HQ Player

[terzinator]

6 hours ago, terzinator said:

 

ah, interesting. Ok, I will test that. THANK YOU!

 

 

YESSSSSS. 

 

This did it. Set to 192k PCM and all is well. I never even thought to change the rate. 

 

THANK YOU.

[mrvco]

2 hours ago, Outlaw said:

Hqplayer 4 is a upgrade that has to be purchased.Hqplaer key for 3 wont work.

 

Wow, I had no idea.  Thanks..  Looks like I'll be sticking with v3 for now.  Not sure what v4 adds over v3, but unfortunately it isn't in the budget at the moment.

[mrvco]

nm, doesn't that my 2012 Mac Mini would benefit.  Makes a new Mac Mini all that much more enticing I suppose.  Thanks @Miska for keeping v3 available.

[Xoverman]

Hi everyone,
does’ anyone have good experience up sampling DSD64 (SACD) to DSD512.
I can’t find any combination that sounds better then bypassing DSD directly to the DAC ( es9038Pro ).
Sending DSD directly to the DAC has better Transients, and deeper soundstage.
Up sampling PCM to DSD has a big Sonic impact . Everything is improved with PCM.

[Yviena]

@Miska Is it possible to maybe get a build where all corepinning/cpu allocation is disabled allowing the OS to do all the thread scheduling, as i want to see if it's maybe better if the OS handles everything as the latest windows 1909 has again changed some stuff, and now with your latest build  with parallel set to both auto, and checked cpu usage is capped to 5% on HQplayer no matter what i do, i wonder if there's some sever threading conflicts going on to create this...

[Miska]

1 hour ago, Yviena said:

@Miska Is it possible to maybe get a build where all corepinning/cpu allocation is disabled allowing the OS to do all the thread scheduling, as i want to see if it's maybe better if the OS handles everything as the latest windows 1909 has again changed some stuff, and now with your latest build  with parallel set to both auto, and checked cpu usage is capped to 5% on HQplayer no matter what i do, i wonder if there's some sever threading conflicts going on to create this...

 

You already have that done through the environment variable?

 

[Yviena]

1 hour ago, Miska said:

 

You already have that done through the environment variable?

 

yeah i do but, somehow this newest windows update has completely fucked it up i think they are doing selective testing with who get's updated build first, no matter what settings i change or do hqplayer is capped to around 3-6% cpu usage now.

 

I think i will need to reinstall older windows now, and disable updates...

[Miska]

23 minutes ago, Yviena said:

yeah i do but, somehow this newest windows update has completely fucked it up i think they are doing selective testing with who get's updated build first, no matter what settings i change or do hqplayer is capped to around 3-6% cpu usage now.

 

Sounds like "fun"

 

Somehow I admire a little the scheduler in macOS because so far it has been doing best job of all three platforms, when not helped at all...

 

[Yviena]

6 hours ago, Miska said:

 

Sounds like "fun"

 

Somehow I admire a little the scheduler in macOS because so far it has been doing best job of all three platforms, when not helped at all...

 

Ok so I managed to find the cause, windows update was just a coincidence, I think it is network issues that made HQplayer cap out at 5% when I use a direct connection now I get the expected CPU usage around 22% but I still hear really soft pops/clicks so I will try another router tomorrow, who knows maybe it's the Ethernet NIC of the pc, or the NAA that has gone bad.  Weird tough that my previous NAA from same brand did not have soft clicks/pops at same CPU usage from upsampling server.

[Xoverman]

@miska & everyone else ,
I'm enjoying the full version of HQPLAYER desktop for quite some time now, 

Up sampling PCM to DSD512 has a big Sonic impact on my system.

But I can’t find any combination that sounds better then bypassing DSD directly to the DAC ( es9038Pro ). Sending DSD64 directly to the DAC has better Transients, and deeper soundstage more space on my system. 

What am I doing wrong? I just can't believe that the es9038pro dos a better job upsample dsd64 then HQPLAYER cold do.

[Miska]

5 hours ago, Yviena said:

Ok so I managed to find the cause, windows update was just a coincidence, I think it is network issues that made HQplayer cap out at 5% when I use a direct connection now I get the expected CPU usage around 22% but I still hear really soft pops/clicks so I will try another router tomorrow, who knows maybe it's the Ethernet NIC of the pc, or the NAA that has gone bad.  Weird tough that my previous NAA from same brand did not have soft clicks/pops at same CPU usage from upsampling server.

