HQ Player

[Miska]

6 hours ago, Yviena said:

@Miska So what are the new modulators/DSD filter for?

 

There were three SDM -> SDM conversion integrators and two rate conversion algorithms before. Now there are two more integrators for total of five options.

 

And one more noise filter (now total 9 options) for DSD -> PCM conversion.

 

[Miska]

5 hours ago, jcn3 said:

Is this built into HQP embedded already or is it coming in a future release?

 

It is in latest, and first version of it was added in 4.20.2.

 

[SwissBear]

On 1/23/2021 at 3:38 PM, Miska said:

It already has the same Crystek CCHD-957 clocks I also use (and referred to earlier), built in, so it should be perfectly fine without external clock:

https://shenzhenaudio.com/products/singxer-sda-6-primary-direct-solution-cchd-957-femtosecond-clock-of-usb-nos-model-dsd512-ak4499-dac-decoder

 

 

Hi Jussi,

 

You expressed your interest to see measurements that objectivise the improvements of 10 MHz clocks to the performance of a DAC. 

 

Here are the measurements made on the same DAC, a SDA-6 Pro by Singxer, without and with a 10 MHz clock. The improvement is significant, knowing that the last points of SINAD measurements are difficult to reach. Please note that both measurements were made at the same output voltage as this is susceptible to impact SINAD measurements.

 

Also, those measurements were made in PCM mode at 48 kHz sampling. So they do not reflect the impact of the NOS mode of this DAC in DSD, which @barrows highlighted here

 

Using your software to upsample to DSD and go through the DAC in NOS mode would be an additional way to improve its performance.

 

 

 

 

As far as J-test is concerned, here is a J-test of a similar DAC (not the same unit though). It is made without the 10 MHz clock.

 

 

As per the constructor, the J-test with the 10 MHz clock would show an improvement of the jitter in the near end of the spectrum, which is understandable as this is where those clocks are performing better than clocks usually used inside the DAC.

 

I will stop hijacking your thread and open a dedicated one for this DAC.

 

 

[Miska]

15 minutes ago, SwissBear said:

As far as J-test is concerned, here is a J-test of a similar DAC (not the same unit though). It is made without the 10 MHz clock.

 

 

As per the constructor, the J-test with the 10 MHz clock would show an improvement of the jitter in the near end of the spectrum, which is understandable as this is where those clocks are performing better than clocks usually used inside the DAC.

 

I will stop hijacking your thread and open a dedicated one for this DAC.

 

Now we need this made with 11.025 kHz tone and zoomed in correctly to +-4 kHz around the test tone. But above result is not very good. But not because of the built-in oscillators phase noise, but because there is some clock pollution (interference) inside the device. Which is unlikely to get better by use of external clock. There shouldn't be those side lobes around the 12 kHz peak.

 

With the CCHD-957 clocks the phase noise part of this plot should also look much better.

 

This plot looks more like having some PLL involved.

 

So as usual, the result above is not limited by the clock oscillators, but board design, PSU and the possible PLL involved.

 

[SwissBear]

30 minutes ago, Miska said:

 

Now we need this made with 11.025 kHz tone and zoomed in correctly to +-4 kHz around the test tone. But above result is not very good. But not because of the built-in oscillators phase noise, but because there is some clock pollution (interference) inside the device. Which is unlikely to get better by use of external clock. There shouldn't be those side lobes around the 12 kHz peak.

 

With the CCHD-957 clocks the phase noise part of this plot should also look much better.

 

This plot looks more like having some PLL involved.

 

So as usual, the result above is not limited by the clock oscillators, but board design, PSU and the possible PLL involved.

 

 

Is it the design of the DAC or the measurement conditions which are not really comparable to yours ? Especially the fact that you are measuring at DSD256 which is significantly improving the jitter ?

 

[Miska]

5 hours ago, SwissBear said:

 

Is it the design of the DAC or the measurement conditions which are not really comparable to yours ? Especially the fact that you are measuring at DSD256 which is significantly improving the jitter ?

 

 

Not sure, but this result is also worse than I get with my Spring 2, even in PCM mode. Only case why I could understand above result is running the May DAC at 48 kHz in NOS mode.

