HOLO Audio MAY DAC

[KenMoreira]

1 minute ago, toddrhodes said:

 

I just use the same filter in both areas. 

 

I will say - when you're doing DSD256 - Sinc L is using a ratio based on what you're converting from and to. And that conversion ratio generally results in a LOT of taps. I don't know the exact math, but I gave up on that thing a long time ago. I think Sinc Mx does a great job filling the gap with it, anyway.

OK I'll try Sinc MX instead of SincM.  Are you guys isn't the standard ec7 or the longer named one? Don't have it loaded up to see off hand 

[KenMoreira]

These seem to be the best I can do with my i5 for now. But it sounds very open, and detailed and natural so thanks 

[KenMoreira]

What is 48K DSD.  You both have it checked 

[KenMoreira]

9 minutes ago, toddrhodes said:

For both DF and Holo DACs (and probably others, these are just the ones I'm familiar with), they will accept DSD at 48k rates, which I believe, in conjunction with "Adaptive Output Rate," makes for an easier load on HQPlayer to convert sample rates. I would be willing to bet that you'll see better performance if you check or gray Adaptive Output Rate and check 48k DSD, then set your DSD rate to 48k 256.

 

Basically what you're telling HQPlayer right now is that if you are playing 24/96, then it should resample to multiples of 44.1 (but in the DSD range, which is much higher but I'm lazy and don't feel like doing the conversion math right now) only. But really, it should be able to do 44.1 rate families in even multiples, and 48k families in even multiples (96, 192, 384 - but again in DSD rates). Checking Adaptive Output Rate allows for this "keep it in the family" feature. There is nothing to be gained by taking 88.2 and converting it to 96, or vice versa - just more work for the machine which might just add unnecessary noise.

 

This reduces the work your machine has to do to resample one rate family to another, which your current settings are forcing it to do :)

 

And without 48k DSD checked, I don't think it will even try to use that rate family for DSD conversion since not all DACs support it. But ours and yours do.

That might explain why some stuff doesn't play when adaptive is on..  You need both setting working together.. 

 

I appreciate the knowledge.  So I'll check that off, and check adaptive and grey out multi 

[KenMoreira]

7 minutes ago, 87mpi said:

I'm a DSD fan too, in fact I built an i9 9900k PC just for that, but For me the best SQ on S3 KTE is definitely the PCM 1536 Sinc-M LNS15.  The sound is more palpable and ALIVE, more 3D. DSD in comparison is more 2D to me, even though it has more black. For the DSD my favorite SETTING is ASDM7EC 256 and  gauss long. I use the direct USB connection from the motherboard (asrock z390m-itx phantom) with Sablon 2020 cable.

I would like to upgrade to a jcat or pink faun usb pcie card with external linear power supply.

Agreed. I find for some recordings like hotel California.  The blackness is preferred but for other albums pcm 

[KenMoreira]

1 hour ago, toddrhodes said:

For both DF and Holo DACs (and probably others, these are just the ones I'm familiar with), they will accept DSD at 48k rates, which I believe, in conjunction with "Adaptive Output Rate," makes for an easier load on HQPlayer to convert sample rates. I would be willing to bet that you'll see better performance if you check or gray Adaptive Output Rate and check 48k DSD, then set your DSD rate to 48k 256.

 

Basically what you're telling HQPlayer right now is that if you are playing 24/96, then it should resample to multiples of 44.1 (but in the DSD range, which is much higher but I'm lazy and don't feel like doing the conversion math right now) only. But really, it should be able to do 44.1 rate families in even multiples, and 48k families in even multiples (96, 192, 384 - but again in DSD rates). Checking Adaptive Output Rate allows for this "keep it in the family" feature. There is nothing to be gained by taking 88.2 and converting it to 96, or vice versa - just more work for the machine which might just add unnecessary noise.

 

This reduces the work your machine has to do to resample one rate family to another, which your current settings are forcing it to do :)

 

And without 48k DSD checked, I don't think it will even try to use that rate family for DSD conversion since not all DACs support it. But ours and yours do.

That might explain why some stuff doesn't play when adaptive is on..  You need both setting working together.. 