 

Ok, good that it wasn't a Windows update breaking things. Maybe there's a lot of packet loss for some reason?

 

For slower NAA devices it is important to have Ethernet flow control (802.3x) enabled to avoid lot of packet loss and resends due to packet loss. At least RaspberryPi 1 & 2 have such clicking problems at higher rates because they have limited bandwidth capability and both USB ports and the Ethernet interface are connected to the same USB bus. RasPi3 should be less problematic in that respect and I believe (although I have not tested) RasPi4 has fixed that altogether. However, RasPi2 works fine through it's I2S interface to DAC boards like HiFiBerry. Some patches have added support for 352.8/384k rates on the I2S, but that in turn is not reliable and the DMA gets easily messed up resulting in channels being swapped, etc.

 

What kind of NAA device do you have and what are you running on it?

 

[Miska]

12 minutes ago, Xoverman said:

@miska & everyone else ,
I'm enjoying the full version of HQPLAYER desktop for quite some time now, 

Up sampling PCM to DSD512 has a big Sonic impact on my system.

But I can’t find any combination that sounds better then bypassing DSD directly to the DAC ( es9038Pro ). Sending DSD64 directly to the DAC has better Transients, and deeper soundstage more space on my system. 

What am I doing wrong? I just can't believe that the es9038pro dos a better job upsample dsd64 then HQPLAYER cold do.

 

I'm running DSD upsampled to DSD256. (but I'm not using an ES9038Pro based DAC)

 

AFAIK, ESS doesn't upsample DSD inputs.

 

If you are using HQPlayer4, you have three settings that affect DSD sources. Modulator you use, and then in DSD Sources settings you have "SDM Integrator" and "SDM Conversion". Have you checked what kind of sonic difference you get with these? Modulator choice should be pretty straightforward, because you likely use the same one you also use for PCM sources. So you are left with two settings in the DSD Sources dialog.

 

[Yviena]

3 hours ago, Miska said:

 

Ok, good that it wasn't a Windows update breaking things. Maybe there's a lot of packet loss for some reason?

 

For slower NAA devices it is important to have Ethernet flow control (802.3x) enabled to avoid lot of packet loss and resends due to packet loss. At least RaspberryPi 1 & 2 have such clicking problems at higher rates because they have limited bandwidth capability and both USB ports and the Ethernet interface are connected to the same USB bus. RasPi3 should be less problematic in that respect and I believe (although I have not tested) RasPi4 has fixed that altogether. However, RasPi2 works fine through it's I2S interface to DAC boards like HiFiBerry. Some patches have added support for 352.8/384k rates on the I2S, but that in turn is not reliable and the DMA gets easily messed up resulting in channels being swapped, etc.

 

What kind of NAA device do you have and what are you running on it?

 

It's the Allo usbridge signature, the previous NAA was their normal usbridge.

[randytsuch]

28 minutes ago, Yviena said:

It's the Allo usbridge signature, the previous NAA was their normal usbridge.

 

You know about the sig dietpi issue with pops/clicks?  I had the problem at 256dsd with my dac, loaded latest version from allo, and do not update.

I also changed clock speed and deleted unused programs.

[randytsuch]

FYI, look at post #913 here

fixing pops

 

I wrote some detailed instructions on what I did, and it helped another person with the popping problem with a sig running NAA.

[3ggerhappy]

 

 

Recently I was not able to play tidal tracks in HQP(using mconnect lite), getting this error:

 

(process:2235): RygelRenderer-WARNING **: 14:18:50.288: Failed to access resource at http://192.168.1.104:12529/tidal/121210884.flac: Connection terminated unexpectedly

 

Using HQP embedded trial, anyone had similar issues? TIA

[ambre]

Hi Jussy, 

 

Since today I am  playing ITunes respectively Apple Music via HQP 4.1.1 by means of the new real time audio function of the HQP 4.1.1.                    A wonderful add-on.😁 

 

My setup is Apple iPad or Mac Book Pro sending via Airplay/Ethernet the music files  to an Airport express (with Optical Audio Out) and connected via a Toslink to a Mac Mini I7 (late 2012 ) with Optical input digital In. 

