[Funnychap]

Assuming you are taking the preamp variable out of the equation..

Venus II is more open/airy sounding, if you use Venus II in OS vs NOS.

If you add the Iris to the Venus II and use the I2S connection, the Venus II is more detailed and the difference between the two DAC's in terms of which one is more more open/airy sways even more in favor of the Venus II (as compared to Spring 3 KTE, via USB, in NOS).

This isn't to say Spring 3 KTE isn't detailed, open or airy- just a relative comparison.

Note I always used Spring 3 KTE via NOS and USB. I always used Venus II in OS via I2S through a Gaia or Iris DDC. So not quite apples to apples.

Which sound you prefer will depend on associated gear and subjective preferences. Both are terrific options.

EDIT: If you eliminate the DDC, and just compared these two DAC's on their own, I'd take the Spring 3 KTE over the Venus II.

Someday I would upgrade my DAC from Denafrips Ares II 12th and Denafrips Iris 12th (love R2R for live/realistic) so I appreciate all the comparisons between a Holo Audio and a higher Denafrips than my Ares II 12th+Iris 12th.
----
My setup:
After I upgraded my mixed use theater/music room's 2 front L/R speakers to GoldenEar BRX with the awesome AMT ribbon tweeter, these tweeters are soo revealing and realistic (gave me goosebumps) that it got me to upgrade the weaker links in my chain:
- Added a 3rd GoldenEar BRX as my center channel speaker.
- Added a DDC from Denafrips Iris 12th which improved the sound for my already great Denafrips Ares II 12th (a DAC with R2R technology for 3D live/realism)
- Eversolo DMP-A6 streamer with native Apple Music for hi-res lossless and ripping my CDs to it.

source=Apple Music (hires lossless, and spatial audio surround music) and Denafrips Ares II 12th dac (R2R technology for 3D live/realistic sound).

I get full range sound with my GoldenEar BRX by pairing them with my 2 sealed subwoofers.
Both music (via Eversolo DMP-A6) and theater (via Arcam AVR5) share the front speakers via a switchbox for RCA: Parasound 2125 powers the GoldenEar BRX speaker and oututs to two sealed subwoofers at 40 Hz crossover (each subwoofer has 1500 watts, 1 woofer, 2 bass radiators, can go down to 14 Hz).
So my front left and right speakers paired with the 2 subwoofers can play full range from 14 Hz to 35 Khz as if they were floorstanding speakers....perfect for music.

But movies lacked the thunderous room shaking bass for explosions, so my other room (family room) had 2 subwoofers and I moved one of them to the Home Theater for 7.3.4 and it's awesome.

 

[Patu]

My i7-11850H struggles at DSD1024 (breaks every 2 secs) but not problem at DSD512.
You would need something around i9-13900K or better but it could also help If you have a NVIDIA card, the HQPlayer can take advantage of it as well by enabling the option "CUDA-offload" relieving some load from the CPUs.

May I ask the exact settings you use in HQP for DSD512? Also is 11850H enough for DSD512x48? I’m in the process of upgrading my server machine to be able to upsample to DSD512 with a proper modulator and filters.

 

[BrainSalat]

May I ask the exact settings you use in HQP for DSD512? Also is 11850H enough for DSD512x48? I’m in the process of upgrading my server machine to be able to upsample to DSD512 with a proper modulator and filters.

I used these (see below), but I have to clarify this was more a CPU/GPU load test to give feedback to other user that had a similar CPU as mine, than trying to give some recommended settings. I don't use the HQP any longer after I have my Spring3 KTE. From while to while I give it a try to the HQP again to just remember why I don't need the HQP anymore. I also explained here https://www.head-fi.org/threads/holo-audio-spring-r2r-dac.810065/post-17644069 why the HQP doesn't have the same effect in R2R DACs as with Delta-Sigma DACs. With my previous DAC (ADI2-PRO), to me, the HQP was a great complement, but not with the Spring3.

 

[Patu]

I used these (see below), but I have to clarify this was more a CPU/GPU load test to give feedback to other user that had a similar CPU as mine, than trying to give some recommended settings. I don't use the HQP any longer after I have my Spring3 KTE. From while to while I give it a try to the HQP again to just remember why I don't need the HQP anymore. I also explained here https://www.head-fi.org/threads/holo-audio-spring-r2r-dac.810065/post-17644069 why the HQP doesn't have the same effect in R2R DACs as with Delta-Sigma DACs. With my previous DAC (ADI2-PRO), to me, the HQP was a great complement, but not with the Spring3.