 

I appreciate the knowledge.  So I'll check that off, and check adaptive and grey out multi 

[KenMoreira]

I have the same, but I haven't used it since I got the spring 3 kte

[KenMoreira]

9 hours ago, toddrhodes said:

For both DF and Holo DACs (and probably others, these are just the ones I'm familiar with), they will accept DSD at 48k rates, which I believe, in conjunction with "Adaptive Output Rate," makes for an easier load on HQPlayer to convert sample rates. I would be willing to bet that you'll see better performance if you check or gray Adaptive Output Rate and check 48k DSD, then set your DSD rate to 48k 256.

 

Basically what you're telling HQPlayer right now is that if you are playing 24/96, then it should resample to multiples of 44.1 (but in the DSD range, which is much higher but I'm lazy and don't feel like doing the conversion math right now) only. But really, it should be able to do 44.1 rate families in even multiples, and 48k families in even multiples (96, 192, 384 - but again in DSD rates). Checking Adaptive Output Rate allows for this "keep it in the family" feature. There is nothing to be gained by taking 88.2 and converting it to 96, or vice versa - just more work for the machine which might just add unnecessary noise.

 

This reduces the work your machine has to do to resample one rate family to another, which your current settings are forcing it to do :)

 

And without 48k DSD checked, I don't think it will even try to use that rate family for DSD conversion since not all DACs support it. But ours and yours do.

well my cpu load dropped to mid 60s in usage. but i still get drop outs at 256x so i guess ill drop to 128

[KenMoreira]

Any one here use dolby atmos for headphones? 

[KenMoreira]

22 minutes ago, toddrhodes said:

 

 

My May is about 4 weeks old now I think, and I've got probably 400+ hours of play on it, 100 hours of listening or thereabouts (totally guessing here as that seems real high now that I type it out) - IMO, it does get better with age. Better bass and improvements in imaging precision, meaning room cues and spatial cues, are what stand out to me. I've only used mine in NOS mode with HQP providing DSD256 to it. I've tried 32x PCM but just haven't locked into it yet. More on that in a sec. But to me, I didn't want to believe it and it DOES sound excellent out of the box. But whether it's acclimation or it "burns in," I have no idea, I just know I appreciate it now - I'm going to say "far more" - than right away. And that's saying a lot because right out of the box it presented an improved experience over my Denafrips Venus II it replaces.

 

More on that last point - I've fought this urge real hard, but I've come to the conclusion that going straight out of my USB port on my i9900KF HQ Player PC... is just as good as what I've become accustomed to using my Zen Stream as an NAA. I still intend to make the PC USB output "better" and try an even better USB cable than I have, but I'm simplifying my setup and now I'm just getting rid of the NAA entirely. I might go HQP Embedded and run Linux on this machine eventually, I'm not sure. But I can say the Titanis USB input is legit. I've tried it a few times and thought "Nah, it just can't be as good" and I'd convince myself it wasn't. But really, it is. And this is the same thing being said for Spring 3 L3 with the same input as well. My May does have the "enhanced" USB input but I'm using an Intel chip so it technically shouldn't matter in my setup.

 

I now have a Curious Evolved USB cable coming to try out, and have moved May KTE back out of my room and it sits right next to the HQ Player PC. In terms of preferring DSD over PCM - I do get what others say about PCM being more "exciting" but I prefer the timbre and tone density of 256DSD upsampled with the ASDM7EC modulator - gauss-long fitter. 

 

In fact I like this whole combo so much, I'm going to sell off the rest of my vinyl that isn't like an heirloom treasure to me, as I honestly can't see playing it anytime soon. No need.

Burn in on my spring was signicant. And the only gear I've noticed it happen on.  

[KenMoreira]

53 minutes ago, wanta911 said:

 

Opinions can only be based on personal experience and in my case - count me in as a semi-sceptic on the issue of burn-in. I don't discount the premise out of hand but I do question the number of hours that sometimes get quoted and at what point the burn-in is psychological versus mechanical. I read on Audiogon where someone said that their Holo DAC went to "another level" @ 3,000 hours of playing time.....it's posts like this that give this hobby a bad name.

