HQ Player

[The Computer Audiophile]

5 hours ago, Luca72c said:

 

... Many of us use windows for a reason... or more than one...

I really can't understand what's the problem using the i7NUC (or x5minipc) with windows, if it works fine. I hope you didn't credit the fairy tale about HQPlayer sounding better on linux than on windows... Jussi himself once answered that on server PC there is no sound difference using windows or linux, the difference may only be in performance (in my experience no difference), but in your case this difference is clearly in favour of windows, so what's the problem?

About NAA, there could be a difference maybe (i can't confirm), but if HQP NAA image doesn't work, why don't you try other linux NAA images like Gentooplayer, for example? Jussi linux images are sometimes a bit too "minimal" to work ok on all machines... Or just stick with windows, find a cooling solution and live peaceful, just enjoy your music avoiding getting stuck in ridicolous paranoia about linux superiority.

I prefer a Linux NAA image because it works like an appliance. There’s nothing to do to it. It just works. Windows is a completely different animal. 

[Luca72c]

6 minutes ago, The Computer Audiophile said:

I prefer a Linux NAA image because it works like an appliance. There’s nothing to do to it. It just works. Windows is a completely different animal. 

 

As you see, it seems in some cases it just doesn't work well. In those cases, if windows NAA works well instead, why should someone change?

Jussi made an hard work to provide us with optimized, functional and good sounding windows versions of HQPlayer apps, why shouldn't we use them when needed/preferred? Not for better/worse sound reasons, for sure

[The Computer Audiophile]

17 minutes ago, Luca72c said:

 

As you see, it seems in some cases it just doesn't work well. In those cases, if windows NAA works well instead, why should someone change?

Jussi made an hard work to provide us with optimized, functional and good sounding windows versions of HQPlayer apps, why shouldn't we use them when needed/preferred? Not for better/worse sound reasons, for sure

If something works for someone and they are happy, they shouldn’t change. Did someone suggest otherwise?

[Flextreme]

   

21 hours ago, Flextreme said:

Recently rediscovered the DSD filter in HQplayer. Love the gigantic (almost overly audiophile) liquid/plastic presentation of Sinc-L/ASDM7/@DSD256 on my AMD 5900x/RTX 3090 pc sourcing a May KTE.

 

Love to try ASDM7EC, but at DSD64 it barely works, higher DSD rates are constantly drop out, even with multicore DSP and Cuda enabled (and working, verified this). My guess my AMD x5900 is the limiting factor (processor or usb limitations?).

 

Did all recent  ZEN 3 platform updates for USB issues, but the fast MAY firmware (30.12) still does not work for me.

 

Anyone here got the May with 30.14 firmware running ASDM7EC at DSD256x on a AMD ZEN3 platform?

 

OK, after some extensive troubleshooting I think if found a solution to run ASDM7EC at DSD256x on a AMD ZEN3 platform without drops. This is probably specific to ZEN3.

 

Solution: Enable multi DSP and Cuda if possible. Fix and limit affinity of HQP to 3 threads on 3 separate logical processors on CCx1.

 

 

 

Create a shortcut with the command below, replace XXX with CMD, cloudfare blocks references to the cmd executable. This will fix HQP to the first 3 CPUs, non SMT.

 

C:\Windows\System32\XXX.exe /c start "C:\Program Files\Signalyst\HQPlayer 4 Desktop" /High /Affinity 15 "C:\Program Files\Signalyst\HQPlayer 4 Desktop\HQPlayer4Desktop.exe"

 

Basically this is the ONLY way I can run ASDM7EC at DSD256x... All other masking/affinity options (more/less threads) will not work for me. At 4 or more logical CPUs, HQP threads seem to start jumping cores and it this seems to leads dropouts. 2 CPU's/Threads via masking is not enough for ASDM7EC. Ultimately,. something looks wrong with the scheduling of threads of HQP on AMD...

 

I'm confident other AMD users run into the same issues as me, hopefully this can help.

 

Edit: forgot to mention that his dramatically reduces CPU temperature, by more than 10c, and therefore my PC runs as quite as it can be. Very nice...

[Mark Labbett]

hi everyone i have 2 systems a main one(minimax dac supreme with dexa op amps) and a man cave set up(minimax dac plus with dexa op amps) both computers are intel i7 4790k 16GB ram music is on usb hdd's

What would be the best filters for

a) up-sampling everything under dsd128 to dsd128 on the supreme

b) down sampling dsd to 176.4 and up-sampling everything under 176.4 to 176.4 on the plus

I know the millage can vary for different people but i am asking from a purely technical/science side.

