Topic: U4 would be the less CPU resources demanding pianoteq piano I have!

I have a powerful laptop where I have pianoteq installed but which I and the members of my family use for multiple purposes, so I have a small oldish netbook (a Sony Vaio, two cores AMD-350 processor 1,60 Ghz with 2 GB of RAM) which stays permanently on top of my electric piano (midi keyboard) and which I ordinarily use as a sound module when playing (just because it is handy and more comfortable to have it there always connected to my keyboard).

I have been trying to play with pianoteq loaded into RealBand (from PGMusic) while backing tracks which I made with Band-in-a-Box run (are being played as an accompaniment). The accompaniments I made use realtracks (not Midi tracks, but wav tracks for each instrument). This demands more resources from the computer than if I just play the piano with pianoteq as a standalone (with no backing accompaniment).

The piano that had fewer problems to satisfactory accomplish this and did not make the CPU to overload easily came to be the U4 if loaded into Pianoteq Stage 5. When trying with Pianoteq Stage 6, the CPU overloaded earlier (more easily) when paying runs over the piano keys with the sustain pedal pressed down.

I tried with the Steinway Model B and the CPU was overloaded much earlier than with the U4 through the runs I played on the keyboard. The same with the Steinway Model D (which I played in demo mode as I don’t have it). And odd enough, I tried with some of the Kivir Proyect pianos (the Pleyel and the Erard) and the legacy pianos which come free with pianoteq, with no better results.
I did not try with the C. Bechstein DG that I also have because my netbook can’t cope with it even if being played by itself with the standalone version of pianoteq (my more powerful laptop can, of course).

I also tried a sampled piano of only one giga (I have bigger ones but I tried with a small one) loaded to a well known sampler (NI) and it was worse. The computer could not cope with the sampled piano and the Band-in-a-Box realtracks being played simultaneously.

Therefore, I am very glad I bought the U4. I wanted an upright in my collection, I love how it sounds (as an upright I mean) and it is seems to be very efficient in terms of CPU resources usage.

I just wanted to share (not much to do during our quarantine) and would appreciate any feedback and opinions regarding the subject.

Last edited by Alejandro55 (29-05-2020 12:52)

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

You can try to increase the latency setting in Pianoteq (or in the ASIO driver settings depending on which ASIO or DirectSound driver you use).  Latency is sometimes controlled by the DAW itself, if you're not running the standalone.  Often increasing latency by a few milliseconds is enough to handle the more demanding libraries in real-time with noticeably impacting playback responsiveness.  Also, keeping the pedal engaged for long periods of time can increase overhead substantially, so sometimes I use the pedal less on loud or fast passages, if I find the CPU is starting to get overloaded as I play.

I also suspect the reason that 5 is faster than 6 is that 6.1 introduced a new ambient air simulation model, which may have increased the processing load somewhat.

Spotify: https://open.spotify.com/artist/2xHiPcCsm29R12HX4eXd4J
Pianoteq Studio & Organteq
Casio GP300 & Custom organ console

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

tmyoung wrote:

You can try to increase the latency setting in Pianoteq (or in the ASIO driver settings depending on which ASIO or DirectSound driver you use).  Latency is sometimes controlled by the DAW itself, if you're not running the standalone.  Often increasing latency by a few milliseconds is enough to handle the more demanding libraries in real-time with noticeably impacting playback responsiveness.  Also, keeping the pedal engaged for long periods of time can increase overhead substantially, so sometimes I use the pedal less on loud or fast passages, if I find the CPU is starting to get overloaded as I play.

I also suspect the reason that 5 is faster than 6 is that 6.1 introduced a new ambient air simulation model, which may have increased the processing load somewhat.

Thank you tmyoung for your input.

I use ASIO4ALL as my sound driver.

I have the ASIO buffer size set to 512 for RealBand, which might be a reasonable setting. Anyway, as it is I can play the U4 simultaneously with the realtracks for the accompaniment with no conflicts, so I am happy with that.

The runs I played with the sustain pedal pressed down were to allow me to detect and foresee which virtual pianos were less demanding of CPU resources so as to use those in order to minimize conflicts (the ones I was facing with some pianos and possible ones) while playing along with my backing tracks. According to my findings, when using my oldish netbook (not my notebook), some pianos end up overloading the system if playing a run up and down the keyboard that I would not normally do, while others would overload the system when playing a run that I might do when playing. In the way I have the ASIO buffer set for RealBand, when playing the U4 simultaneously with the realtrack, the system would finally be overloaded if I play an extra long run (that is, a nonstop run until the system becomes overloaded), but not if I play as I normally do.