 

Burn-in is also used by fanbois on forums to try and discount another person's dislike of a particular piece of equipment that they themselves DO like or recommended: "oh you need to burn it in for 500 hours and then it will all fall into place". I bought the original Auralic Vega DAC when it first came out and in my particular setup at the time I found it fatiguing after an hour or so each time I used it. Forum advice went from 250 to 500 to 750 to 1000 hours of burn-in each time I noted my dissatisfaction. After 1000 hours, it sounded exactly the same as it did within the first 2 hours - like a pencil being pushed into my ear drum 🤣. In hindsight I can understand that system matching and/or the infamous ESS Sabre glare was to blame (yes I do subscribe to that in older applications).

 

I am not discounting the argument for burn-in on an R2R DAC as it makes some sense, at least compared to a Delta Sigma application and I am not discounting out of hand some of the impressions here, as they at least all seem to be around the same time frame. The logical side of me questions exactly what is burning in after say 50 or even 100 hours and it will be interesting to experience it for myself when my Spring arrives later this week 🤙

Burn in on these is real. They start out bright with almost no bass at all. You will see. 

[KenMoreira]

4 minutes ago, Phil_Bog said:

 

In that case, can you think of anything that could explain or solve the harshness I'm experiencing? 

Burn in.  The dac is siblant in symbals etc at first 

[KenMoreira]

7 minutes ago, camott said:

 

For whatever reason, I have found plugging the May's power supply directly into the wall sounds better than through a power conditioner.

 

 

With HQPlayer server directly connected to the DAC maybe this is true .... but using an HQPlayer NAA endpoint with an LPS is about as low power/low noise as is possible. All it does is relay the stream from remote server, be it 44/48 or heavily upsampled. I use a low power Fitlet2 with optical input to isolate any ethernet noise. With an LPS of course.

 

On 9/24/2021 at 7:39 AM, 87mpi said:

Yes, 256 and 7EC modulator. 

 

I come from an older 3 Pc networked system, a nuc i5 for roon server, i9 9900k for HQP, nuc i3 fanless as endpoint NAA with LPS Keces P8, another LPSs on router and switch. 

 

I removed everything when I heard the difference compared to the PC i9 9900k plugged directly into USB in the Dac. 

Sounds noticeably better.

 

May/Spring 3 KTE USB input is truly works wonders! 

 

Now I use GentooPlayer OS (Linux OS with a myriad of options and customizations.  You can simply activate or deactivate different servers like HQP desktop or Embedded, LMS, Roon, mpd, different players, kernels, ram system etc .. You can really do whatever you want, the.  All with a spectacular and simple web interface, unlike Audiolinux for example) and only HQP embedded  and olny local files, zero streaming service. I use the network only to control the player from the tablet with the HQPcontrol v4 app.  Once in play you can also unplug the network cable, so finally zero network influence! 

 

I will not go back for sure! 

 

On 9/24/2021 at 8:06 AM, Zauurx said:

 

I am taking the opportunity of this post to... do a little disgression.
I don't have a May but a Pegasus... no Zen Stream but an Allo USBRidge SIG with which I can switch from PCM 1.5
However I am interested in this Zen Stream to know if it could be better than the Allo (+ Shanti).
What power supply do you use it with?
The USBridge with GentooPlayer is excellent but this is true for many equipment until you compare them.
I use 100% HQP for DSD 256 / sync.gauss.long (or xla) and ASDM7EC (core i5 8400 / Win10 LTSC).

USB is silent to 170 dB. No jitter or noise.. So you must be listening louder than a jet engine to hear noise at all and need that isolation from Ethernet 

[KenMoreira]

1 minute ago, Extreme_Boky said:

I also like plugging the power cables straight into the outlets.

 

You found the solution that works for you, which is good.

 

The only time I can hear the clear depth perception in sound, as well as the completely natural harmonic presentation that I strive to have (same as vinyl), is if I do not oversample at all. Sometimes (with some 44.1kHz material) I choose DSD64fs... or 88.1/96kHz. 

Dac, amp or both?  I understand an amp may need more on demand current. Curious how a DAC is pulling so much.  Not impossible just curious maybe 

[KenMoreira]

16 minutes ago, Extreme_Boky said:

May DAC has class A output stage. 60W is a lot... but it is a part of the design to keep R-2R and especially oscillators warm (sweet spot for the lowest phase noise). Also, class A circuitry has low distortion and it can be tuned to produce a pleasant H2 harmonic distortion spectrum (well know to valve-gear lovers)

 

The power cables I use are all solid core, mostly pure copper... but also a combination of copper ribbons and annealed silver ribbons. As my system became more revealing, I completely removed silver plated stuff (I had some silver and then gold plated solid copper cables as well... long time ago).