Thanks

[lpost]

3 hours ago, The Computer Audiophile said:

I prefer a Linux NAA image because it works like an appliance. There’s nothing to do to it. It just works. Windows is a completely different animal. 

The only time the Linux NAA is problematic is when the device has multiple NICs. In this case hqplayer image with only networkaudiod (hqplayerd disabled) is best as it will auto bridge the interfaces. Just remember to rename the host to avoid network conflicts.

[k6davis]

I've got the new poly-sinc-ext3 filter running comfortably converting Redbook and DSD64 to EC DSD256 on my i7-9700K. I'm really enjoying the sound. It's a nice refinement of poly-sinc-ext2 which I was already very happy with. My rig is not capable of converting Hi-Res PCM to poly-sinc-ext3 EC DSD256, so I have the Nx filter set to poly-sinc-ext2.

 

With the above settings, DSD64 is converted using poly-sinc-ext2. 

 

Jussi, would you consider allowing DSD64 content to be converted with the 1x filter setting? I understand that DSD64 is a higher bit rate than 1x PCM, but it's a light load for the DSP - lighter than Redbook. Plus the vast majority of my (and maybe most people's ??) DSD content is SACD based DSD64. 

 

To ask it in another way, could the difference between 1x and Nx be more like "easy to process" (including DSD64) and "more difficult to process" ?

[JTS]

On 5/5/2021 at 9:44 AM, Miska said:

Use linear phase brickwall filter at the DAC side, this won't have problem in combination with HQPlayer filters.

@Miska I did a peruse of the various filters in combination with the Soekris 2541 set to the Linear Phase Brickwall filter. I am going to paste a few of my observations here from a little listening report I did for another board:

 

Quote

- 192KHz vs 96KHz PCM are very close sounding. The 96KHz is a hair, a touch, a smidge "rounder" sounding.
- the treble extension improves with upsampling. More sparkle, more air.
- the imaging and staging become very precise
- the stage depth increases, as does the 3D qualities of the individual performers.
- transient response becomes even quicker
- midrange presence is a little less "natural" sounding.
- depending on the filter, the sound becomes more clear and snappy. It can cross over into dry and technical depending on the filter.
- Of the range of filters that I tried, my favourites were Poly-Sinc LP and, as always, Poly-Sinc XTR LP
- I find I am paying a bit more attention to the mid-upper range of the music with upressing.
- Without upressing, I am listening more to the mid-lower range.
- With both modes, the treble is always pleasing.
- With upressing, the 2541 renders all the details of complex passages in the music - there is no confusion here.
- The sound is a touch more "natural" without the upressing. I will need to experiment a bit more with dithers to see if I can improve this.
- With upressing, you lose a bit of micro-subtlety in the mids and bass.

 

Are there other dithers that would be appropriate to try when upsampling to 96K or 192K? I love all the improvements that upsampling are bringing to the table but want to chase a bit more of the "natural" quality the 2514 exhibits without upsampling. These comes from midrange tone and all the little round edges in micro details when listening without upsampling.

[ted_b]

1 hour ago, k6davis said:

With the above settings, DSD64 is converted using poly-sinc-ext2. 

 

?  DSD is not converted using ANY filter...it uses modulators and a wide or narrow SDM conversion.  The ext2 or 3 (or any PCM filter) is only used on PCM sources, not DSD64.  Unless you are talking DSD-to-PCM, which I doubt, cuz you are converting PCM to DSD.

[Zauurx]

 

I don't understand. With the latest version, the number of threads keeps increasing... more threads because more actions... OK. Like the apodising counter.
But I have more instability... some crashes... for the first time, a Roon crash (on the same machine) (??).

With version 4.103.. I don't move 118 threads .. in 4.110, I start at 250... goes up to 800 and more !?!

 

[Jean Paul D]

@Miska 

Issues with 4.11 on M1:

 

fade out a few seconds before the end of a track (experienced with pcm to dsd)

need to restart between records (pcm of different SR)

[Jean Paul D]

Client (new version installed from non ARM package since installed on an older MBP) works erratically 

[Jean Paul D]

Wait ! and probably apologies : I have a faulty pair of convo filters : might be the culprit provoking unwanted behaviour (but the fading at end of tracks)

[Jean Paul D]

solved (but fade out a few seconds before end of track)

[fredg_31]

Hello, with the latest version, some Qobuz files aren’t played (white noise instead), if I play another song and retry the faulting song this time it plays perfection.

 

[k6davis]

5 hours ago, ted_b said:

?  DSD is not converted using ANY filter...it uses modulators and a wide or narrow SDM conversion.  The ext2 or 3 (or any PCM filter) is only used on PCM sources, not DSD64.  Unless you are talking DSD-to-PCM, which I doubt, cuz you are converting PCM to DSD.