Nevertheless, your recommendation was useful.
For my netbook, and for the standalone version of Pianoteq 5, I have the ASIO Buffer size set to 256 samples, and with that setting I do not have nor had problems to play the U4 or the Steinway Model B. But I was having problems with Pianotq 6 if playing the C. Bechstein C, so following your recommendation, I pushed the buffer size up from 256 samples to 384 and “vualá”, now I can play the Bechstein too, without me perceiving any undesirable nor unbearable latency (maybe someone with better hearing than mine might, but not me).

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

P.s.
I have just tried and found out that if within RealBand, in the audio settings, I lower down the sample rate from 44,1 Hz to 22,05 Hz while keeping the ASIO buffer size in 512 samples, I am able not only to play the U4 simultaneously with the realtracks loaded into the program, but also the Steinway Model B if loading Pianoteq 6 (not just Pianoteq 5 as before) into a midi track in order to play along with it. I can also play the C Bechstein DG, but with this one I get crackles if playing too many (well, not that many) notes with the sustain pedal pressed down.

I guess that when doing what I mention I ought to be lowering down the audio quality, but well, If doing so, I can choose which piano to play and not get stuck with the one (not that there is anything wrong with this, as if I had a real upright piano, that would be the only one I could play as well -with only one preset to choose from-).

Regarding the acoustic Pianoteq pianos that I have and their CPU resources usage, it came apparent to me that the U4 is the one that consumes less, then the Steinway Model B follows, and the C Bechstein DG would be the most demanding.

Last edited by Alejandro55 (29-05-2020 20:06)

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

Alejandro55 wrote:

I have a powerful laptop where I have pianoteq installed but which I and the members of my family use for multiple purposes, so I have a small oldish netbook (a Sony Vaio, two cores AMD-350 processor 1,60 Ghz with 2 GB of RAM) which stays permanently on top of my electric piano (midi keyboard) and which I ordinarily use as a sound module when playing (just because it is handy and more comfortable to have it there always connected to my keyboard).

What is your Pianoteq performance index saying? (I love that measurement, so useful!)

What operating system are you using? (How) have you performance tuned it to maximize low latency throughput? Especially if you are just using it for this single purpose, there is a lot you can do. Low latency DSP and cpu frequency scaling are not compatible with each-other. So if you can lock your cpu at the highest frequency while playing you will get greater headroom before xruns/DSP overload.

IIRC my core2duo laptop got about 25 on the index. My new i5 6265u scores much better.

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

ethanay wrote:
Alejandro55 wrote:

I have a powerful laptop where I have pianoteq installed but which I and the members of my family use for multiple purposes, so I have a small oldish netbook (a Sony Vaio, two cores AMD-350 processor 1,60 Ghz with 2 GB of RAM) which stays permanently on top of my electric piano (midi keyboard) and which I ordinarily use as a sound module when playing (just because it is handy and more comfortable to have it there always connected to my keyboard).

What is your Pianoteq performance index saying? (I love that measurement, so useful!)

What operating system are you using? (How) have you performance tuned it to maximize low latency throughput? Especially if you are just using it for this single purpose, there is a lot you can do. Low latency DSP and cpu frequency scaling are not compatible with each-other. So if you can lock your cpu at the highest frequency while playing you will get greater headroom before xruns/DSP overload.

IIRC my core2duo laptop got about 25 on the index. My new i5 6265u scores much better.

My small netbook which I leave on top of my piano keyboard and keep connected to it, is working on Windows 7 32 bits. I already described its other characteristics. CPU Frequency 1.600 Mhz.

In that computer, I tried setting the internal sample rate to 22050 Hz for both Pianoteq 5 and 6, keeping the ASIO bufer size set to the values I mentioned above (256 samples for pianoteq 5, 384 samples for version 6 and 512 samples for RealBand). Doing that, I can play every piano (normal playing) not getting the CPU overloaded (at the expense of sound quality?).
Having done that, with Pianoteq 5 I do not overload the system when playing the U4 or the Seintway Model B no matter how crazily I might try to hit keys while pressing the sustain pedal.
Different with Pianoteq 6, with which if I hit keys like mad while pressing the sustain pedal, the U4 does not collapse but the Stainway Model B does and the C Bechstein DG does it sooner. None does now if I play normally.

These are the Pianoteq performance index numbers that I read (if forcing the pianos’ polyphony to the maximum I could) that interest you: Pianoteq 5: fluctuating from 10 to 12 but reaching sometimes 14 as a maximum. Pianoteq 6: does not go over 10. Pianoteq 5 loaded into RealBand: does not go over 11.
You are obviously much more audio technical oriented than me, so you might conclude something from the numbers I gave you.

