Hello everyone,
I want to run Pianoteq Standalone on a Intel Celeron N4100. First of all does Pianoteq work on this processor (or more generally on Intel Celeron processors) and how well does it work? Can it run on high settings? When not is it really a problem (sound quality) when in doesn't run on the highest settings?
Best regards
Xeran
You will probably have to cut the max polyphony quite a lot (64 or 32 voices), and perhaps run at lower sample rate...
The N4100 was released Q4 of 2017 as the top end mobile Celeron. Should be interesting to see how well it runs Pianoteq. Where did you get the computer?
You will probably have to cut the max polyphony quite a lot (64 or 32 voices), and perhaps run at lower sample rate...
I think the polyphony is not the biggest problem. A too low buffer size would be a problem.
Where did you get the computer?
I don't have it. That's because i'm asking. This is the mini computer I thought off https://www.kickstarter.com/projects/13...le_created
Yeah, you should be able to run it at 128 samples buffer size... depending on audio interface.
Can a better audio interface reduce the stress on the cpu? I never had an audio interface. I hope that's not an stupid question.
Well, depending on how well-written ASIO drivers are for it. RME are legendary for this (check out Babyface).
Is there any real experience with its 2016 predecessor N4200 already? It is very similar to the brandnew N4100. If the N4200 works, N4100 should too.
With 6 W TDP both can be passively cooled. But it would be interesting to hear from first hand, whether 1.1 GHz "normal" cpu-freq is enough. In the past often >= 2 GHz were recommended for Pianoteq.
In burst mode 2.4 - 2.5 GHz are possible with those N4xxx - but without thermal problems?
I am optimistic because my much older N2930 with 1.83 - 2.16 GHz works flawless in a fanless netbook with a standard linux kernel (64 samples buffer size, onboard pci-soundcodec).
Well, depending on how well-written ASIO drivers are for it. RME are legendary for this (check out Babyface).
That's confusing me, because in my other thread here: http://www.forum-pianoteq.com/viewtopic.php?id=5416
I was told that the process of generating the signal takes place completely in the cpu, and the audio Interface is just responsible for the digital / analog conversion:
Pianoteq runs in the computer and produces the sound as digital samples, (for example 48000 samples of 24 bits each per second) but these bits have to be converted to audible sound through a DAC (digital to analog converter). There is one in your computer, that can be of lower quality and more prone to noise (due to others circuits around it) than if you buy an external USB card, that also furnishes a number of MIDI input channels for you keyboard.
So you need a fast enough computer for pianoteq's computation of the samples, but you get better audible results using a good external soundcard.
Well, it's not quite like that. Audio interface serves as a bridge between what the CPU computes and how it will sound on your speakers. And actually, your OS stands in between that bridge, too, and this is where drivers come into play, to give CPU a more direct access to the audio interface, than going through OS hardware abstraction layers. This has impact on latency - the better the drivers an audio interface has, the lower the latency. And of course, audio interface CAN slightly color the sound that the CPU produces, because different DACs have different specifications.
I had the chance to test a fanless budget notebook with Intel N4200 (Apollo Lake) today.
Nice Performance index 43:
I used the Blues Demo of Pianoteq Trial v6.0.3 and the default instrument settings on Debian/Stretch Linux with KDE plasma-dektop.
The onboard soundcodec ALC255 accepted a minimum buffer size of 128 samples. Major tweak was setting the cpu scaling_governor from "powersave" to "performance" via /etc/rc.local.
Cheers