 

Amplifiers should not be demanding/causing the current surges directly from mains supply (except when you turn them ON), if their power supply sections are design correctly. 

 

 

 

 

I use a cheap furman conditioner.  I was just wondering why some people prefer dacs and amps out of the wall 

[KenMoreira]

Shit yeah, with how much stuff costs in this hobby. 

That's an easy buy 

[KenMoreira]

Trit

5 minutes ago, Extreme_Boky said:

It is to do with how each HiFi audio unit interferes with others, predominantly via the same, ground wire - back towards the main junction point (where your power meter is...and then further back towards the transformer station)  

 

... if a same conditioner box is used, then a single ground wire (from that box) will be feeding the cumulative rubbish back towards the main junction point.

 

However, if you choose a separate power outlets in the wall, the amount of ground feed will double. In this case, the more - the merrier. 

 

Also, the ground leads should be large in diameter, and the lead should be multistrands... So, in this case, the solid core (for ground wire) is not a good option.

 

 

 

Truthfully don't know if I ever have noticed a difference or looked for one 

[KenMoreira]

2 hours ago, Miska said:

 

At least technically, DSD side of Spring 3 is at least as good as the PCM side...

 

Waiting for a but.. 

[KenMoreira]

Do any of you ever use nos lol

[KenMoreira]

2 minutes ago, Ozan Bolat said:

what are your favorite filters ? I reassessed PCM on my Spring 3 after burning it with a feed of SDM exclusively and to my surprise I love PCM too now while out of the box SDM sounded much better. If there's a rational different from a change of my mood, I'd point that SDM is more maths inside the computer, less physicality inside the Holo

What is sdm

[KenMoreira]

Is it still worth it to use the higher dsd sample rates now that theirs a 20kHz filter?  Or with that eliminating any 'measurable' noise being omitted... Should i lower my settings and try for L sync? 

[KenMoreira]

On 6/20/2022 at 8:26 PM, musicman1 said:

I finally was able to use DSD 1024X with various Modulators such as AMSDM7 512+fs without any issues. It still appears I may need to upgrade my CPU if I want to use the EC modulators, however I haven't tried them all. I'm also using convolution filters from Accurate Sound so I'm sure that's not helping things. One thing I've noticed using Roon and HQPlayer is sometimes things are very finnicky and I have to play with things to get everything to work, regardless if I previously changed anything or not.

 

I was wondering if there should be a specific order on turning on hardware and starting software. I usually turn on in this order - Headphone amp - Holo May DAC - HQPlayer - Roon. Is there a better order to do this? Also I have ordered the Holo Audio Serene KTE with hopes it can bring my Audio levels up a bit considering I'm losing several dB by upsampling to DSD. 

Guessing you have hefty CPU just to do what you are. 

 

I wonder about dual socket server boards lol

[KenMoreira]

46 minutes ago, ArcticSapien said:

@barrows I was driven up the wall and gave up some months back, I'm now restarting my problem solving. I've narrowed it down to:

* unmanaged switch? Thought some other forums identified this as an issue. 

 

* Sonore opticalRendu? Read from other pages that the oR has limitations. I see that you're using the oM+SigRendu without problems. 

 

* network issues... but the dropouts go away when I use less demanding modulators, so it doesn't seem to be a network problem.

But I use Google Nest WiFi to connect to the switch>oR>Holo May DAC (yes bummer... I wasn't able to run a fibre cable across to my music room). So, this could all be resolved with a direct cable link!

 

Gotta try eliminate these possible causes one by one! 

You could consider "power line" of course using that will induce noise.  How much is and weather its audible who knows. 
but at least You would be wired. 

[KenMoreira]

1 minute ago, Miska said:

 

That's the case for those filters, but the list ordering is not based on how much computational power is needed, but just grouped as features. And I tend to add new things to the end, unless they are grouped with something earlier.

 

so what IS the most demanding now? sync  L L ? or Li? whichever that is lol 

[KenMoreira]

optimal 
'dsd source settings' ?