Thank you for posting the settings.

Interesting view on HQP with Spring 3. I use the Spring 3 also and it has clearly superior sound quality when upsampling to DSD256+ than with PCM upsampling or NOS. HQP upsampling to DSD256+ gives you the best sound quality with Holo DACs and IMO there's no question about it. I've experimented very much with different options. Going NOS is the worst possible option with Holo DACs.

I also don't understand the view that R2R DACs wouldn't benefit of upsampling same way as Delta Sigma DACs. Even the HQP developer/owner Jussi Laako uses Spring 3 with HQP.

 

[borkenarrou]

Thank you for posting the settings.

Interesting view on HQP with Spring 3. I use the Spring 3 also and it has clearly superior sound quality when upsampling to DSD256+ than with PCM upsampling or NOS. HQP upsampling to DSD256+ gives you the best sound quality with Holo DACs and IMO there's no question about it. I've experimented very much with different options. Going NOS is the worst possible option with Holo DACs.

I also don't understand the view that R2R DACs wouldn't benefit of upsampling same way as Delta Sigma DACs. Even the HQP developer/owner Jussi Laako uses Spring 3 with HQP.

Interesting, will be great if you can share your HQPlayer setting, I tried PCM upsampling it was good enough for me, haven't tried DSD upsampling much as it make my PC sluggish, what is the CPU usage you are getting.

 

[Leto Dal]

Thank you for posting the settings.

Interesting view on HQP with Spring 3. I use the Spring 3 also and it has clearly superior sound quality when upsampling to DSD256+ than with PCM upsampling or NOS. HQP upsampling to DSD256+ gives you the best sound quality with Holo DACs and IMO there's no question about it. I've experimented very much with different options. Going NOS is the worst possible option with Holo DACs.

I also don't understand the view that R2R DACs wouldn't benefit of upsampling same way as Delta Sigma DACs. Even the HQP developer/owner Jussi Laako uses Spring 3 with HQP.

I genuinely prefer my May in NOS mode it just sounds smoother and more analog than anything I tried. But if you have a setting for HQ Player you would suggest, I am very curious to try

 

[Patu]

Interesting, will be great if you can share your HQPlayer setting, I tried PCM upsampling it was good enough for me, haven't tried DSD upsampling much as it make my PC sluggish, what is the CPU usage you are getting.

I upsample to DSD256(x48) currently. Settings I use:

Modulator: ASDM7EC-super
Filters: Oversampling 1x: poly-sinc-gauss-long (xl is great also) and Nx: poly-sinc-gauss-hires

Other options aren't that relevant but they've been configured like Jussi Laako instructed is the best way for Holo Spring 3. When upsampling to PCM, the DAC bits setting for Holo should be set to 20 bits btw.

My server cpu is i5-11400T which is much weaker CPU than your i7-11850H. You should easily be able to run DSD256 with the settings I described above. Even DSD512 with the better modulators. Modulator choice is at least as relevant as filter choice. I've found ASDM7EC-super to be the best and ASDM7ECv3 close second. These filters can be found in the latest version of HQP. Overall DSD (256+) has more refined, full, "analog" and three-dimensional sound than PCM, which in comparison can sound raw, edgier and flatter.

I genuinely prefer my May in NOS mode it just sounds smoother and more analog than anything I tried. But if you have a setting for HQ Player you would suggest, I am very curious to try

Please see above the settings I've found working best for me.

 

[Louisiana]

Interesting, will be great if you can share your HQPlayer setting, …

These are my settings for DSD and I love them:

With PCM I use poly sinc gauss xla, LNS15, max. PCM Rate, in your case 1536.

 

[BrainSalat]

Interesting view on HQP with Spring 3. I use the Spring 3 also and it has clearly superior sound quality when upsampling to DSD256+ than with PCM upsampling or NOS. HQP upsampling to DSD256+ gives you the best sound quality with Holo DACs and IMO there's no question about it. I've experimented very much with different options. Going NOS is the worst possible option with Holo DACs.

