Best CPU for hqplayer

[bogi]

6 hours ago, sledwards said:

Could the user evaluate the difference using the trial version of Pro? Record a track using DSD 256 and 512 with ASDM7ECv2 modulator. Would that produce a fair comparison

I don't see a technical reason why it wouldn't be a valid comparison.

[bogi]

10 hours ago, SwissBear said:

Do you mean that the processor is protected against overheating and that there is no risk of "frying" it ?

You cannot fry your processor. On OS level, thermal throttling is followed by thermal shutdown if not sufficient. If thermal shutdown on OS level would not come into action then CPU has its own built-in thermal shutdown, which cannot be disabled.

 

I just looked in HWiNFO64 tool for my thermal limits. Throttling on OS level should happen when CPU temperature reaches 100 °C or when GPU reaches 67 °C. At DSD512 my usual CPU temp is 96 °C and GPU temp is about 60 °C.
 

 

 

For me it is more important that my notebook does not get too loud than to care about temperatures which are under limits. With HQPlayer 4.20.2 my notebook was louder, but the CPU and GPU temperatures at DSD512 were the same.

[bogi]

3 hours ago, juanitox said:

is there really a gap in quality bettween  DSD256  vs DSD512 Upsampling with the same Dac and filters ?   when it would start to decrease  512DSD ?  1024DSD ?  or it 's always an upgrade ?

It is often an upgrade, but not always. It depends on DAC used. It may depend on listener preferences (some aspect may improve but some other aspect may get worse).

[bogi]

44 minutes ago, MarekBoro said:

Otherwise this is only "psychoacoustic audiofilia nervosa" 🤣

Better not to continue this way. It seems you don't have listening experience and you don't understand effects of computer generated noise. If we would stay in digital domain (but who is listening to bits?) you could be true. Bits are only bits. But most of members here have experience with effects of noise to audio reproduction quality.

[bogi]

4 hours ago, cpcat said:

When I took the windows boot disk out of the machine it has reliably made the connection, at least so far.

That's unrelated. Maybe reboot helped.

[bogi]

8 hours ago, cpcat said:

bogi, I respect your opinion- however, I have replicated this.

ok ... Maybe you have a boot manager installed on NvME disk which appears as 1st disk in BIOS boot order and that makes the difference when you pull NvME out. BIOS is looking at your inserted disks in some order to find the first bootable one. When it is found, boot process starts from that disk. If that disk contains a boot manager, then you may be able to start an OS from other disk from boot manager on the 1st disk.
You could try to change boot order of your disks in BIOS.

Other possibility is to use a BIOS hotkey to get into BIOS Boot Menu and to select other than the 1st disk to boot from. That's the way to skip boot manager of your current disk 1 and to boot directly form other disk.
I am still curious if a boot manager can disable LAN interface ... but I don't have much experience with Linux/Windows dual boot.

[bogi]

7 hours ago, b0bb said:

When you add or remove the NVME device the identifier for ethernet changes and the network stack is no longer able to find it.

The ethernet interface name is hard coded in /etc/netplan/*.yaml.

This is done at during the OS install.

So it is Ubuntu bug. I added NVMe drive to my notebook and I installed my home OS on it. My work OS on other drive (Windows) was not affected.

[bogi]

37 minutes ago, ArcticSapien said:

The GPUs are eye-watering $$... perhaps it's more cost-effective to rebuild a dedicated HQP Desktop using Ubuntu! (I'm on Windows now).

 

Already a basic RTX model gives nice performance. On Windows my RTX A2000 4GB Laptop GPU works without issues for PCM to DSD512 (even with PEQ like tilt filter and headphone crossfeed) with all filters except few (not enough power for the new sinc-short, -medium, -long and not enough GPU RAM for few extreme long ones like sinc-L, sinc-LI, sinc-Mx, non-2s poly-sinc-long). Everything other works. On DSD256 4GB of GPU RAM is enough for sinc-LI, sinc-Mx. I did not try something like convolution with many taps since I am listening on headphones.

 

DSD64 to DSD512 works well when I don't use PEQ. So for room EQ convolution filters and DSD sources A2000 seems to be not enough powerful.