 

Thanks Ted. I wasn't aware of that. 

 

Roon incorrectly reports that poly-sinc-ext2 is being used during DSD-to-DSD upconversion. I see now that the HQP Client reports "wide", just as you said. 

[Account Closed]

Jussi, I had to revert to 4.10.3 from 4.11 on my M1 system:

 

Problems using ASDM7EC/DSD256

 

1. Dropouts using ext2 (perfect on 4.10.3)

2. Had to restart HQP to get play started (reported above)

3. Track faded before finish (reported above)

 

Thanks for looking into these.

[luisma]

On 3/4/2021 at 3:05 PM, Miska said:

 

 

I actually now have two separate locations, 15 km apart, both with >100 Mbps fiber internet, now I need to set up VPN to be able to access also between the locations.

 

Hi Jussi:

 

You have probably seen this but just in case.

https://www.wireguard.com/

• Much faster than IPSEC and OpenVPN
• Included within the Linux kernel
• Runs in any device, even Raspberry Pi's
• Seems like a good case for your VPN if you don't have it working already :)

Take care

 

[The Computer Audiophile]

2 minutes ago, luisma said:

Hi Jussi:

 

You have probably seen this but just in case.

https://www.wireguard.com/

• Much faster than IPSEC and OpenVPN
• Included within the Linux kernel
• Runs in any device, even Raspberry Pi's
• Seems like a good case for your VPN if you don't have it working already :)

Take care

 

I’d setup a site to site VPN that goes router to router. Then you can avoid any overhead on the computer. 

[Account Closed]

41 minutes ago, The Computer Audiophile said:

I’d setup a site to site VPN that goes router to router. Then you can avoid any overhead on the computer. 

FWIW, I would be very very careful. Once you open a router to outside access (even VPN to VPN), all kinds of potentially bad things become possible. I used to have mine setup so I could access it offsite via its remote access app but the more I read the more I became convinced that it wasn't worth the risk so I shut it all off and only do local network access.

[The Computer Audiophile]

6 minutes ago, bobflood said:

FWIW, I would be very very careful. Once you open a router to outside access (even VPN to VPN), all kinds of potentially bad things become possible. I used to have mine setup so I could access it offsite via its remote access app but the more I read the more I became convinced that it wasn't worth the risk so I shut it all off and only do local network access.

Almost all businesses use VPNs, including banks. I use them all over the world. As long as it is setup right, the risk is fairly low. 

[Quadman]

On 5/11/2021 at 7:32 AM, Miska said:

If a track has let's say under 100 indications, it likely has some clipping but doesn't necessarily need an apodizing filter. If you get over 1000 indications, you'd certainly want to use an apodizing filter.

Jussi, I am a bit confused with the apodizing limit counter in 4.11, btw I think 4.11 sounds excellent, better than 4.10.3.  Last night was the first night with 4.11.  I used only apodizing filters for my entire listening session, yet the limit counter was showing some very high numbers.  Below are 2 screen shots.  One, 1.43 minutes into a track and already I have 1143 limit breakers, another only 12 seconds into a track and already 284 limit breakers.  I saw but did not capture a 6+ minute track with over 3500 limit breakers.  All this using apodizing flters.  So what does this mean if I am getting so many limit breakers with Apodizing filters.  My HQP volume was generally -12 to -9 last evening.  I use Roon as my interface and send generally redbook to HQP.  I7-10700K no cuda offload.

 

1:43 into song 1143 limit breakers

 

12 seconds into first track, I think vol was -16 at the time

[micheloupatrick]

As I understand it, the apodizing counter reflects the content of the track before it is upsampled : so it seems that at least for these tracks, you’re right about using an apodizing filter (although using an apodizing filter on a file that doesn’t need it is OK too).

[ted_b]

I read the screen shot has having NO limiters, just tons of apodizing flags.

 

My problem is I have NO idea what one is supposed to do with this data?  And I assume the apodiing filters are the ones with the "X" in the APO column, so if we use ext2 are we set, regardless?

[austinpop]

On 5/11/2021 at 6:32 AM, Miska said:

sinc-Mx is constant time, meaning that it has variable number of taps but it's length (delay) is constant for each rate multiple. It has 1M taps at 16x output rate, 0.5M taps at 8x output rate, and 16M taps at 256x output rate.

 

So if output rate is set at 16FS, does that mean Sinc-Mx and Sinc-M will be equivalent, both at 1M taps? Irrespective of input sample rate, between 1FS and 8FS?