First, about the explanation of my link I have to clarify (I did it before as well after posting that) that my comments about the Spring3+HQP are only valid on the PCM upsampling context, where you are still using the Spring3's R2R ladder circuit workflow (same as for NOS). And in that case, we seems to agree (for you: equally worst, for me: no benefits, yeah, it sounds different, but better??? mmmm, so it is more up to your preferences).

About the HQP DSD upsampling workflow, you need to take in mind that the Spring3 has two separated circuits, one for PCM streams (the R2R ladder) and one for DSD Streams. For the second, HoloAudio, so far I know, has not publish how this work internally, but is not using/reusing a R2R ladder (like others R2R DACs than convert first from DSD-to-PCM to reuse thier R2R ladder), they have a novel circuit that transform directly from DSD to Analog using some kinds of resistor ladder construction (but I could guess this is not using oversampling either).....from their website:

Please note, the word ‘R2R’ is the name of a kind of architecture of resistor ladders. This architecture requires less resistors yet still is capable to deliver more than acceptable performance. Most DACs in the industry use this architecture. The DSD module of Spring3 does not use R2R architecture for the DSD, more specifically it uses a very specific architecture which is optimized to perform DSD to analog. So to be entirely accurate the DSD part of Spring is using resistor ladders, not R2R..

When I got my Spring3 I had the chance to test it side by side against my ADI-2 PRO and the Chord TT2 (loaned) for around 2 weeks. As expected, the TT2 and Spring3 were clearly superior, but while the TT2 and Spring3 shared several qualities (I liked both) there was one feature where the Spring3 had a notorious edge over the other contenders and it was the DSD native support. All my DSD albums sounded way better on the Spring3.

So here some final thoughts.... Is it really about HQP's magic PCM to DSD conversion? Then this could be verified on others R2R DACs......or is it about Spring3's magic DSD implementation? How about to convert PCM to DSD using other tools (foobar, Roon), that could be also an interesting test. I personally have not tried too much HQP DSD upsampling, will give it a try with your recommended settings....but so far I can tell you I am not happy with the noise the fans of my computer are doing now on my very quite listening room.

I also don't understand the view that R2R DACs wouldn't benefit of upsampling same way as Delta Sigma DACs.

What did you don't understand from my link? In simple words. Delta-Sigma DACs require to perform oversampling (always, it's part of their pipeline) and the HQP can do that better using state of the art oversampling filters (a clear benefit). But oversampling is not part of R2R DACs, it's an optional preprocessing of the input stream. About the benefits/drawback of this preprocessing for the R2R ladder I won't be so sure.

Even the HQP developer/owner Jussi Laako uses Spring 3 with HQP.

How convenient

 

[anjocaido]

First, about the explanation of my link I have to clarify (I did it before as well after posting that) that my comments about the Spring3+HQP are only valid on the PCM upsampling context, where you are still using the Spring3's R2R ladder circuit workflow (same as for NOS). And in that case, we seems to agree (for you: equally worst, for me: no benefits, yeah, it sounds different, but better??? mmmm, so it is more up to your preferences).

About the HQP DSD upsampling workflow, you need to take in mind that the Spring3 has two separated circuits, one for PCM streams (the R2R ladder) and one for DSD Streams. For the second, HoloAudio, so far I know, has not publish how this work internally, but is not using/reusing a R2R ladder (like others R2R DACs than convert first from DSD-to-PCM to reuse thier R2R ladder), they have a novel circuit that transform directly from DSD to Analog using some kinds of resistor ladder construction (but I could guess this is not using oversampling either).....from their website:


When I got my Spring3 I had the chance to test it side by side against my ADI-2 PRO and the Chord TT2 (loaned) for around 2 weeks. As expected, the TT2 and Spring3 were clearly superior, but while the TT2 and Spring3 shared several qualities (I liked both) there was one feature where the Spring3 had a notorious edge over the other contenders and it was the DSD native support. All my DSD albums sounded way better on the Spring3.