I have not problems at all with my other notebook, but that one I plug to my keyboard piano only now and again as my wife or son might want to use it while I play / practice piano. With this one I can as well load some big sampled pianos I have (the bigger being 14 GB) and can play them having no conflict and I can also play some virtual electric pianos that I have that only work on a 64 bits system. By the way, I also have both Pianoteq’s electric piano collections but although I love them, I tend to play much more using acoustic piano sounds.
This more potent notebook that I have is a Hewlett Packard Intel Core(TM) i5-5200U CPU 2.2 GHz, 8 GB RAM, with Windows 10 64 bits.

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

Alejandro55 wrote:
ethanay wrote:
Alejandro55 wrote:

I have a powerful laptop where I have pianoteq installed but which I and the members of my family use for multiple purposes, so I have a small oldish netbook (a Sony Vaio, two cores AMD-350 processor 1,60 Ghz with 2 GB of RAM) which stays permanently on top of my electric piano (midi keyboard) and which I ordinarily use as a sound module when playing (just because it is handy and more comfortable to have it there always connected to my keyboard).

What is your Pianoteq performance index saying? (I love that measurement, so useful!)

What operating system are you using? (How) have you performance tuned it to maximize low latency throughput? Especially if you are just using it for this single purpose, there is a lot you can do. Low latency DSP and cpu frequency scaling are not compatible with each-other. So if you can lock your cpu at the highest frequency while playing you will get greater headroom before xruns/DSP overload.

IIRC my core2duo laptop got about 25 on the index. My new i5 6265u scores much better.

My small netbook which I leave on top of my piano keyboard and keep connected to it, is working on Windows 7 32 bits. I already described its other characteristics. CPU Frequency 1.600 Mhz.

In that computer, I tried setting the internal sample rate to 22050 Hz for both Pianoteq 5 and 6, keeping the ASIO bufer size set to the values I mentioned above (256 samples for pianoteq 5, 384 samples for version 6 and 512 samples for RealBand). Doing that, I can play every piano (normal playing) not getting the CPU overloaded (at the expense of sound quality?).
Having done that, with Pianoteq 5 I do not overload the system when playing the U4 or the Seintway Model B no matter how crazily I might try to hit keys while pressing the sustain pedal.
Different with Pianoteq 6, with which if I hit keys like mad while pressing the sustain pedal, the U4 does not collapse but the Stainway Model B does and the C Bechstein DG does it sooner. None does now if I play normally.

These are the Pianoteq performance index numbers that I read (if forcing the pianos’ polyphony to the maximum I could) that interest you: Pianoteq 5: fluctuating from 10 to 12 but reaching sometimes 14 as a maximum. Pianoteq 6: does not go over 10. Pianoteq 5 loaded into RealBand: does not go over 11.
You are obviously much more audio technical oriented than me, so you might conclude something from the numbers I gave you.

I have not problems at all with my other notebook, but that one I plug to my keyboard piano only now and again as my wife or son might want to use it while I play / practice piano. With this one I can as well load some big sampled pianos I have (the bigger being 14 GB) and can play them having no conflict and I can also play some virtual electric pianos that I have that only work on a 64 bits system. By the way, I also have both Pianoteq’s electric piano collections but although I love them, I tend to play much more using acoustic piano sounds.
This more potent notebook that I have is a Hewlett Packard Intel Core(TM) i5-5200U CPU 2.2 GHz, 8 GB RAM, with Windows 10 64 bits.


Glad to know that.

The Eletric Pianos package also requires little CPU.

Respeito, Esforço e Sabedoria

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

Historic pianos also demand less CPU, I discovered it by accident when using a raspberry pi 3

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

Professor Leandro Duarte wrote:
Alejandro55 wrote:
ethanay wrote:

What is your Pianoteq performance index saying? (I love that measurement, so useful!)

What operating system are you using? (How) have you performance tuned it to maximize low latency throughput? Especially if you are just using it for this single purpose, there is a lot you can do. Low latency DSP and cpu frequency scaling are not compatible with each-other. So if you can lock your cpu at the highest frequency while playing you will get greater headroom before xruns/DSP overload.

IIRC my core2duo laptop got about 25 on the index. My new i5 6265u scores much better.

My small netbook which I leave on top of my piano keyboard and keep connected to it, is working on Windows 7 32 bits. I already described its other characteristics. CPU Frequency 1.600 Mhz.

