I bought fanless mini PC (W11 Pro) with an Intel Celeron N5105 processor running at 2+GHz,
to use as a low noise source running JRiver to feed my 8-channel Topping DM7.
I just ran it a bit yesterday and I thought I heard a few minuscule hiccups or dropout.
Anybody has an insight into what sort of processor/speeds would be required (48kHz, 6 channel FLAC with JRiver convolver).
Jan
https://www.amazon.com/MeLE-Quieter2Q-Processor-Quad-Core-Computer/dp/B0BYZ91RFL/ref=sr_1_3
to use as a low noise source running JRiver to feed my 8-channel Topping DM7.
I just ran it a bit yesterday and I thought I heard a few minuscule hiccups or dropout.
Anybody has an insight into what sort of processor/speeds would be required (48kHz, 6 channel FLAC with JRiver convolver).
Jan
https://www.amazon.com/MeLE-Quieter2Q-Processor-Quad-Core-Computer/dp/B0BYZ91RFL/ref=sr_1_3
Last edited:
Just check your CPU load at task-manager and you will actually know.
But basically any CPU for the last 10 years is enough for audio.
When doing some other tasks, 4 cores at least is really nice to have.
But basically any CPU for the last 10 years is enough for audio.
When doing some other tasks, 4 cores at least is really nice to have.
Audio playback is not at all demanding for CPUs.
Try increasing the buffer size in JRiver. When using ASIO4ALL, it can also be done in the ASIO4ALL control panel.
With a dedicated ASIO driver for your soundcard you should be able to adjust it there if it has a control panel.
This example is for MusicBee:
Hugo
Try increasing the buffer size in JRiver. When using ASIO4ALL, it can also be done in the ASIO4ALL control panel.
With a dedicated ASIO driver for your soundcard you should be able to adjust it there if it has a control panel.
This example is for MusicBee:
Hugo
N5105 has way more than enough CPU horsepower. Often times it is the operating system getting in the way, or I should say not getting out of the way because it is doing other things.
I own multiple Beelink U59 Pro mini-PCs with the N5105 to run DSP crossovers for my loudspeaker systems. I can use USB audio devices with up to 20 channels at 96kHz without a hiccup under Ubuntu. Before that I used less powerful J4105, N4100, even back to the J1900 SOC or Raspberry Pi 3 and they could all handle 8 channels of audio up to 192kHz via USB2 without a problem. Audio is just not all that demanding from the PC.
I own multiple Beelink U59 Pro mini-PCs with the N5105 to run DSP crossovers for my loudspeaker systems. I can use USB audio devices with up to 20 channels at 96kHz without a hiccup under Ubuntu. Before that I used less powerful J4105, N4100, even back to the J1900 SOC or Raspberry Pi 3 and they could all handle 8 channels of audio up to 192kHz via USB2 without a problem. Audio is just not all that demanding from the PC.
That software must be doing too much.
I measure <1% utilization of one CPU core on lossless audio decompression and playback (stereo 44.1KHz/16-bit).
Ed
I measure <1% utilization of one CPU core on lossless audio decompression and playback (stereo 44.1KHz/16-bit).
Ed
@jan.didden did not say exactly what he is doing with JRiver...
There is a big difference between the CPU demand from decoding and playing back audio, doing IIR DSP, and doing FIR DSP and lots of room correction with filters that have a high number taps.
Even with CPU-intensive tasks Jan may just need to tweak a few thinks like the buffer settings (as mentioned above).
There is a big difference between the CPU demand from decoding and playing back audio, doing IIR DSP, and doing FIR DSP and lots of room correction with filters that have a high number taps.
Even with CPU-intensive tasks Jan may just need to tweak a few thinks like the buffer settings (as mentioned above).
I was happy with a P-III at 733 MHz for audio. 256MB RAM, 320 kbps, Winamp series 4!
In 2003...
Two things:
the OS, and whether you are playing music, or doing complex stuff like multiple file encoding.
See the task manager or its Linux equivalent to check.
How much RAM?
8 GB or at least 4 is okay, DDR4 or more.
Other bottlenecks could be latency to storage media (in the same case or other location).
In 2003...
Two things:
the OS, and whether you are playing music, or doing complex stuff like multiple file encoding.
See the task manager or its Linux equivalent to check.
How much RAM?
8 GB or at least 4 is okay, DDR4 or more.
Other bottlenecks could be latency to storage media (in the same case or other location).
I am using JRiver with DSP processing that I developed with Acourate.
It is a combination of a 2-way system with pretty heavy room correction.
But I will look at the ASIO settings as Hugo suggested (I use a Topping DM7's ASIO driver).
There may be more involved because the W11 start up is very slow also.
It has 8GB memory so that should be OK. The music sits on a microSD in the PC.
Jan
It is a combination of a 2-way system with pretty heavy room correction.