 

But in order to say anything about phase noise amounts that external clock is supposedly improving, you need to zoom in the frequency range. Because that is in the width of the central lobe base.

 

Here's Spring 2 running at 1.4112M sampling rate from HQPlayer, no NAA, from my Xeon workstation's regular USB connection:

 

Couple of  low level side lobes, but especially the phase noise is very low here too.

 

[Yviena]

@Miska Have you tried doing ryzen specific optimizations yet with your 5800x? 

 

Maybe using the cache, or something differently will increase performance, or even changing the compiler used.

[Miska]

12 minutes ago, Yviena said:

@Miska Have you tried doing ryzen specific optimizations yet with your 5800x? 

 

Maybe using the cache, or something differently will increase performance, or even changing the compiler used.

 

I will see for Linux version, maybe primarily for HQPlayer Embedded.

 

Have you tried HQPlayer Embedded on Fedora 33 Minimal (from the server installer)? And HQPlayer Desktop on Fedora 33 Workstation?

 

[jacquesr]

23 hours ago, Miska said:

There seems to be an occasional timeout problem when stopping playback, that sometimes makes Roon stop between trakcs. I will make a new build tomorrow to fix this.

 

Hello Jussi,

When I pause Roon for several minutes (not sure how many) then I have to turn my Aqua Linq on/off.

I suppose this is the issue you mentioned. Hope that helps.

[Miska]

1 minute ago, jacques_racine said:

When I pause Roon for several minutes (not sure how many) then I have to turn my Aqua Linq on/off.

I suppose this is the issue you mentioned. Hope that helps.

 

No, that shouldn't be necessary. Just restarting playback from Roon. So when the problem appears, Roon shows paused state or fails to start playback. Then you can just try to start the playback again and it proceeds. In worst case you can restart HQPlayer which will drop the NAA connection and establish a new one.

 

[toddrhodes]

I am experiencing the issue of playback completely stopping on track switch - so it plays fine, no issues for X tracks, then a new track is to begin and it just never starts playing.

 

Using latest Roon on a core server, HQP Desktop 4.9 on a secondary machine feeding latest HQP NAA Image (Pi4).

 

It was happening on the 4.8 build as well, so not sure if it's different or what, but just putting it out there. It's really only been an "issue" I've noticed while letting some new components play and warm up and such. Otherwise, that's not my typical use case to just let Roon Radio play and play and play.

[AudioDoctor]

Maybe I have missed it someplace, but is there a way to be alerted to updates to HQPlayer other than checking the website?

[toddrhodes]

1 minute ago, AudioDoctor said:

Maybe I have missed it someplace, but is there a way to be alerted to updates to HQPlayer other than checking the website?

 

I learned about it on Facebook, on the HQP "page." Not everyone does social media though so, not sure about other avenues.

[Miska]

1 hour ago, AudioDoctor said:

Maybe I have missed it someplace, but is there a way to be alerted to updates to HQPlayer other than checking the website?

 

Twitter and Facebook. So it is opt-in in this way...

 

There are links on the left pane here: https://www.signalyst.com/consumer.html

[Miska]

1 hour ago, toddrhodes said:

I am experiencing the issue of playback completely stopping on track switch - so it plays fine, no issues for X tracks, then a new track is to begin and it just never starts playing.

 

This problem was introduced in 4.9 with the NAA related changes, shouldn't be in 4.8. I've been trying to fix it, but it is hard to verify if a fix is working since it happens so infrequently... I have a new test build now and if it works without issues for me, I will release it shortly.

 

[AudioDoctor]

17 minutes ago, Miska said:

 

Twitter and Facebook. So it is opt-in in this way...

 

There are links on the left pane here: https://www.signalyst.com/consumer.html

 

Thanks Miska.

[ray-dude]

1 hour ago, Miska said:

Twitter and Facebook. So it is opt-in in this way...

 

 

I use https://visualping.io to monitor the website for changes, and email me when it does.

[Yviena]

7 hours ago, Miska said:

 

I will see for Linux version, maybe primarily for HQPlayer Embedded.

 

Have you tried HQPlayer Embedded on Fedora 33 Minimal (from the server installer)? And HQPlayer Desktop on Fedora 33 Workstation?

