HQ Player

[ericuco]

46 minutes ago, CheapSplurge said:

Which do I download and I just put it on the pi SD card? After I erase and format the SD card that goes into the pi? And the HQP image is the only thing that should be in the SD card for the pi? 

If you are just using the pi as a HQP NAA, then you use one of the pi images from https://www.signalyst.eu/bins/naa/images/

 

As Chris mentions above, download image, unzip it, write image file to SD card using Etcher, plug the SD card into your pi, power up and your done. You can’t log-in to the pi with the NAA image so you can’t install any additional software like Roon endpoint.

[CheapSplurge]

With all of your help I got it done.

Thabk you guys so much. 

[3ggerhappy]

Hi Guys, 

 

how do you think this mini pc would perform with hqplayer embedded, also is the spec decent enough for its price?

 

https://www.hystou.com/gaming-mini-pc-f7-gtx-1650

 

 

[Miska]

1 hour ago, 3ggerhappy said:

Hi Guys, 

 

how do you think this mini pc would perform with hqplayer embedded, also is the spec decent enough for its price?

 

https://www.hystou.com/gaming-mini-pc-f7-gtx-1650

 

 

 

The i7 CPU version seems quite decent. But the challenge with such small powerful devices is fan noise. Since the fans are small they need to rotate at high speeds which means they tend to get loud...

 

[Hammer]

Hi, just if anyone’s using the Topping D90 MQA with the latest firmware 1.24, seems using Default DAC bits In Settings produces and error.  If I set to 32 bits, then it works fine...

[Miska]

33 minutes ago, Hammer said:

Hi, just if anyone’s using the Topping D90 MQA with the latest firmware 1.24, seems using Default DAC bits In Settings produces and error.  If I set to 32 bits, then it works fine...

 

Which OS and what kind of error?

 

[niner93]

I need some help here for my drop out issue.  I have played around available settings (Buffer time, Multicore DSP, ...) but nothing fix the drop out.  I really appreciate any comment or suggestion.

 

I built a new machine, among other things, to run 6 or 8 channel crossover with HQplayer.  The computer spec is below:

- HQPlayer 4.6.0

- AMD Ryzen Threadripper 3970x

- Gigabyte TRX40 Aorus Master motherboard

- 64 GB (4 x 16 GB) DDR3200

- 1TB M2 SSD

- EVGA RTX2080Ti

- Custom water cool loop

- Windows 10 pro for Workstation build 2004 (also tried previous release)

- Clean environment, basically Hqplayer is the only app installed

- Default settings, no overclocking

 

I will use Merging Hapi DAC (meaning RAVENNA connection).  But for now, I'm testing everything with a exaSound e28 8-channel DAC.  Convolution is disabled to debug the drop out issue.  Filters are poly-sinc-ext2, ASDM7EC and 48kx256.

 

With ASDM7:
- e28 as 8 channel DAC, and Hqplayer to output all channels. Audio is perfect, no drop out at all.

With ASDM7EC:
- e28 as 2 channel DAC, and Hqplayer to output 2 channels. Audio is perfect, no drop out at all.
- e28 as 8 channel DAC, and Hqplayer to output only 2 channels. Audio dropped out (pause) every 15 seconds or so.  The timing of the drop out is fairly consistence.
- e28 as 8 channel DAC, and Hqplayer to output all channels. Audio dropped out almost every second.  CPU temp is only 56C max.

 

[niner93]

I forgot to include the matrix pipeline setting

[StreamFidelity]

59 minutes ago, niner93 said:

I need some help here for my drop out issue.

 

Just my thoughts:
- If you go out with 48x256, why didn't you activate 48k DSD?
- Have you checked the latencies? Look LatencyMon.

[Miska]

3 hours ago, niner93 said:

With ASDM7:
- e28 as 8 channel DAC, and Hqplayer to output all channels. Audio is perfect, no drop out at all.

 

Yes, this is expected to work.

 

3 hours ago, niner93 said:

With ASDM7EC:
- e28 as 2 channel DAC, and Hqplayer to output 2 channels. Audio is perfect, no drop out at all.

 

I'm actually slightly surprised that this works with this CPU at all, but great that it does! Probably it can reach high enough boost clocks for those few cores.

 

3 hours ago, niner93 said:

- e28 as 8 channel DAC, and Hqplayer to output only 2 channels. Audio dropped out (pause) every 15 seconds or so.  The timing of the drop out is fairly consistence.

 

This increases load a little bit, can be seen also from the graphs and seems to be enough to just make it miss deadlines every now and then. Note that WiFi drivers tend to be latency offenders. So could be something as simple as periodic scanning for networks.

 

3 hours ago, niner93 said:

- e28 as 8 channel DAC, and Hqplayer to output all channels. Audio dropped out almost every second.  CPU temp is only 56C max.

 

This is sort of expected since stereo case is already pretty tight on resources, higher number of cores is loaded so likely individual cores don't get as high boosts. And memory bandwidth / cache loads are increased too, so resources are tighter.

 

If you have some tool to check per core clocks, it could provide some info whether it is limited boost or something else being bottleneck.

 

[Theobetley]

I have inquirees into the Sonore link (no response yet) but Miska maybe you have some input on this. I turn on my dac then Hqp but I show no connectivity to my dac in Hqp settings. I go into the Mysonicorbiter site and when I try to click on Manage within My S... it shows cant open site error. Any ideas?

[R1200CL]

11 minutes ago, Theobetley said:

I have inquirees into the Sonore link (no response yet) but Miska maybe you have some input on this. I turn on my dac then Hqp but I show no connectivity to my dac in Hqp settings. I go into the Mysonicorbiter site and when I try to click on Manage within My S... it shows cant open site error. Any ideas?