 

Sounds marvelous and a real extension of my music library.

 

Ps. I am able to upsample it to DSD 256 with " regular " NS/ Dither filters  and tot DSD 128 with the ASDM5EC or ASDM7EC.

 

Please have a look at the attached screen-copies.

 

 

[Miska]

2 hours ago, 3ggerhappy said:

Recently I was not able to play tidal tracks in HQP(using mconnect lite), getting this error:

 

(process:2235): RygelRenderer-WARNING **: 14:18:50.288: Failed to access resource at http://192.168.1.104:12529/tidal/121210884.flac: Connection terminated unexpectedly

 

Using HQP embedded trial, anyone had similar issues? TIA

 

Usually these kind of errors appear if the play queue is outdated (old URIs on the list). You could browse to the content again and attempt to initiate playback from there again.

[JTS]

Hi, I have just ordered an UPboard Gateway to use as an NAA as recommended on the Signalyst website. It should be arriving in the coming week. I am wondering if the HQplayer image can be installed onto the onboard eMMC storage opposed using a USB key. If so how might this be achieved? Do i need to load a live distro first and use some linux-fu to install the HQPlayer NAA image onto the onboard eMMC?

[Miska]

4 hours ago, JTS said:

Hi, I have just ordered an UPboard Gateway to use as an NAA as recommended on the Signalyst website. It should be arriving in the coming week. I am wondering if the HQplayer image can be installed onto the onboard eMMC storage opposed using a USB key. If so how might this be achieved? Do i need to load a live distro first and use some linux-fu to install the HQPlayer NAA image onto the onboard eMMC?

 

Yes, that's one way to do it, with either "bmap-tool" or "dd", of which bmap-tool is faster and safer. I have not tested it much, because all my units have Debian installed on the eMMC.

 

One convenient way to have NAA image in easy to update way is to put it on a microSD card and use a small microSD-USB adapter. But there are also some very small USB memory sticks that barely even come out of the port.

 

Since NAA image boots to RAM, if you want, you can safely remove the storage media once booted up.

 

[JTS]

2 hours ago, Miska said:

 

...there are also some very small USB memory sticks that barely even come out of the port.

 

 

I think this will be the way to go. As you mention, it'll make it very easy to update. It'll also make it easy to play a bit with DietPi, too. Thanks, Jussi.

[Yviena]

@Miskai noticed something weird, everytime i need to re detect naa, and play music my cpu usage is at 27% with the your latest build leading to stuttering  with multi core + cuda checked, but if i i do these steps

1.Exit  hqplayer, turn off corepinning via variable

2. Start hqplayer, and play music,

3. Exit hqplayer 

4. turn on corepinning

5. start hqplayer, and play music 

The cpu usage settles down to 22.5-24.5% and music plays without dropouts.

 

Any idea why doing these steps helps?

 

[Miska]

10 hours ago, Yviena said:

@Miskai noticed something weird, everytime i need to re detect naa, and play music my cpu usage is at 27% with the your latest build leading to stuttering  with multi core + cuda checked, but if i i do these steps

1.Exit  hqplayer, turn off corepinning via variable

2. Start hqplayer, and play music,

3. Exit hqplayer 

4. turn on corepinning

5. start hqplayer, and play music 

The cpu usage settles down to 22.5-24.5% and music plays without dropouts.

 

Any idea why doing these steps helps?

 

Just restarting HQPlayer doesn't do the same? Sounds crazy, since that procedure does essentially nothing from HQPlayer perspective. But it could be that the OS does something a bit differently after such procedure. Maybe that kicks some CPU core boosts a bit higher?

 

[Yviena]

7 hours ago, Miska said:

 

Just restarting HQPlayer doesn't do the same? Sounds crazy, since that procedure does essentially nothing from HQPlayer perspective. But it could be that the OS does something a bit differently after such procedure. Maybe that kicks some CPU core boosts a bit higher?

 

Yeah it does sound crazy but process explorer confirms it each time that CPU usage does actually go down from 26-27 to 22-24.5% hwinfo64 reports the same max boost frequencies though.

 

Though total CPU usage is still better on 4,1,0,1, and I don't need to do the procedure as above.

[Outlaw]

Latest release windows desktop find cpu % is lower.Still have issue when starting music have to hit play twice for music to play.