[BShaw]

Which is the reason I have never contemplated trying any DSD setting higher than 256 in HQPlayer. Until now. Which presents DSD at a level that can't quite match the transparency and snap of LAN PCM768. But now that I did go ahead and try it, each higher level of DSD rate sounds better and better all the way to 2048. even though the A26 display shows that it can't even show a value of input that is higher than DSD1024. The highest PCM value that is shown is 768 even though it will play higher rates.

I would now say that LAN DSD 2048 is the best that the A26 can sound. And I haven't even tested any of the higher performance high rate modulators yet. Still using ASDM7EC-super.

I don’t understand how dsd 2048 is delivering subjectively better sound than dsd 256, when now the A26 is downsampling to…. dsd256?

 

[EMINENT]

Which is the reason I have never contemplated trying any DSD setting higher than 256 in HQPlayer. Until now. Which presents DSD at a level that can't quite match the transparency and snap of LAN PCM768. But now that I did go ahead and try it, each higher level of DSD rate sounds better and better all the way to 2048. even though the A26 display shows that it can't even show a value of input that is higher than DSD1024. The highest PCM value that is shown is 768 even though it will play higher rates.

I would now say that LAN DSD 2048 is the best that the A26 can sound. And I haven't even tested any of the higher performance high rate modulators yet. Still using ASDM7EC-super.

The A26 is not capable of higher than 512. This is why the display shows 24.5M with 48k rate x 512 and if playing 44.1x512 would show 22.5M.
44.1x1024 = 45.1584M and a 48x1024 would show 49.152M.
The most the A26 can render are 512 rates.

 

[sendler2112]

I don’t understand how dsd 2048 is delivering subjectively better sound than dsd 256, when now the A26 is downsampling to…. dsd256?

Apparently the filter in HQPlayer works better during the conversion from PCM 24/ 48 to DSD 2048 than to DSD 256 even though the Gustard A26 display can't confess to receiving anything higher than LAN DSD512, and the dac chips then supposedly truncate anything higher to DSD256.

I would LOVE to set up a public ABX trial. I hear a slight difference between DSD 48x2048 and DSD 44.1x2048. 1024, 516, 256, they all sound different. DSD516 with the ASDM7EC-super 512+modulator is a big step up from DSD256 with the ASDM7EC-super.

Maybe next year at CanJam NYC? I appreciate that the designer is working hard and has been very generous with support on the other forums, but I assume they are making a good buck (a "finally" situation no doubt) at $300 a pop for an HQPlayer license. Maybe they will see the value in setting up a booth with Holo or somebody at the Canjams to get even more people hooked.

 

[HudsonJames]

Which is the reason I have never contemplated trying any DSD setting higher than 256 in HQPlayer. Until now. Which presents DSD at a level that can't quite match the transparency and snap of LAN PCM768. But now that I did go ahead and try it, each higher level of DSD rate sounds better and better all the way to 2048. even though the A26 display shows that it can't even show a value of input that is higher than DSD1024. The highest PCM value that is shown is 768 even though it will play higher rates.

I would now say that LAN DSD 2048 is the best that the A26 can sound. And I haven't even tested any of the higher performance high rate modulators yet. Still using ASDM7EC-super.

This isn’t how your dac works. It can’t play any higher rates other that DSD256 or PCM768. Those are its maximum limits.
The chip that processes the incoming signal downsamples it to DSD256. It can not do anymore.
Doesn’t matter what you are sending it.

Downsampling in HQplayer is one thing, but forcing your Dac to do this internally introduces a lot of negative effects.

 

[sendler2112]

That could be the case with A26, but not sure if same with R26.
R26 I found the U18 which worked well with X26 to sound thin with the same i2s in comparison to standalone USB or Lan. The U18 being not very good as discovered and sought to eliminate the extra box and gain galvanic isolation with the Holo.
X26/i2s/U18 I preferred PCM 7xx with sinc-Mx and LNS15 but could not test DSD as it was broken and hissed.
R26 preferred LAN DSD512 with ext2 and AMSDM7 512+fs.

So, I did try the PCM 7xx with LNS15 and sinc-Mx back and forth with TPDF and poly-sinc-gauss-xla. All 4 interchangeably. And still prefer p-s-g-xla with TPDF. I also prefer ASDM7 512+ to AMSDM7 512+or ASDM7EC-super. Buck convention and theory and try DSDxx at 2048.

 

[sendler2112]

This isn’t how your dac works. It can’t play any higher rates other that DSD256 or PCM768. Those are its maximum limits.
The chip that processes the incoming signal downsamples it to DSD256. It can not do anymore.
Doesn’t matter what you are sending it.

Downsampling in HQplayer is one thing, but forcing your Dac to do this internally introduces a lot of negative effects.

Did you try it?

 

[HudsonJames]

Did you try it?

There isn’t any good reason to try this.
The chip attempts to filter all the extra away to the dsd256 rate its hardware maxes out at.
You are just overloading the digital filter.
This kind of internal downsampling can cause all types of bad stuff, aliasing artifacts, quantization noise, ringing, blurred transients, etc etc.

 

[Hensema]

I hear a slight difference between DSD 48x2048 and DSD 44.1x2048. 1024, 516, 256, they all sound different. DSD516 with the ASDM7EC-super 512+modulator is a big step up from DSD256 with the ASDM7EC-super.