You may edit your post and explain where your HQPlayer is resting, and what DAC. If you’re using SonicTransporter you need to add “:8088/config” after the IP. (But I guess you’re able to access HQPlayer). If not, please see my reply the Sonore Rendu thread. 

[WilliamWykeham]

I've searched all through the forums, can't find a solution to my problem, am wondering if anyone here knows what to do:

 

I have a JCAT Network card that I've set up on my single-box computer with HQP and RoonServer, in Windows Server 2019.  

 

When I bridge the ethernet ports on the network card and connect this way to either of my endpoints (Allo Digione and Allo USBridge), neither one is recognized in HQP.  

 

HQP does recognize these endpoints when they are connected in a non-bridged manner, to a switch connected to my router.  It's only when I bridge that I get issues.  

 

I've checked via my router and the endpoints are both indicating an active connection to my network when bridged.  

 

I don't think I need to change the IP address on my endpoint; I'm using they assigned IP address.  

[Miska]

20 minutes ago, WilliamWykeham said:

HQP does recognize these endpoints when they are connected in a non-bridged manner, to a switch connected to my router.

 

This is the way it is supposed to be connected. Why are you trying to do something else?

 

[WilliamWykeham]

4 minutes ago, Miska said:

 

This is the way it is supposed to be connected. Why are you trying to do something else?

 

 

For curiosity's sake.  

 

Based on JCAT's Scenario 1 Scheme here: https://jcat.eu/featured/net-card-femto/

 

And after reading the "Novel way to massively improve..." thread on this site. 

[Miska]

13 minutes ago, WilliamWykeham said:

For curiosity's sake.  

 

Based on JCAT's Scenario 1 Scheme here: https://jcat.eu/featured/net-card-femto/

 

And after reading the "Novel way to massively improve..." thread on this site. 

 

You can use such at the NAA side if you like. But NAA itself should be connected to the regular home network with a good switch (Cisco and HPE for example make such).

 

Otherwise you tend to end up with problems like you describe.

 

[niner93]

6 hours ago, StreamFidelity said:

 

Just my thoughts:
- If you go out with 48x256, why didn't you activate 48k DSD?
- Have you checked the latencies? Look LatencyMon.

Thanks for your suggestions.

 

I tried 44.1x256 and I still have drop out.

I ran LatencyMon, and the system doesn't seem to have an issue with latency.

[niner93]

4 hours ago, Miska said:

 

Yes, this is expected to work.

 

 

I'm actually slightly surprised that this works with this CPU at all, but great that it does! Probably it can reach high enough boost clocks for those few cores.

 

 

This increases load a little bit, can be seen also from the graphs and seems to be enough to just make it miss deadlines every now and then. Note that WiFi drivers tend to be latency offenders. So could be something as simple as periodic scanning for networks.

 

 

This is sort of expected since stereo case is already pretty tight on resources, higher number of cores is loaded so likely individual cores don't get as high boosts. And memory bandwidth / cache loads are increased too, so resources are tighter.

 

If you have some tool to check per core clocks, it could provide some info whether it is limited boost or something else being bottleneck.

 

Thanks Miska for the reply.

 

A couple updates:

1) I switched over to use the Merging HAPI DAC over Ethernet (RAVENNA), and the drop out is still there.

2) I turned off WiFi, unfortunately it didn't help.

 

I used HWiNFO to capture core clock while the system is running above configuration.  Please let me know if it helps.

[Miska]

45 minutes ago, niner93 said:

Thanks Miska for the reply.

 

A couple updates:

1) I switched over to use the Merging HAPI DAC over Ethernet (RAVENNA), and the drop out is still there.

2) I turned off WiFi, unfortunately it didn't help.

 

I used HWiNFO to capture core clock while the system is running above configuration.  Please let me know if it helps.

 

Also note that HAPI cannot do DSD at 48k multiples, only at 44.1k.

 

You could compare the clock figures to the stereo case to see if stereo case that doesn't have dropouts runs those few cores at higher clocks than here.

 

[niner93]

4 hours ago, Miska said:

 

Also note that HAPI cannot do DSD at 48k multiples, only at 44.1k.

 

You could compare the clock figures to the stereo case to see if stereo case that doesn't have dropouts runs those few cores at higher clocks than here.

 

Thanks for the tip.  Yes, I found out very quickly that hapi doesn't support 48 in DSD.

 

I looked at the core clock while running 2 channel output.  You are right, in 2-ch output configuration, the core clock frequencies are a little bit higher.  Not by much, but they are higher.  Probably right on the edge.

[StreamFidelity]

9 hours ago, niner93 said:

I ran LatencyMon, and the system doesn't seem to have an issue with latency.

 

Yes, but the NT Operating System Kernel is at the top. The lower the latencies, the better. An example of my audio PC:

 

9 hours ago, niner93 said:

I used HWiNFO to capture core clock while the system is running above configuration.

 

Many processes (142) are active. The less the better, an example:

[Miska]

When looking at those Windows load graphs, it is useful to right-click the graph and select it to also show kernel/system times in addition to user space. This way you get better picture of what is going on.

 

As long as latencies stay below millisecond range, they are unlikely to cause issues. Another aspect is how frequently the system hits the longer latency paths, and thus the cumulative effect. But those are shown to some extent through the system time loads I'm talking about above.

 

[CheapSplurge]

What settings in HQplayer has more revealing, brighter presentation for chord dacs? 

[numlog]

try normal poly-sinc

[Account Closed]

48 minutes ago, CheapSplurge said:

What settings in HQplayer has more revealing, brighter presentation for chord dacs? 

The brightest, punchiest, sharpest transient and generally most in your face filter is probably poly-sinc short mp.