[bogi]

A friend whom I know throw Slovak hifi forum wants to build a HQPlayer computer which he wants to run headless. He wants to play to NAA. The machine will be placed in the listening room, but least at a distant place. So it needs to be something quiet, but not necessarily passive. He wants to run convolution and he wants to try DSD with suitable DAC. I would like to get some suggestions, these alternatives are viable:

- DSD256 without CUDA

- DSD512 without CUDA

- DSD512 at first maybe a bit restricted on filters and later with added GPU

 

About modulators and filters ... at least EC-light, but better if all EC modulators could work, all higher quality short to middle length filters, no need to fight for extra long ones like sinc-L and similar ... to keep good price to performance ratio.

 

He asked me for opinion about official Intel NUC 12 with i7-1260P ... that reflects his current thinking. How much can be reached from such a NUC? Do you know better alternatives in similar price range?

[bogi]

Miska and StreamFidelity, thanks for your suggestions! He is currently undecided, needs some time.

[bogi]

11 hours ago, Asdfgh said:

Can anyone recommend a great Black Friday deal on a windows laptop that might work ok for 512 EC’s? Im an apple guys so whats good in the windows world is a bit of a mystery to me! Thanks everyone 🙏🏻

 

Look at my signature. It is a notebook capable of PCM to DSD512. RTX A2000 helps a lot.

[bogi]

43 minutes ago, Asdfgh said:

Do you Jedi’s think this thing would do 512 with ASSM7ECv2?
 

https://www.pbtech.co.nz/product/NBKASU360503/ASUS-Vivobook-16X-K3605ZC-16-WUXGA-120Hz-RTX-3050?qr=popular_related_products

 

For modulator it looks ok for me. I cannot compare 3050 to my A2000, but notebook from my sig is able to handle all modulators at 512 and almost all filters.

 

sinc-L, sinc-LI, sinc-M, sinc-long, poly-sinc-hb-l, won't work (mostly because of 4GB GRAM limit). sinc-M and sinc-LI should work at DSD256. So mostly long filters are affected. With improved clarity of my chain after adding a galvanic isolator I clearly perceive transient smear with long filters (Miska and copy_of_a mentioned that more times) so I don't see it as a disadvantage that some long filters don't work for me at 512. I am not often listening to classical music and for genres with faster transients middle to short filters are more suitable.

 

poly-sinc-long-*, poly-sinc-xtr-* don't work too at 512 (they work at 256) on my nb but their -2s versions work at 512.

[bogi]

2 hours ago, Carousel said:

Does CUDA acceleration work in Trial mode?

Yes. Search for nvidia-smi for example here - you can look at CUDA usage with it.

 

2 hours ago, Carousel said:

Should I disable hyperthreading in the BIOS for example???

No. General guidelines are:
- load "Optimized defaults"
- select "XMP Memory Profile 1" if available

 

2 hours ago, Carousel said:

Crank up the heat on the physical cores???

That would help if you are experiencing thermal throttling - it does not look as your case.

 

You can try multicore=“1” as alternative to multicore=“auto”.

[bogi]

7 hours ago, wakalaaaa said:

If I only want to upsample dsd256, 13900H 1370P, CPU, is it okay?  I hope to use sinc L, 7ec super.

Did you already listen to sinc-L, do you know how does it sound? Did you compare it to other filters by listening? Is classical music the genre you prefer?

[bogi]

3 minutes ago, wakalaaaa said:

I tried sinc -L, dsd 128 on mac mini m2, I like its transparent tone and good separation.  I rarely listen to classical music, sometimes I listen to movie soundtracks.

OK. I only wanted to know if you know and like it. Because, at least technically, long filters are more suitable to classical music than music with fast transients, since they introduce transient smear with fast transients. Miska mentioned it more times.

 

What's your DAC? Is it connected through USB? Do you use any kind of isolation from computer noise?

 

Why I am asking? I preferred long filters in the past. I perceived fuller more weighty sound with nice timbre of instruments. With short filters I perceived missing impact. Then I bought a galvanic isolator. That cleaned up sound and I started to perceive transient smear with long filters more intensively. I changed my preference to middle to short filters, whose sound improved with added isolation.

[bogi]

CPU power of 13900H seems to be sufficient for sinc-LI at DSD256. Maybe others will yet add their experience.

On my laptop with nVidia A2000 GPU and 4GB of GRAM sinc-LI works at at DSD256. AFAIK sinc-L requires twice so much RAM. So it looks to me 16GB of RAM on computer without nVidia GPU should be sufficient for sinc-L at DSD256.

 

You can also compare sound of sinc-LI and sinc-L. sinc-LI is easier reachable, it possibly could work at DSD256 with your current Mac Mini. My other tip for you to try and compare with sinc-L is poly-sinc-hb-m.