The big step you are hearing might be the 512+ modulator. Might be interesting to try it on DSD256. In my experience the 512+ modulators sound excellent at 256 rate.

 

[sendler2112]

There isn’t any good reason to try this.
The chip attempts to filter all the extra away to the dsd256 rate its hardware maxes out at.
You are just overloading the digital filter.
This kind of internal downsampling can cause all types of bad stuff, aliasing artifacts, quantization noise, ringing, blurred transients, etc etc.

Yeah. I almost made the same mistake and held off trying anything higher than DSD256 for several weeks based on comments like these. Relegating yourself to DSD256 is unfortunately noticeably inferior to PCM768. Which is bettered still slightly by DSD2048. There is obviously some advantage to the internal processing of the filter in HQPlayer that allows the most transparency when working at this rate, before it even leaves the PC. Or maybe some benefit to the transmission in any following digital cables. Or some advantage in the interaction with the onboard clock. Even if it will inevitably undergo a simple integer truncation in the AK4499EX chip set.

What do you have to lose in trying DSD2048 from HQPlayer with the Gustard A26? 20 minutes and 5 cents of electricity?

 

[HudsonJames]

Yeah. I almost made the same mistake and held off trying anything higher than DSD256 for several weeks based on comments like these. Relegating yourself to DSD256 is unfortunately noticeably inferior to PCM768. Which is bettered still slightly by DSD2048. There is obviously some advantage to the internal processing of the filter in HQPlayer that allows the most transparency when working at this rate, before it even leaves the PC. Or maybe some benefit to the transmission in any following digital cables. Or some advantage in the interaction with the onboard clock. Even if it will inevitably undergo a simple integer truncation in the AK4499EX chip set.

What do you have to lose in trying DSD2048 from HQPlayer with the Gustard A26? 20 minutes and 5 cents of electricity?

I think once you spend more time with your Dac and HQPlayer you’ll get a better understanding.
I do not share your view that PCM sounds better on the A26 for one.
I would triple check your settings and make sure you are doing a proper level matched test.
Even if your settings and levels are matched the chip still has a built in dsd reduction that must be compensated for for you to properly level match.

One of the reasons for this brilliant software is to remove all (or as much as possible) of the internal digital processing done by the dac.
What you are doing is not that. You are forcing the dac to downsample by overloading its filter.
I’m quite positive that a output measurements would show lots of nasty things this is doing.

Subjectively if you still think this sounds better more power to you. I’m not going to argue that. I love speed and going fast but I don’t mow my lawn at 70mph. Sure you can, I bet it’s fun, but I’m not going to tell others to try it.

 

[m-i-c-k-e-y]

Just an OT thought..

Thinking about upsampling to DSD2048 in feeding a DSD512 DAC.

This idea has the same principle as "Supersampling" for graphics in your computer monitor. Where images/video (esp in games) are upsampled and displayed on their current resolution. Try displaying a 1080p wallpaper in a 1080p monitor and the same wallpaper in 4K/8K. The later is better.

Experiment: Download these wallpapers in 1080p, 4K and 8K resolution. Compare images fullscreen (especially nearby bushes/trees). Later has more definition displaying on your screen.

NOT SAYING that upsampling DSD2048 to a DSD512 DAC is good (for now, to your DAC longterm). More experimentation is needed. But the idea is not new.

 

[LXMc]

I've got a question in regards to a slightly different HQPlayer use-case than most. Hoping someone here has some insight..

A while back I had asked about using HQPlayer in conjunction with Virtual Audio Cable (VAC) in order to route music from a non-Roon/Qobuz streaming platform (Deezer at the time) through HQPlayer. In that scenario I locked the VAC line-in to 16/44.1 since that's all Deezer was spitting out and all was good. Now that Tidal is getting rid of MQA and has bundled their hi-res files into their normal subscription I've been trying it out again but I had some questions regarding the possibility of my setup accommodating variable sample rates with Tidal's bit-perfect capability (something Deezer doesn't do).

As far as I can tell, VAC seems like it can handle the automatic changing of sample rates between 16/44.1 up to 24/192. When I enable cable range and channel mixing in VAC and I go to HQPlayer, it gives a spread of options that I have to select from. It seems like selecting the original 'audio:default/44100/2' works as it used to with 16/44.1 song, but as soon as Tidal switches to something higher it gives me an error (in tidal) telling me to switch it to 'high - 44.1' playback. I then tried changing HQPlayer to the maximum available input: 'audio:default/192000/2' and it appears to be allowing playback on all the songs I've tried that vary in sample rate. So technically it works - But I don't know if this is 'correct' or if HQPlayer is just being fed a 24/192 rate regardless or what Tidal is feeding VAC.

These are my VAC settings and how that comes through in HQPlayer:

When playing content, even though the song playing was 24/48 - HQPlayer appeared to just be receiving a straight 192khz rate so I'm assuming this is not working as I actually want and adjusting sample rate on the fly dependent on the song.

I realize this is probably a long-shot hoping someone has some insight as this likely not how most people use the software, but figured I'd ask. As a fallback I can always limit Tidal to 'high' 16/44.1 but it would be nice if I could utilize the wider range of sample rates available.

thanks!