So here some final thoughts.... Is it really about HQP's magic PCM to DSD conversion? Then this could be verified on others R2R DACs......or is it about Spring3's magic DSD implementation? How about to convert PCM to DSD using other tools (foobar, Roon), that could be also an interesting test. I personally have not tried too much HQP DSD upsampling, will give it a try with your recommended settings....but so far I can tell you I am not happy with the noise the fans of my computer are doing now on my very quite listening room.


What did you don't understand from my link? In simple words. Delta-Sigma DACs require to perform oversampling (always, it's part of their pipeline) and the HQP can do that better using state of the art oversampling filters (a clear benefit). But oversampling is not part of R2R DACs, it's an optional preprocessing of the input stream. About the benefits/drawback of this preprocessing for the R2R ladder I won't be so sure.


How convenient

I've already tested HQPlayer with Spring3 but I prefer to use jplay, the sound in my opinion and ears , is more natural.
I've also tested jriver and foobar and none seemed better than jplay for playing music. I think that the fact that jplay turns off most of the Windows functions greatly benefits the sound quality.
One detail that can help a lot is building a Fanless PC like I did. There are no noises in the circuit or in the room.

 

[Patu]

First, about the explanation of my link I have to clarify (I did it before as well after posting that) that my comments about the Spring3+HQP are only valid on the PCM upsampling context, where you are still using the Spring3's R2R ladder circuit workflow (same as for NOS). And in that case, we seems to agree (for you: equally worst, for me: no benefits, yeah, it sounds different, but better??? mmmm, so it is more up to your preferences).

PCM upsampling does sound better than NOS though, but it's worse than DSD.

About the HQP DSD upsampling workflow, you need to take in mind that the Spring3 has two separated circuits, one for PCM streams (the R2R ladder) and one for DSD Streams. For the second, HoloAudio, so far I know, has not publish how this work internally, but is not using/reusing a R2R ladder (like others R2R DACs than convert first from DSD-to-PCM to reuse thier R2R ladder), they have a novel circuit that transform directly from DSD to Analog using some kinds of resistor ladder construction (but I could guess this is not using oversampling either).....from their website:

Yes I know this. I don't really care how they do it if it just sounds better. Of course it feels little bit silly to not use the R2R part of the DAC since that's kind of the thing with Holo but I'm sure they know what they're doing with the DSD section. Of course the R2R side is in use when I use my setup with TV and movies.

So here some final thoughts.... Is it really about HQP's magic PCM to DSD conversion? Then this could be verified on others R2R DACs......or is it about Spring3's magic DSD implementation? How about to convert PCM to DSD using other tools (foobar, Roon), that could be also an interesting test. I personally have not tried too much HQP DSD upsampling, will give it a try with your recommended settings....but so far I can tell you I am not happy with the noise the fans of my computer are doing now on my very quite listening room.

Probably both. I first upsampled to DSD256 with Roon's built-in DSP. Then I installed HQP trial version and did some thorough testing. It didn't take long to realize that HQP does the upsampling way better than Roon. The sound quality is much better.

I have my server machine in another room and it's pretty quiet even when stressed. I stream to Holo Red which sits next to the Spring 3. So the fan noise is not an issue for me.

What did you don't understand from my link? In simple words. Delta-Sigma DACs require to perform oversampling (always, it's part of their pipeline) and the HQP can do that better using state of the art oversampling filters (a clear benefit). But oversampling is not part of R2R DACs, it's an optional preprocessing of the input stream. About the benefits/drawback of this preprocessing for the R2R ladder I won't be so sure.


How convenient

Well it seems that you have mixed the reason why upsampling is the thing with NOS DACs. The whole idea of doing the upsampling before the signal enters the DAC, is that the NOS DAC itself is not doing it at all. So to achieve the best possible sound quality (objectively and IMO subjectively also) you need to do the upsampling with a software like HQP or Roon DSP. With a regular delta-sigma DACs the signal is always processed by the DAC anyway, so it's less beneficial to do any processing before the signal enters the DAC. However, even upsampling DACs do benefit of HQP processing, but not as much as NOS DACs like Holo.

I have AKM4493 based DAC (FiiO K7) also and while upsampling to DSD256 with HQP does improve the sound quality, it's not as obvious as it is with Holo Spring 3.