In that computer, I tried setting the internal sample rate to 22050 Hz for both Pianoteq 5 and 6, keeping the ASIO bufer size set to the values I mentioned above (256 samples for pianoteq 5, 384 samples for version 6 and 512 samples for RealBand). Doing that, I can play every piano (normal playing) not getting the CPU overloaded (at the expense of sound quality?).
Having done that, with Pianoteq 5 I do not overload the system when playing the U4 or the Seintway Model B no matter how crazily I might try to hit keys while pressing the sustain pedal.
Different with Pianoteq 6, with which if I hit keys like mad while pressing the sustain pedal, the U4 does not collapse but the Stainway Model B does and the C Bechstein DG does it sooner. None does now if I play normally.

These are the Pianoteq performance index numbers that I read (if forcing the pianos’ polyphony to the maximum I could) that interest you: Pianoteq 5: fluctuating from 10 to 12 but reaching sometimes 14 as a maximum. Pianoteq 6: does not go over 10. Pianoteq 5 loaded into RealBand: does not go over 11.
You are obviously much more audio technical oriented than me, so you might conclude something from the numbers I gave you.

I have not problems at all with my other notebook, but that one I plug to my keyboard piano only now and again as my wife or son might want to use it while I play / practice piano. With this one I can as well load some big sampled pianos I have (the bigger being 14 GB) and can play them having no conflict and I can also play some virtual electric pianos that I have that only work on a 64 bits system. By the way, I also have both Pianoteq’s electric piano collections but although I love them, I tend to play much more using acoustic piano sounds.
This more potent notebook that I have is a Hewlett Packard Intel Core(TM) i5-5200U CPU 2.2 GHz, 8 GB RAM, with Windows 10 64 bits.


Glad to know that.

The Eletric Pianos package also requires little CPU.

marcos daniel wrote:

Historic pianos also demand less CPU, I discovered it by accident when using a raspberry pi 3


With Pianoteq 6 and keeping the same settings (ASIO buffer size and internal sample rate) with which I made my last trials (as mentioned in my post previous to this one), it did not seem to me that the electric pianos have a low demand on CPU resources. I can easily crush them and get my system to be overloaded if playing (fooling around I ought to say, rather than “playing”) with the intention of doing so. It didn’t (the system did not become overloaded) when playing those instruments as I would normally use them. I guess that a lot of effects (reverb, chorus, delays and maybe equalizing) might be involved to feature their sound which I expect might very demanding of CPU resources.

Regarding the Historical Pianos (the free ones at least, I haven’t tried the ones from the Karsten collection), they appear to be as Marcos Daniel said, they seem to be uncrushable (no matter what I did, the system did not become overloaded). It seems they have a much lower demand on CPU resources than the other more demanding pianos. Therefore, I could give them a try just for the sake of doing so, but I foresee that I won’t be making much use of them as it is presumable that they come free for a reason, and they would not properly suit (for being historical) the styles of music I prefer playing.

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

I can confirm that Kremsegg 1 & 2, Erard 1922 and Pleyel 1926 are less resources hungry.
Erard, Peyel and Bechstein (from Kremsegg) are modern grands, suitables for most music styles (I think). They don't sound as good as the grands you buy singly but they are nicer to me than internal sounds of my digital pianos.

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

You cannot expect miracles from a 1.6 GHz CPU.

Hard work and guts!

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

marcos daniel wrote:

I can confirm that Kremsegg 1 & 2, Erard 1922 and Pleyel 1926 are less resources hungry.
Erard, Peyel and Bechstein (from Kremsegg) are modern grands, suitables for most music styles (I think). They don't sound as good as the grands you buy singly but they are nicer to me than internal sounds of my digital pianos.

The same on my side. Any of the pianoteq’s pianos sound better than the internal sound of my digital piano.

Since I decided to learn to play the piano (already being quite a mature grown up man) and bought my digital piano, I have been in a quest to find the ultimate piano sound, something that would make me feel as if I was playing a real piano. On those days I bought a sound module, the Emu Proformance (which I don’t have any more as my son broke it) and later on a Kurzweil KMP-1 which I still have screwed underneath my digital piano and connected to it. Then came my search for a convincing virtual piano (my first better sounding one was the one from the Super Quartet dxi/vst from Edirol and afterwards the GIGA pianos for GigaSampler, but those and others following were not yet good enough).

It was not until nowadays that I came to be happy with a piano sound and feel I can finally settle down. Computers are much more powerful these days and huge advances had been reached on manipulating and featuring sound. Pianoteq sounds and feels to me quite as if I am playing a real piano and I have a few very convincing sampled pianos that sound good. And it is not that I can’t make comparisons, because I had the opportunity of playing a few fairly good sounding uprights and two nice Grand pianos while I was taking piano lesson.