[bogi]

3 minutes ago, Carousel said:

50-70% with sinc medium.

What is your HQPlayer version? Because the newest contains update of sinc-short, sinc-medium and sinc-long - they are now 2 stage (for DSD output) and are easier on resources. I can easily do sinc-medium on notebook from my signature below.

For sinc-long you have probably not enough GPU RAM.

[bogi]

49 minutes ago, Carousel said:

Is it technically possible to upscale a PCM 24-48 source file to DSD256 using sinc filters like sinc-M or sinc-Mx? My set-up can do a 16-44 source file, but not a 24-48 source file.

To understand this behavior, you need to look into chap. 4.5 of HQPlayer 5 Desktop manual, which is installed together with application files. You can find that these sinc-* filters contain 'integer' in the 'ratio' column of the filter table. It means these filters allow only integer upsampling ratio. For example then you are upsampling from 44.1k to DSD256 (44.1*256=11.29MHz), the upsampling ratio is 256. But when you are upsampling from 48k to 44.1*256, upsampling ratio is non integer and therefore HQPlayer refuses to perform the operation. No hardware update helps on that.

 

What could help you is a DAC which supports so called 48k based DSD rates, for example 48*256=12.28 MHz. Otherwise the only possibility for you is to use a more or less similar filter which allows any conversion ratio, like poly-sinc-long-lp-2s, poly-sinc-ext3, poly-sinc-hb-l, poly-sinc-gauss-long. Worth of try could be also middle length filters like poly-sinc-hb-m, poly-sinc-lp-2s and with hires material poly-sinc-gauss-hires-lp.

[bogi]

25 minutes ago, Carousel said:

ticking 48K DAC in HQP does not seem to make it into one though!

To have it really tested, you need yet to set max. DSD rate in HQPlayer settings to 48k based one and also I recommend to set [x] Adaptive rate which switches output rate between 44.1k and 48k families according source rate. Then try to play 48k based content. If it does not work, just undo these changes.

Maybe there is an exaSound S88 user who could tell you if 48k based DSD rates work with that DAC.

[bogi]

12 minutes ago, blackhawk579 said:

Can you critque my HQ Payer/ Roon Server silent PC build (below)?

 

Not a critique but a basic question. Did you consider using HQPlayer NAA (I don't know what's your DAC)? Then you could place your audio computer in any other room. Then it wouldn't be needed for it to be so silent like a comp placed in a listening room. That would make it also less expensive and cooling would less restrict its performance.

NAA advantage is not only the fact it can be placed in a different room than HQPlayer computer, so HQPlayer computer can be configured to play (not at once) to more NAAs placed in dirrerent rooms. Important advantage of low power NAA computer is is that it isolates DAC from high power and thus noisy HQPlayer computer.

[bogi]

3 hours ago, dericchan1 said:

everything else other than hqplayer desktop seem to be not working well with RPI5

That's an opposite pole to bug report. 🙂 It rarely occurs in software world.

[bogi]

@blackhawk579  for DSD512 and your intended use (I assume 2ch stereo) that IMO should be ok. Less powerful setups are able of DSD512 too. But for other work on computer it is good to have some spare resources. Only if you are fan of few longest and hardest to process filters I would suggest more graphics RAM (I would be OK with those 8GB).

DSD1024 could work with most of filters too on that setup (except of long filters which require more RAM on graphics card).

 

The only questionable thing for me is quality of cooling under constant high load. The worst case would be thermal throttling occurring periodically. May be you can find some reviews or you can reach other users to be sure on this point.

[bogi]

Some tips to consider (I don't own such, wait for other suggestions):

Used i3-1215u  https://www.ebay.de/itm/364626305013?hash=item54e56bdff5:g:cLsAAOSwgNlleFyE
New: i5-13500H https://www.ebay.de/itm/256361121182?hash=item3bb050459e:g:OCEAAOSwAB1ld0bc
Review https://www.techradar.com/pro/chuwi-corebox-5th-mini-pc-review

https://www.youtube.com/watch?v=CHc_zLoG5t4

 

[bogi]

@dericchan1  What's the price of Pi5 with such a cooling?

[bogi]

@dericchan1  You may yet try to play with 'Blocks per cycle' value.

 

Are you playing to 48k DSD capable DAC? (Cyan2 ?). Do you have 'Adaptive rate' checked?