I'm missing the point of "How convenient" comment. Jussi Laako has a list of recommended hardware to be used with HQP on Signalyst website: https://www.signalyst.com/consumer.html

It lists the DACs which support "native DSD" or "direct DSD" (I'm not sure of the correct terminology here) and give you the best results when used with HQP. Holo DACs fit the bill.

 

[BrainSalat]

PCM upsampling does sound better than NOS though, but it's worse than DSD.

"does sound better"... whatever it does mean is your own subjective opinion and it's all right, we all have different ears and preferences. Everyone should do what it sounds better for you. I prefer my Spring3 on NOS, because "to me" it sounds more natural (or less digital) and I don't perceive any benefits, like more clarity or better layering, etc....like I perceived from my previous delta-sigma DACs with HQP.

Now that I have tried a bit the DSD up-sampling with your recommended settings (thanks for that) I have to say I noticed some differences and I liked it in some cases (electronics), like an increased resolution and more holographic extended soundscape, but in other cases I feel the tonal balance of NOS is more accurate (acoustic music) and the soundstage was more realistic (like there was not need to make it bigger). Using DSD up-sampling kind of make me remember the DNA of delta-sigma DACs (less analog). I think depending on the mood and music styles I could choice between NOS and DSD. On the other hand I don't like my laptop making that fan noise.

Yes I know this. I don't really care how they do it if it just sounds better. Of course it feels little bit silly to not use the R2R part of the DAC since that's kind of the thing with Holo but I'm sure they know what they're doing with the DSD section. Of course the R2R side is in use when I use my setup with TV and movies.

I completely agree, it sounds crazy that for getting the best of your R2R DAC, don't use your R2R DAC....but remember my first comment was referring to PCM upsampling only ( I know i was not clear enough).

Probably both. I first upsampled to DSD256 with Roon's built-in DSP. Then I installed HQP trial version and did some thorough testing. It didn't take long to realize that HQP does the upsampling way better than Roon. The sound quality is much better.

Yes also completely agree, HQP is the best up-sampler out there, but the thing I wanted to compare was NOS vs DSD (by any other software conversion, Roon, foobar, etc) to really compare the two internal circuits of the Spring3 (PCM vs DSD), but probably is not easy to make an apple vs apple comparison.

I have my server machine in another room and it's pretty quiet even when stressed. I stream to Holo Red which sits next to the Spring 3. So the fan noise is not an issue for me.

Cool, that would be optimal. I use the Spring3 for headphone listening which I feed with my laptop. A dedicated server machine would be to overkill for my use case.

Well it seems that you have mixed the reason why upsampling is the thing with NOS DACs. The whole idea of doing the upsampling before the signal enters the DAC, is that the NOS DAC itself is not doing it at all. So to achieve the best possible sound quality (objectively and IMO subjectively also) you need to do the upsampling with a software like HQP or Roon DSP.

That's exactly what I explained in my link and on "simple words". . BTW there are also hardware solutions like the M-Scaler (dedicated) and Chord DACs.

With a regular delta-sigma DACs the signal is always processed by the DAC anyway, so it's less beneficial to do any processing before the signal enters the DAC. However, even upsampling DACs do benefit of HQP processing, but not as much as NOS DACs like Holo.

That's not true. Delta-Sigma DACs always up-sample the input stream to a fixed rate (the maximal rate supported by the DAC), at least the input stream already come at that sample rate, in which case the internal up-sampler get bypassed. That's why you want to set HQP's sample rate to the maximal supported by your DAC. And that is exactly the reason why delta-sigma DACs get so benefited from HQP in contrast to R2R DACs.

I have AKM4493 based DAC (FiiO K7) also and while upsampling to DSD256 with HQP does improve the sound quality, it's not as obvious as it is with Holo Spring 3.

also a delta-sigma DAC, agree from beginning!

I'm missing the point of "How convenient" comment. Jussi Laako has a list of recommended hardware to be used with HQP on Signalyst website: https://www.signalyst.com/consumer.html

It lists the DACs which support "native DSD" or "direct DSD" (I'm not sure of the correct terminology here) and give you the best results when used with HQP. Holo DACs fit the bill.

nnnaa....just joking!. because you were quoting a comment of the owner of the product he is promoting!

