PT fixes it...
My system changed from when I last "figured" all this out:
- faster machine (i3T vs Celeron)
- PT 6 vs 7
- faster MIDI to USB interface
- different audio drivers on Linux
Glad that helped.
It sounds like your driver stability sounds a bit shaky. Maybe an update will help?
It is a bit odd to title a thread "Performance for lower systems".
I don't remember exactly but I'm sure I meant title to be "Performance of lower end systems".
(I knew 'lower' was wrong but couldn't remember the exact expression "lower end"
the "of" / "for" is an error I make often with small English words - they all look the same to me...!)
and this just became very relevant as v8 seems more demanding on the CPU!!! ??? I'll start another thread if I can confirm that...
I'm getting a new i7 core tomorrow - should have ordered an i9??? to be determined...
Key Fumbler wrote:It is a bit odd to title a thread "Performance for lower systems".
I don't remember exactly but I'm sure I meant title to be "Performance of lower end systems".
(I knew 'lower' was wrong but couldn't remember the exact expression "lower end"
the "of" / "for" is an error I make often with small English words - they all look the same to me...!)and this just became very relevant as v8 seems more demanding on the CPU!!! ??? I'll start another thread if I can confirm that...
I'm getting a new i7 core tomorrow - should have ordered an i9??? to be determined...
No, a new i7 is going to chew up multiple simultaneous instances with no problems. Absolutely no need for an i9!
If planning on running a single instance without silly latency demands a user can get away with a lot less. i5, i3, Celeron and lower besides with performance concessions.
This from the FAQ:
What kind of CPU (processor) do I need?
Pianoteq is CPU intensive software for it computes the sound in realtime. Many modern computers already offer a fully sufficient CPU. We recommend any multi-core CPU, listed on this page: https://www.cpubenchmark.net/high_end_cpus.html
By restricting the internal sample rate in the Options menu, you can work with less powerful CPU's. The Linux version of Pianoteq 7 and later lets you work with ARM RISC based CPU's used in smaller Linux based consumer electronic devices.
No, a new i7 is going to chew up multiple simultaneous instances with no problems. Absolutely no need for an i9!
well - we can agree that I've been very confused with the performance issues...
I start a new thread if I have problems...
so far seems to me that the performance took a VERY bit hit from v7 to v8
v8 sound seems better - I don't think I'll need the layering to balance lower and high registers to my liking and audio system...
My i7 it not new:
i7-6700T
!! Status: Discontinued !!
Launch Date Q3'15 !!!
and is the "T" version: 35W (so it can be cooled silently) not the full power version...
and it's not arriving tomorrow (miss read the tracking of the parcel...)
it is on the list!
edit: borderlines unusable to me in the same machine that was fine for version 7
That CPU should have no problems. On paper slightly faster than mine.