 

No I have not tried fedora, only Ubuntu with newest upstream kernel/updates.

[austinpop]

I am experiencing random NAA crashes with HQP 4.90. Restart of NAA fixes it.

[chipvn]

On 12/15/2020 at 9:55 AM, bibo01 said:

@Miskaand other users,

Do you have any experience or recommendations for nuc motherboards, like NUC10i7FNB, using i7-10710U cpu? What are the limits of such systems? Is it possible 44.1->DSD128 with any modulator? Possible ->DSD256?

Thanks  

Just to update you as well as those who might interested on small fanless NUC system.

I have tested before with the NUC8i5BEH in Akasa fanless case (with seem like to work well with total system power consumption not more than 25W) with the intention to find a compact fanless system to be used with a Raspberry Pi based NAA/Roon Bridge.

 

With HQPlayer desktop 4.5, this NUC could handle 44.1 to DSD128x44.1 with poly-sinc-ext2 / poly-sync-xtr-2s - ASDM7EC easily and total power consumption is less than 25W. It could handle 48/96 well also.

For 192 (the heaviest) to DSD128x44.1 with -sinc-ext2 / -sync-xtr-2s and ASDM7EC, the total power consumption might go to about 30W and the fanless case might not be able to coupe with on long run to keep the CPU away from thermal throttling (still can use without drop off / noise). To bring total power consumption down to around 25W with ASDM7EC and DSD128x44.1, I need to switch to none -2s (none -xtr) filter like poly-sync-mp-2s.

 

This NUC8i5 fanless could do ASDM7 with 44.1/48/96 to DSD256x44.1 easily with most common poly-sync filters included poly-sync-xtr and -sync-ext2. For ASDM7, to do 192 to DSD256x44.1 within 25W total power consumption, it is better not to use -sync-xtr or sync-ext2 cause total power consumption will be around 30W.

[sgb]

10 hours ago, Miska said:

 

Twitter and Facebook. So it is opt-in in this way...

 

There are links on the left pane here: https://www.signalyst.com/consumer.html

There are some of us (or am I the only one) HQPlayer users that don't use Twitter or Facebook. I am certinly not going to get a Twitter or Facebook account just to know when there is a new version of HQPlayer. 

[Jud]

4 hours ago, sgb said:

There are some of us (or am I the only one) HQPlayer users that don't use Twitter or Facebook. I am certinly not going to get a Twitter or Facebook account just to know when there is a new version of HQPlayer. 

Don’t know about your OS/distribution, but when I update the Linux distributions on which I run HQP, they tell me if there’s a new version of any installed program, including HQP, available. I look for any OS/app updates as a matter of routine anyway, so HQP updates come along for the ride.

[kelvinwsy]

3 minutes ago, Jud said:

Don’t know about your OS/distribution, but when I update the Linux distributions on which I run HQP, they tell me if there’s a new version of any installed program, including HQP, available. I look for any OS/app updates as a matter of routine anyway, so HQP updates come along for the ride.

I have Twitter but hate using it. Only spotted Jussi's twit that there is a verison 4.90 for Hqplayer 4 by chance9Email feed LOL)

After upgrading to ASDM EC 512, I needed to up the Turbo on my 10700k to 4.4 and 4.3 Ghz for the 1st 4 and then 2nd 4 cores respectively for a stutter free playback. The SQ jump is certainly worth the slight effort. The high frequencies are even more crystalline clear all the way to the Heavens! Gives a greater sense of detail, sound stage width, depth and separation. Highly recommended for all users to do the update

Enjoy the Music

Kelvin.

[asdf1000]

2 hours ago, kelvinwsy said:

After upgrading to ASDM EC 512

 

Say what now 😃

 

Are you the newly crowned King of HQPlayer

[Miska]

6 hours ago, sgb said:

There are some of us (or am I the only one) HQPlayer users that don't use Twitter or Facebook. I am certinly not going to get a Twitter or Facebook account just to know when there is a new version of HQPlayer. 

 

Front page on my web site is alternative solution.

 

Many people don't interactively use the application. In addition, many times the updates depend on third party hardware/software vendor (SGC, PinkFaun, Antipodes, iCat, Euphony). So it is not clear to make the application itself notify about updates.