 

[Patu]

"does sound better"... whatever it does mean is your own subjective opinion and it's all right, we all have different ears and preferences. Everyone should do what it sounds better for you. I prefer my Spring3 on NOS, because "to me" it sounds more natural (or less digital) and I don't perceive any benefits, like more clarity or better layering, etc....like I perceived from my previous delta-sigma DACs with HQP.

Of course it's just my opinion. I don't feel the need to write "IMO" in every other sentence.

Now that I have tried a bit the DSD up-sampling with your recommended settings (thanks for that) I have to say I noticed some differences and I liked it in some cases (electronics), like an increased resolution and more holographic extended soundscape, but in other cases I feel the tonal balance of NOS is more accurate (acoustic music) and the soundstage was more realistic (like there was not need to make it bigger). Using DSD up-sampling kind of make me remember the DNA of delta-sigma DACs (less analog). I think depending on the mood and music styles I could choice between NOS and DSD. On the other hand I don't like my laptop making that fan noise.

If analog sound, to you, means less musical information available, harder to follow music, softer/rounder passages and so on then yes, I understand the preference for NOS. I even tried the NOS mode once again, inspired by this conversation. It's very easy to listen to and I understand the fascination towards it. With HQP DSD upsampling though, I find more open, spacious and holographic soundstage. The smooth and enjoyable overall balance is still there but with better resolution and separation of different instruments. Bass sounds bigger and hits harder. Music is just easier to follow.

I completely agree, it sounds crazy that for getting the best of your R2R DAC, don't use your R2R DAC....but remember my first comment was referring to PCM upsampling only ( I know i was not clear enough).

Yeah no problem.

That's not true. Delta-Sigma DACs always up-sample the input stream to a fixed rate (the maximal rate supported by the DAC), at least the input stream already come at that sample rate, in which case the internal up-sampler get bypassed. That's why you want to set HQP's sample rate to the maximal supported by your DAC. And that is exactly the reason why delta-sigma DACs get so benefited from HQP in contrast to R2R DACs.

Yes you're correct, you should upsample to the highest supported rate before feeding the signal to the DAC (with delta-sigmas).

In the end, I guess we just have to disagree on the subject. Developer of HQP also disagrees with you on this subject.

 

[BrainSalat]

If analog sound, to you, means less musical information available, harder to follow music, softer/rounder passages and so on then yes, I understand the preference for NOS.

We definitively have different perceptions on how the Spring3 sounds, I absolutely love how my Spring3 sounds in NOS mode and I think Jeff Zhu and his team has made a stunning work with the Spring3 and May. And by analog sound I meant the sound of Vynil and Ree-to-Reel. To me, Holo DACs (and probably others R2R DACs as well) is what get closer to the previous two without the drawback of these technologies. It's like Hires audio and Vynil would have conceived a baby.

In the end, I guess we just have to disagree on the subject. Developer of HQP also disagrees with you on this subject.

I also would invite you to read what Kitsune thinks about this subject: https://www.kitsunehifi.com/nosvsos/.
Did you know Kitsune removed OS from Spring3 (it was present in Spring2 and May). But if you like OS then I totally agree HQP is the best way to go.

 

[Patu]

We definitively have different perceptions on how the Spring3 sounds, I absolutely love how my Spring3 sounds in NOS mode and I think Jeff Zhu and his team has made a stunning work with the Spring3 and May. And by analog sound I meant the sound of Vynil and Ree-to-Reel. To me, Holo DACs (and probably others R2R DACs as well) is what get closer to the previous two without the drawback of these technologies. It's like Hires audio and Vynil would have conceived a baby.


I also would invite you to read what Kitsune thinks about this subject: https://www.kitsunehifi.com/nosvsos/.
Did you know Kitsune removed OS from Spring3 (it was present in Spring2 and May). But if you like OS then I totally agree HQP is the best way to go.

It isn’t just OS vs NOS. You can do oversampling/upsampling in so many ways with different software and settings. There’s also the topic of PCM vs DSD, two different sound signatures completely. The upsampling which is done with a regular integrated DAC chip is way inferior to software based (HQP) upsampling or something like M-Scaler. So even with Spring 2 it is beneficial to use something like HQP instead of the integrated upsampling option. To my knowledge, May still has the upsamplinh option(?)

But yeah it’s pointless to debate preferences. NOS sounds ok to me also but IMO you lose much of the potential.