EvilDragon wrote:

You cannot expect miracles from a 1.6 GHz CPU.

No, I wasn’t expecting miracles from a 1.6 Ghz CPU, and specially not for it being a netbook and not even a notebook, but got one anyway and Pianoteq helped to make the Miracle come true.

I can have that insignificant netbook (which I originally bought to be able to carry around with me for working purposes) perfectly suited to fit on top of my digital piano and always keep connected to it, allowing me to use as it a sound module and to be able to play a virtual instrument that sounds as a real piano, and even to trigger simultaneously some accompaniments to play along with.

I have several virtual pianos, but being Pianoteq the one I prefer, it happens to be the one that I can use with that small netbook. Other virtual pianos that I have and I can play using my netbook do not satisfy me. I have some big sampled pianos that sound quite good, but my netbook can’t cope with them (my other notebook which I now and again connect to my digital piano obviously does). The only smaller sampled piano that I have that my netbook can manage and sounds reasonably well (though not enough and not as good as Pianoteq) is Sampletekk Black (powered by Kontakt).

So, miracles, sometimes, might certainly happen.

Last edited by Alejandro55 (01-06-2020 17:33)

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

Alejandro55 wrote:
marcos daniel wrote:

I can confirm that Kremsegg 1 & 2, Erard 1922 and Pleyel 1926 are less resources hungry.
Erard, Peyel and Bechstein (from Kremsegg) are modern grands, suitables for most music styles (I think). They don't sound as good as the grands you buy singly but they are nicer to me than internal sounds of my digital pianos.

The same on my side. Any of the pianoteq’s pianos sound better than the internal sound of my digital piano.

Since I decided to learn to play the piano (already being quite a mature grown up man) and bought my digital piano, I have been in a quest to find the ultimate piano sound, something that would make me feel as if I was playing a real piano. On those days I bought a sound module, the Emu Proformance (which I don’t have any more as my son broke it) and later on a Kurzweil KMP-1 which I still have screwed underneath my digital piano and connected to it. Then came my search for a convincing virtual piano (my first better sounding one was the one from the Super Quartet dxi/vst from Edirol and afterwards the GIGA pianos for GigaSampler, but those and others following were not yet good enough).

It was not until nowadays that I came to be happy with a piano sound and feel I can finally settle down. Computers are much more powerful these days and huge advances had been reached on manipulating and featuring sound. Pianoteq sounds and feels to me quite as if I am playing a real piano and I have a few very convincing sampled pianos that sound good. And it is not that I can’t make comparisons, because I had the opportunity of playing a few fairly good sounding uprights and two nice Grand pianos while I was taking piano lesson.

EvilDragon wrote:

You cannot expect miracles from a 1.6 GHz CPU.

No, I wasn’t expecting miracles from a 1.6 Ghz CPU, and specially not for it being a netbook and not even a notebook, but got one anyway and Pianoteq helped to make the Miracle come true.

I can have that insignificant netbook (which I originally bought to be able to carry around with me for working purposes) perfectly suited to fit on top of my digital piano and always keep connected to it, allowing me to use as it a sound module and to be able to play a virtual instrument that sounds as a real piano, and even to trigger simultaneously some accompaniments to play along with.

I have several virtual pianos, but being Pianoteq the one I prefer, it happens to be the one that I can use with that small netbook. Other virtual pianos that I have and I can play using my netbook do not satisfy me. I have some big sampled pianos that sound quite good, but my netbook can’t cope with them (my other notebook which I now and again connect to my digital piano obviously does). The only smaller sampled piano that I have that my netbook can manage and sounds reasonably well (though not enough and not as good as Pianoteq) is Sampletekk Black (powered by Kontakt).

So, miracles, sometimes, might certainly happen.

Have you tried reducing the internal sample rate? Minimally changes the sound, but drastically reduces processor usage.

Respeito, Esforço e Sabedoria

Re: U4 would be the less CPU resources demanding pianoteq piano I have!

Professor Leandro Duarte wrote:
Alejandro55 wrote:
marcos daniel wrote:

I can confirm that Kremsegg 1 & 2, Erard 1922 and Pleyel 1926 are less resources hungry.
Erard, Peyel and Bechstein (from Kremsegg) are modern grands, suitables for most music styles (I think). They don't sound as good as the grands you buy singly but they are nicer to me than internal sounds of my digital pianos.

The same on my side. Any of the pianoteq’s pianos sound better than the internal sound of my digital piano.

Since I decided to learn ...............

Have you tried reducing the internal sample rate? Minimally changes the sound, but drastically reduces processor usage.

Yes, I already did. I mentioned that a few posts above this one. It worked.