But I will look at the ASIO settings as Hugo suggested (I use a Topping DM7's ASIO driver).
There may be more involved because the W11 start up is very slow also.
It has 8GB memory so that should be OK. The music sits on a microSD in the PC.
Jan
https://support.microsoft.com/en-us...-windows-da2f9573-6eec-00ad-2f8a-a97a1807f3dd
Realy boring, but somtimes it has to be done.
Realy boring, but somtimes it has to be done.
W10 was a headache, always sending data over the net, tying up resources.
W11 is said to be better, I would disable the net connection or restrict it.
These versions of Windows have heavy anti piracy and data collection, you will find it difficult to use pirated programs, and they tell Big Brother what you are doing.
That uses up data, and processor time.
There are programs called 'decrapifier' or similar, those help reduce these issues.
I would not use the computer at the same time for other work, until I got used to it. Keep it disconnected from internet after the initial Windows genuine process, validating the install in your name.
For fun, see the disk space taken by a fresh install of Windows, XP was 1 GB, 7 was 4, 10 is ... well, look it up!
W11 is said to be better, I would disable the net connection or restrict it.
These versions of Windows have heavy anti piracy and data collection, you will find it difficult to use pirated programs, and they tell Big Brother what you are doing.
That uses up data, and processor time.
There are programs called 'decrapifier' or similar, those help reduce these issues.
I would not use the computer at the same time for other work, until I got used to it. Keep it disconnected from internet after the initial Windows genuine process, validating the install in your name.
For fun, see the disk space taken by a fresh install of Windows, XP was 1 GB, 7 was 4, 10 is ... well, look it up!
If normal windows cleaning don't help we can also get some guidelines from the windows users on stage
https://gigperformer.com/docs/ultimate-guide-to-optimize-windows-for-stage/index.html
https://gigperformer.com/docs/ultimate-guide-to-optimize-windows-for-stage/index.html
Sometimes the trouble is that the PC is too powerful for the application (Like music playback with light filtering). Then "standard windows setup" aggressivly tries to save power and suddenly there are to few resources left when other programs interfere.
Ideally you'd use a real time OS for audio, mainstream OS's aren't really optimized for low latency, but for ultra-high throughput. People see the similar timing issues with driving CNC hardware from stock processors running stock OS's. Without low-latency you are dependent on having large enough buffers and not being delayed too long by other hardware drivers.
Even a fast microcontroller is powerful enough for most straightforward audio DSP processing.
Even a fast microcontroller is powerful enough for most straightforward audio DSP processing.
I put my PC into sleep mode each night and only do a full restart every week or so.
Helps a lot.
Hugo
In fact, a small FM module can give decent sound from a USB /SD Card store... it costs a little over a dollar here.
Anything more powerful is an improvement...that little chip is all that is really needed.
Some people have made cell phones and tablets work as audio devices, with large memory cards.
In sum, the processor is adequate, the OS needs work.
Anything more powerful is an improvement...that little chip is all that is really needed.
Some people have made cell phones and tablets work as audio devices, with large memory cards.
In sum, the processor is adequate, the OS needs work.
Have you considered using Linux e.g. Ubuntu? There is a JRiver version for Linux.
https://www.resplendence.com/latencymon is mentioned in the stage site. It is able to find a lot of dodgy behaviour in windows. And peace of mind if it dont.
Free home use version
Free home use version
I'm not a computer guy. I shudder by the thought to install and work with an operating system totally foreign to me.
You know when I set up this new W11 miniPC, Windows offered to configure the system as my other system.
Just a single yes and all of a sudden my OneDrive was there with all my work files, all apps on the taskbar, installed if it could, ready to download if it must.
I like that, eating from my nose while the stupid PC does my bidding 😎
Jan
Checking the JRiver CPU consumption in the Task Manager while playing would tell more about the actual strain on the PC. DSP load can be very low or it can be running at the edge.
This should be the setting https://wiki.jriver.com/index.php/ASIO . You definitely want the large buffers (as large as possible). Directly to an ASIO driver (no ASIO4All/Flexasio in between). If no ASIO drivers were available, WASAPI Exclusive again with maximum buffers offered.
Do you have windows installed on the NVMe SSD drive or the eMMC drive? Typically eMMCs are considerably slower. But that should not hinder the playback fluency much.
IMO there is one more thing. In your post you were not sure about the glitches. Therefore it would be perfect if you could verify there are some/no glitches not by listening, but by reading some log. If JRiver outputs directly to the driver with no extra SW in between (i.e. directly ASIO or WASAPI excl), it should know about the glitches (unless the glitches occur in the USB audio -> USB core driver communication which IMO is quite unlikely). IMO asking at jriver forum about options for tracking underruns in some logs (IIUC JRiver is a paid-for SW which should offer corresponding customer support) could help. Once you can "see" the events of underruns, it's much easier to monitor improvements.